AllCoreDevs - Testing #094
Transcript
- jochem-brouwer (Fairphone)
Gm gm
- Pooja Ranjan
gm
- Antoine James
Gm everyone
- Mario Vega
gm!
- Toni Wahrstätter
is it only me or did we always see the chat from the last call
- Justin Traglia
Reacted to "is it only me or did we always see the chat from the last call" with 👍
- Mario Vega
Replying to "is it only me or did..." same
- Mario Vega
Reacted to "is it only me or did we always see the chat from the last call" with 👍
- Justin Traglia
- Parithosh Jayanthi
Reacted to "is it only me or did we always see the chat from the last call" with 👍
- Parithosh Jayanthi
Replying to "is it only me or did..." Its so that you don’t miss the great memes that were sent after you quit the call
- Kevaundray Wedderburn
Replying to "is it only me or did we always see the chat from the last call" I still see it (on mobile)
- Kevaundray Wedderburn
Reacted to "Its so that you don’t miss the great memes that were sent after you quit the call" with 😁
- Mario Vega
Reacted to "Its so that you don’t miss the great memes that were sent after you quit the call" with 😁
- Toni Wahrstätter
Reacted to "Its so that you don’t miss the great memes that were sent after you quit the call" with 😁
- Pooja Ranjan
Reacted to "Its so that you don’t miss the great memes that were sent after you quit the call" with 😁
- Stefan Starflinger
which clients are ready?
- Stefan Starflinger
for snap/2
- Justin Traglia
Reacted to "Its so that you don’t miss the great memes that were sent after you quit the call" with 😁
- Karim T. (matkt)
We have a first version on Besu we want to test
- Iván | ethrex
We are ready to test
- Kevaundray Wedderburn
Can we already do deep reorgs with hive?
- Toni Wahrstätter
is hive glam-ready yet?
- Toni Wahrstätter
This has been the discussion of last week iirc
- FLCL (Nethermind)
nethermind needs one more pr, merging soon
- Karim T. (matkt)
Ok for besu
- Stefan Starflinger
ah and can clients please prioritize getting eth_simulate and eth_estimateGas fixed?
- spencer
Replying to "is hive glam-ready..." some nits to iron out: https://hive.ethpandaops.io/#/group/glamsterdam
- Toni Wahrstätter
Reacted to "some nits to iron out: https://hive.ethpandaops.io/#/group/glamsterdam" with ❤️
- Mario Vega
Reacted to "some nits to iron out: https://hive.ethpandaops.io/#/group/glamsterdam" with ❤️
- Iván | ethrex
Replying to "ah and can clients p..." We fixed eth_estimateGas, but I have a stalled PR with. the implementation of eth_simulate. I’ll see to resume it
- Mario Vega
Is inclusion list equivocation slashable?
- Toni Wahrstätter
Replying to "Is inclusion list equivocation slashable?" no but it can then be ignored iirc
- Stefan Starflinger
Reacted to "We fixed eth_estimat..." with 👍
- Toni Wahrstätter
why are we doing this so accurately whereas other limits are just like 10mib
- felipe
Reacted to "some nits to iron ..." with 👍
- Kevaundray Wedderburn
Replying to "Is inclusion list equivocation slashable?" Yeah not slashable, but proof of the equivocation is needed to ignore it
- Toni Wahrstätter
I see, yeah seems fine!
