AllCoreDevs - Testing #087
Transcript
- Stefan Starflinger
hello people not on holiday
- Pooja Ranjan
gm
- Yassine Ferhane
gm
- spencer
gm gm
- spencer
lets go glamsterdam 🦩
- potuz
GM. Question for CL devs, do you guys queue blocks that arrive more than the clock disparity time in the future? That is a block that arrives more than 500ms early, do you guys queue it or just ignore
- Dustin
Nimbus ignores
- potuz
Thanks this seems to me the rational thing to do
- Nico Flaig
Reacted to "lets go glamsterdam ..." with 🐻❄️
- Dustin
If the EL wasn't able to reorg that's on the EL, not the CL
- Will Corcoran
Reacted to "lets go glamsterdam 🦩" with 🐻❄️
- Eitan Seri-Levi
LH ignores as well
- potuz
Reacted to "LH ignores as well" with ❤️
- Marius van der Wijden
Were passing all
- Justin Traglia
Are CL clients passing alpha.12 tests?
- Edgar | ethrex
ethrex is passing all
- Mario Vega
Reacted to "ethrex is passing all" with 🔥
- Enrico Del Fante (tbenr)
Replying to "Are CL clients passi..." Our devnet-7 should pass all GLOAS
- Eitan Seri-Levi
Replying to "Are CL clients passi..." LH is passing all alpha 12 tests
- Justin Traglia
Reacted to "Our devnet-7 should ..." with 🔥
- Justin Traglia
Reacted to "LH is passing all al..." with 🔥
- Marius van der Wijden
The consume/engine tests for geth (41) are failing because of mismatched error messages, not real failures
- Mario Vega
Reacted to "The consume/engine tests for geth (41) are failing because of mismatched error messages, not real failures" with 👍🏼
- Eitan Seri-Levi
I think lodestar was passing
- terence
Replying to "Are CL clients passi..." Same with prysm
- Stefan Starflinger
Reacted to "Same with prysm" with 👍
- Justin Traglia
Reacted to "Same with prysm" with 🔥
- Dustin
Gas limit?
- Dustin
200M seems fine -- enough to exercise 7688, otherwise similar as devnet 6
- Ben Adams
300M? 🤔
- Justin Traglia
Replying to "200M seems fine -- e..." Agreed. As long as we can submit more than 8192 deposits.
- Stefan Starflinger
Replying to "300M? 🤔" I'm open to it
- Justin Traglia
Replying to "300M? 🤔" Not opposed.
- Marius van der Wijden
Replying to "300M? 🤔" sure
- Toni Wahrstätter
300m at payload exec deadline at 6s sounds reasonable
- Daniel Lehrner (Besu)
Seems okay
- Stefan Starflinger
then lets do it
- Toni Wahrstätter
This is now a CL decision though
- potuz
Notice that you guys are calling Els only but CLs have now a strong saying on gas limits since it’ s mostly about attestation deadlines
- potuz
jinx
- Justin Traglia
Reacted to "Notice that you guys..." with 😄
- terence
Ship it
- Jun Song
Reacted to "Ship it" with 👍
- Satyajit
Reacted to "Ship it" with 👍
- Stefan Starflinger
Reacted to "Ship it" with 👍
- Eitan Seri-Levi
Reacted to "Ship it" with 👍
- Justin Traglia
Reacted to "Ship it" with 👍
- Mario Vega
https://github.com/ethereum/pm/issues/2151#issuecomment-4953484951
- Justin Florentine (Besu)
BOOOOOOOOO!! (the community has spoken)
- Justin Traglia
https://ethereumfoundation.zoom.us/j/89441658268?pwd=yjqf4TheEb1R7ZSuesCdOAy5tdyVk7.1
- Justin Traglia
Replying to "https://ethereumfoun..." This is the CL breakout call link ^
- Will Corcoran
Reacted to "This is the CL breakout call link ^" with ❤️
- Toni Wahrstätter
Replying to "BOOOOOOOOO!! (the co..." Polar bear voter?
- wolovim
summarizing: EL breakout stays here, CL goes to the new link ^
- Toni Wahrstätter
So 350M gas limit?
- potuz
lol
- Edgar | ethrex
Se ha reaccionado a "So 350M gas limit?" con 😂
- spencer
Reacted to "So 350M gas limit?" with 😂
- Mario Vega
- FLCL
Reacted to "lol" with 😂
- Stefan Starflinger
Replying to "So 350M gas limit?" 301M
- wolovim
Reacted to "So 350M gas limit?" with 😂
- potuz
Where’ s the CL breakout @jtraglia?