- Jihoon
Replying to "Is inclusion list eq..." We can provide a proof so technically it's slashable but includers don't get any incentive so we don't slash IL equivocators
- Jihoon
If we introduce some incentive mechanism for includers in the future, we'd probably have one
- Toni Wahrstätter
Replying to "Is inclusion list equivocation slashable?" is the proof of equivovation still required to bring the full ILs or are IL hashes enough
- Justin Traglia
Talking about https://github.com/ethereum/execution-apis/pull/793
- Łukasz Rozmej
I reviewed it and left few comments, but I would love to merge it soon
- Jihoon
Replying to "Is inclusion list eq..." You'd need full ILs to verify signatures
- Tamaghna Chourdhuri
Reacted to "I reviewed it and left few comments, but I would love to merge it soon" with ❤️
- Justin Traglia
Reacted to "I reviewed it and left few comments, but I would love to merge it soon" with ❤️
- Toni Wahrstätter
Replying to "Is inclusion list equivocation slashable?" yeah, we could have better used SSZ to not require the full IL but since it's an edge case it might just be fine
- Mikhail Kalinin
Why it is still a Draft? We need Marius to move the status
- Łukasz Rozmej
Reacted to "Why it is still a Draft? We need Marius to move the status" with 👍
- Jihoon
Replying to "Is inclusion list eq..." FYI, attester slashing also takes full attestations
- Tamaghna Chourdhuri
Replying to "Why it is still a Dr..." Yeah he is ooo I have dropped him a dm last week
- Parithosh Jayanthi
He’s on vacation
- Fredrik
Coming back the 1st
- Mikhail Kalinin
I will do a review this week
- Kevaundray Wedderburn
Probably Mikhail, Felix or Matt
- Toni Wahrstätter
Replying to "Is inclusion list equivocation slashable?" right yeah, but attestations are constant size and not, in theory, up to mib (not the case with focil as it's 8kib max) but in theory txs are dynamic size
- Jihoon
Replying to "Is inclusion list eq..." No, max size of IL txs is 8KiB. So it's also bounded.
- Toni Wahrstätter
Replying to "Is inclusion list equivocation slashable?" yeah that's what i meant
- Toni Wahrstätter
Replying to "Is inclusion list equivocation slashable?" but up to the 8kib it's still dynmic. compared to an attestation that isn't
- Jihoon
Replying to "Is inclusion list eq..." Well, if it's bounded, it's constant anyway
- Jihoon
Replying to "Is inclusion list eq..." It's O(1)
- spencer
The self destruct one I think we can leave
- Justin Florentine (Besu)
and is there even a realistic option to refuse it, or is it just an update that clarifies a gap?
- spencer
The other maybe we should consider, I'm still looking at it
- Mikhail Kalinin
Reacted to "Yeah he is ooo I have dropped him a dm last week" with 👍
- Justin Florentine (Besu)
devnet10 is launching quick, in a couple days, right?
- Parithosh Jayanthi
Devnet-9, in a day
- Parithosh Jayanthi
But it isn’t public
- Parithosh Jayanthi
Like it’ll only be there for days
- Kevaundray Wedderburn
Replying to "Is inclusion list equivocation slashable?" My understanding is that the average and worse case size for an attestation is the same, where the average size of an IL and the worse case are not necessarily the same because of the dynamic property Toni mentioned
- Kevaundray Wedderburn
Replying to "Is inclusion list equivocation slashable?" Though agree that the worse case is bounded
- Jihoon
Replying to "Is inclusion list eq..." Yeah my point is it's O(1) by defintion
- Jihoon
Replying to "Is inclusion list eq..." Not Θ(1)
- Parithosh Jayanthi
@clients ?
- Luis Pinto | Besu
I haven’t looked into these 3 issues personally
- Iván | ethrex
Reacted to "I haven’t looked int..." with ☝️
- Justin Florentine (Besu)
i think since this is being described as an edge case, that means a devnet would be the best way to discover and test it.
- Stefan Starflinger
@Maria Silva can you share the links to the issues for everyone?
- Mario Vega
- Maria Silva
https://github.com/ethereum/pm/issues/2204#issuecomment-5461653738
- Stefan Starflinger
Reacted to "https://github.com/e..." with 👍
- Mario Vega
Reacted to "https://github.com/ethereum/pm/issues/2204#issuecomment-5461653738" with 👍
- Justin Traglia
CL breakout: https://ethereumfoundation.zoom.us/j/89441658268?pwd=yjqf4TheEb1R7ZSuesCdOAy5tdyVk7.1
- Toni Wahrstätter
execution-apis PR #869: https://github.com/ethereum/execution-apis/pull/869
- Mario Vega
- Mario Vega
https://github.com/ethereum/pm/issues/2204#issuecomment-5461653738
- Mario Vega
- Kevaundray Wedderburn
Does it break their usecases or make it more expensive?