- Stefan Starflinger
Reacted to "So 350M gas limit?" with 😂
- spencer
Replying to "So 350M gas limit?" oh please this
- Mario Vega
Reacted to "So 350M gas limit?" with 😂
- Stefan Starflinger
Replying to "Where’ s the CL brea..." https://github.com/ethereum/pm/issues/2151#issuecomment-4908406237 https://ethereumfoundation.zoom.us/j/89441658268?pwd=yjqf4TheEb1R7ZSuesCdOAy5tdyVk7.1
- wolovim
Reacted to "https://github.com/ethereum/pm/issues/2151#issuecomment-4908406237 https://ethereumfoundation.zoom.us/j/89441658268?pwd=yjqf4TheEb1R7ZSuesCdOAy5tdyVk7.1" with 👆
- spencer
Its merged! :D
- Stefan Starflinger
Does this break anything? Should we add this to devnet-7 or 8?
- Stefan Starflinger
ah so its already in devnet-7 :)
- spencer
Reacted to "ah so its already ..." with 👍
- Mario Vega
Reacted to "ah so its already in devnet-7 :)" with 👍
- Stefan Starflinger
added to the spec
- jochem-brouwer
One point process-wise, can the spec https://notes.ethereum.org/@ethpandaops/glamsterdam-devnet-7 point to EIP commit? So it is clear what EIP version should be used
- Stefan Starflinger
Replying to "added to the spec" https://notes.ethereum.org/@ethpandaops/glamsterdam-devnet-7
- Stefan Starflinger
Replying to "added to the spec" updated
- Stefan Starflinger
Reacted to "One point process-wi..." with 👍
- FLCL
Does anyone have sprase blobs in their branches
- Stefan Starflinger
Replying to "One point process-wi..." we tried that for devnet-6, but the spec gets outdated quickly that way. Will discuss this with the team again on wed
- Maria Silva
And it was unintented
- Dragan Rakita
Was disconnected, adding floor gas to block space is good to limit amount of data.
- Edgar | ethrex
Respondiendo a "Does anyone have s..." on ethrex we have a pr for sparse blobs but not currently on the glam 7 branch
- spencer
Very much in favor of small tweaks, I'd still like to re read all the EIPs (8037/8038/2780) with the new changes Ideally no more spec changes (consensus breaking) though, unless it makes sense!
- Maria Silva
I like more execution gas, actually
- spencer
Replying to "I like more execut..." same!
- Gary Rong
Fine for this renaming tweak, also think execution gas is better
- jochem-brouwer
Heeft gereageerd op "Fine for this rena..." met 👍
- Stefan Starflinger
sgtm
- FLCL
Replying to "Does anyone have sprase blobs in their branches" do you plan to add it there?
- FLCL
ty cu
- Stefan Starflinger
ty
- Marius van der Wijden
bye
Call summary
Targets
- •devnet-7 launch: end of day July 13 or morning July 14 - 00:15:47
Decisions
- •devnet-7 gas limit set to 300M - 00:18:37
- •EIP-7937 floor-cost block-level fix included in devnet-7 (already in spec 7.2) - 00:25:01
- •EIP terminology cleanup (execution gas rename, log cost consolidation) to proceed for devnet-8; no consensus rule changes - 00:34:13
Highlights
- Testing Progress:
- ·EL tests (release 7.2): Nethermind, Geth, ethrex passing most; Besu failing ~40 (repricing); Reth branch ready tomorrow morning - 00:11:44
- ·Geth 41 consume/engine test failures are mismatched error messages, not real failures - 00:14:07
- ·CL tests (alpha.12): Nimbus, Lodestar, Prysm, Lighthouse all passing; Grandine status unknown - 00:14:39
- El Breakout Eip Cleanup:
- ·Proposal to rename 'regular gas' to 'execution gas' in EIP-7937 for forward-compatibility; no consensus change, spec text only - 00:29:47
- ·Proposal to consolidate log costs in EIP-8037 (remove separate log transfer cost, absorb into base + transfer-with-value); net values unchanged, targets devnet-8 - 00:32:22
- El Breakout Eip 7937 Fix:
- ·EIP-7937 bug: blocks up to 8x expected max size were constructible due to floor cost applied only at tx level, not block level - 00:23:47
- ·Fix already merged in spec release 7.2 and included in devnet-7; only 2-3 tests affected - 00:25:01
- Fork Status And Schedule:
- ·Glamsterdam devnet-6 running at ~80%+ participation overall - 00:08:57
- ·Lodestar/Erigon and Lodestar/Besu nodes stuck after fork; Lodestar/Besu restarted and recovered - 00:09:47
- ·devnet-7 imminent: launching after call or end of day; half-million exited validators, 0x02 credentials, EIP-8282 contract deployed manually - 00:15:37
- ·devnet-7 gas limit set to 300M (raised from 200M by consensus) - 00:16:04
Action Items
- •Maria Silva - Draft EIP cleanup PRs (execution gas rename, log cost consolidation) for devnet-8 - 00:29:47
Key decisions
devnet-7 gas limit set to 300M (raised from 200M by consensus)
devnet-7gas limit was raised from 200M to 300M by consensusEIP terminology cleanup (execution gas rename, log cost consolidation) to proceed for devnet-8; no consensus rule changes
PFIEIP terminology cleanup was proposed for devnet-8