- Stefan Starflinger
can't they just deploy their contract and use it for whatever they are doing
- Mario Vega
Replying to "Does it break their ..." Iiuc just makes it more expensive because you are paying for a state gas for a contract that was not created
- Kevaundray Wedderburn
I think given self destruct has been unofficially deprecated, it’s unclear to me that we should add this in
- spencer
If we are potentially fully removing self destruct in Hegota (its PFI'd) then maybe we don't add this one.
- felipe
Reacted to "I think given self..." with 👍
- felipe
Reacted to "If we are potentia..." with 👍
- Maria Silva
Replying to "Does it break their ..." The issue is not breaking, is making more expensive
- Guru
Agree. Given that self destruct has been marked deprecated for a while now, don’t think this should change
- Iván | ethrex
Same for us
- Luis Pinto | Besu
Talking personally - I feel it’s best to have this all cleaned up instead of relying on the removal of selfdestruct. Someone in the community might oppose to that
- FLCL (Nethermind)
need time
- felipe
Reacted to "Agree. Given that ..." with ➕
- felipe
It was officially deprecated, no?
- Justin Florentine (Besu)
anything near self-destruct gives me the willies
- Maria Silva
Pawel unmuted
- Łukasz Rozmej
Reacted to "anything near self-destruct gives me the willies" with 😱
- spencer
Reacted to "anything near self..." with 💯
- FLCL (Nethermind)
let us check if it is very small safe change
- Mario Vega
Reacted to "Pawel unmuted" with 👍🏼
- draganrakita
Reacted to "anything near self-destruct gives me the willies" with 💯
- felipe
Reacted to "anything near self..." with 💯
- Kevaundray Wedderburn
I measured the storage savings from self destruct and it seemed pretty insignificant(unless I made a mistake); I think it was less than 200MB a year
- lightclient
are people using this functionality?
- Maria Silva
Avg. of 1500 accounts per day
- Maria Silva
State savings is minimal, yes
- Kevaundray Wedderburn
Replying to "are people using this functionality?" Yeah I found that two companies were using it for user deposits
- Mario Vega
Reacted to "anything near self-destruct gives me the willies" with 💯
- Kevaundray Wedderburn
Replying to "are people using this functionality?" Generate a deposit address, have user send funds to it, then create account and sweep it to companies hot/cold wallet
- Kevaundray Wedderburn
Replying to "are people using this functionality?" As opposed to having N accounts per user which implies N private keys to custody
- draganrakita
Cool pic
- spencer
Reacted to "Cool pic" with ➕
- Kevaundray Wedderburn
Reacted to "Cool pic" with ➕
- Iván | ethrex
Reacted to "Cool pic" with 🐻❄️
- spencer
Reacted to "Cool pic" with 🐻❄️
- Kevaundray Wedderburn
Reacted to "Cool pic" with 🐻❄️
- felipe
Reacted to "Cool pic" with 🐻❄️
- felipe
Reacted to "Cool pic" with ➕
- Maria Silva
I think it would be at end of transaction
- Maria Silva
If Ire call correctly
- Kevaundray Wedderburn
Reacted to "I think it would be at end of transaction" with 👍
- draganrakita
Reacted to "I think it would be at end of transaction" with 👍
- Luis Pinto | Besu
Reacted to "I think it would be at end of transaction" with ➕
- Mario Vega
Personally I’m now leaning towards making a decision today because of complexity, if we want to make the change we need to start ASAP, and it seems like majority is leaning towards no-go
- Guru
Replying to "I think it would be ..." Then this would make it the first case to refill state gas at the end of tx, right? I think all other end of tx refunds are execution gas
- Guru
Reacted to "Personally I’m now l..." with 👍
- felipe
Reacted to "Personally I’m no..." with 👍
- Stefan Starflinger
Reacted to "Personally I’m now l..." with 👍
- spencer
Reacted to "Personally I’m no..." with 👍
- Maria Silva
Reacted to "Personally I’m now leaning towards making a decision today because of complexity, if we want to make the change we need to start ASAP, and it seems like majority is leaning towards no-go" with 👍
- draganrakita
Reacted to "Personally I’m now leaning towards making a decision today because of complexity, if we want to make the change we need to start ASAP, and it seems like majority is leaning towards no-go" with 👍
- Kevaundray Wedderburn
I think we also had 3298 in H*?
- spencer
Replying to "I think we also ha..." tcreate?
- Guru
Replying to "I think we also had ..." refund removal
- Kevaundray Wedderburn
Reacted to "refund removal" with 👍
- spencer
Reacted to "refund removal" with 👍
- spencer
Reacted to "tcreate?" with 👎
- draganrakita
Imo it is more important to ship Glamsterdam and decrease any risk for that
- spencer
Reacted to "Imo it is more imp..." with 💯
- Guru
Reacted to "Imo it is more impor..." with 💯
- felipe
Reacted to "Imo it is more imp..." with 💯
- felipe
Reacted to "Imo it is more imp..." with ➕
- Stefan Starflinger
Reacted to "Imo it is more impor..." with ➕
- Kevaundray Wedderburn
I think there is no alternative?
- Mario Vega
- FLCL (Nethermind)
any other simpler solution?
- Maria Silva
You can still do it, but it will cost more gas
- felipe
Yeah SD should really highlight to us the friction points on how to deprecate an instruction and how to solidify this process for the future. We need better process for this, one that is clear.
- Iván | ethrex
Reacted to "Imo it is more impor..." with ➕
- Luis Pinto | Besu
No matter how much you deprecate people will use it until you pull the plug, being making it expensive or not supporting it
- Luca Donno | L2BEAT
Reacted to "No matter how much you deprecate people you use it until you pull the plug, being making it expensive or not supporting it" with 👍
- spencer
Reacted to "You can still do i..." with 💯
- felipe
Replying to "Yeah SD should rea..." We should try to set a precedent with this, ideally.
- spencer
Replying to "You can still do i..." for me this is feature not a bug on the assumption we are deprecation SD
- spencer
Reacted to "No matter how much..." with 👍
- spencer
Reacted to "We should try to s..." with 💯
- Fabio Di Fabio
Reacted to "No matter how much you deprecate people will use it until you pull the plug, being making it expensive or not supporting it" with 👍
- Kevaundray Wedderburn
I think making a feature more expensive is also a part of the deprecation cycle to deter new usecases from reaching for it
- felipe
Reacted to "I think making a f..." with 👍🏼
- felipe
Reacted to "I think making a f..." with 💯
- Guru
Reacted to "I think making a fea..." with 💯
- milen | Erigon
Replying to "I think making a fea..." As long as you have an alternative first
- Iván | ethrex
Reacted to "I think making a fea..." with 👍🏼
- spencer
Reacted to "I think making a f..." with 💯
- Luis Pinto | Besu
Are we talking about this one right? https://github.com/ethereum/EIPs/pull/12265
- Maria Silva
Reacted to "Are we talking about this one right? https://github.com/ethereum/EIPs/pull/12265" with 👍
- Guru
Reacted to "Are we talking about..." with 👍
- Mario Vega
- Luis Pinto | Besu
Yeah exactly, it’s only for state
- Luis Pinto | Besu
So the freed gas can only be used for state, right? Or maybe I’m misremembering 😅 the spec has changed so much
- Justin Florentine (Besu)
gas golfers gonna hate that
- Maria Silva
So the freed gas can only be used for state, right? No, we changed this
- spencer
Maybe we can also async chat in the 8037 r&d channel?
- Luis Pinto | Besu
Reacted to "So the freed gas can only be used for state, right? No, we changed this" with 👍
- draganrakita
Can we get example of usage, similar what we had in Svalbard?
- Luis Pinto | Besu
Reacted to "gas golfers gonna hate that" with 🕳️
- jochem-brouwer (backup)
thanks all :D
Call summary
Targets
- •Glamsterdam devnet-9 (non-finality) launches September 1, 2026 (tomorrow) - 00:08:05
- •Sepolia shadow fork: mid-to-end September 2026 (contingent on client releases) - 00:13:22
- •EIP-8037 cross-frame bug (PR #12265) decision: ACDC call, September 4, 2026 - 00:28:16
Decisions
- •DFI for Glamsterdam: EIP-8037 self-destruct ephemeral refund (PR #12249) — no-go; complexity and self-destruct deprecation trajectory cited - 01:01:26
- •Snap2 shadow fork starts on Sepolia (not mainnet); mainnet shadow fork follows later - 00:13:22
Highlights
- Focil Updates:
- ·Barnabas proposes raising signed inclusion list P2P message size bound from 8 KB to 41 KB to account for dynamic transaction list; no decision yet - 00:13:55
- ·Barnabas proposes new gossip rule rejecting inclusion lists with empty transactions; to be discussed in CL breakout - 00:18:06
- ·Engine API GetInclusionList endpoint: minor change to ensure it never returns empty transactions; EL teams asked to review - 00:19:00
- Eip 8037 Issues:
- ·EIP-8037 self-destruct refund (EIPs/PR#12249): majority opposes adding state refund for ephemeral create+self-destruct pattern; decided no-go for Glamsterdam - 00:37:33
- ·EIP-8037 cross-frame state gas bug (EIPs/PR#12265): fix prevents regular-gas-to-state-gas laundering across call frames; consensus-breaking spec change, decision deferred to ACDC - 01:02:06
- Testing Progress:
- ·Sepolia shadow fork planned after devnet-9 stabilizes and client releases are ready; mid-to-end September target - 00:09:10
- ·Mainnet shadow fork to follow Sepolia; needed for Snap2 block-access-list healing and heavy EL state ePBS testing - 00:09:57
- ·Besu has first Snap2 implementation ready to test on shadow fork; ethrex also ready - 00:11:05
- Rest Ssz Engine Api:
- ·REST SSZ Engine API PR (execution-apis #793) is spec-complete; needs client reviews to unblock merge into SSZ branch - 00:19:39
- ·PR still in Draft because Marius is on vacation; Justin to contact repo maintainers to promote status - 00:22:14
- Fork Status And Schedule:
- ·Glamsterdam devnet-8 running well; minor orphan blocks, builders/validators onboarded - 00:06:09
- ·Nimbus EL, Reth, and Grandine still need to move to trunk branches before devnet-9 - 00:07:15
- ·Devnet-9 (non-finality network, ~1,000 nodes) smoke-testing now; targets launch tomorrow morning EU time - 00:08:05
- ·EIP-8037 spec bug (cross-frame state gas refill) could be consensus-breaking; risks ~1-month delay to Sepolia if a new devnet cycle is required - 00:28:16
Action Items
- •All EL teams - Review and approve execution-apis PR #793 (REST SSZ Engine API) to unblock merge - 00:22:39
- •Maria Silva - Post EIP-8037 PR #12265 discussion to Discord EIP-8037 R&D channel; gather async opinions before ACDC - 01:13:01
Key decisions
DFI for Glamsterdam: EIP-8037 self-destruct ephemeral refund (PR #12249) — no-go; complexity and self-destruct deprecation trajectory cited
DFIComplexity and self-destruct deprecation trajectory citedSnap2 shadow fork starts on Sepolia (not mainnet); mainnet shadow fork follows later
Testing sequence established for Snap2 shadow forksEIP-8037 cross-frame state gas bug (PR #12265) decision deferred to ACDC
Consensus-breaking spec change regarding regular-gas-to-state-gas laundering; decision deferred to ACDC call on September 4, 2026
EIPs discussed
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.