AllCoreDevs - Execution #234

gm

Red added read.ai meeting notes to the meeting. Read provides AI generated meeting summaries to make meetings more effective and efficient. View our Privacy Policy at https://www.read.ai/pp Type "read stop" to disable, or "opt out" to delete meeting data.

Gm

Gm

gm

gm gm

gm - will get started in just about a min

Reacted to "gm - will get starte..." with 👍

Heeft gereageerd op "gm - will get star..." met 👍

Reacted to "gm - will get star..." with 👍

Reacted to "gm - will get star..." with 🐙

Reacted to "gm - will get star..." with 🐙

Reacted to "gm - will get star..." with 👍

Reacted to "gm - will get starte..." with 🐙

Reacted to "gm - will get starte..." with 👍

Reacted to "gm - will get starte..." with 👍

Reacted to "gm - will get starte..." with 👍

Good luck Nixo! ❤️

Reacted to "Good luck Nixo! ❤️" with ❤️

Reacted to "Good luck Nixo! ❤️" with ❤️

Reacted to "Good luck Nixo! ❤️" with ❤️

Reacted to "Good luck Nixo! ❤️" with ❤️

Heeft gereageerd op "Good luck Nixo! ❤..." met ❤️

Reacted to "gm - will get starte..." with 🐙

yay, nixo ftw

Reacted to "Good luck Nixo! ❤️" with ❤️

Reacted to "Good luck Nixo! ❤️" with ❤️

Reacted to "Good luck Nixo! ❤️" with ❤️

Reacted to "Good luck Nixo! ❤️" with ❤️

Heeft gereageerd op "yay, nixo ftw" met 🔥

Reacted to "yay, nixo ftw" with 🔥

Reacted to "Good luck Nixo! ❤️" with ❤️

Reacted to "gm - will get starte..." with 👍

Reacted to "Good luck Nixo! ❤️" with ❤️

Reacted to "Good luck Nixo! ❤️" with ❤️

Reacted to "Good luck Nixo! ❤️" with ❤️

A lot of those got addressed, would like to see them so we can discuss them

Reacted to "Good luck Nixo! ❤️" with ❤️

@FLCL is looking at this from Nethermind

Reacted to "Good luck Nixo! ❤️" with ❤️

This sounds like we're missing tests for those edge cases

Reacted to "Good luck Nixo! ❤️" with ❤️

Reacted to "This sounds like we'..." with ➕

the CL/EL devnets should just merged

Reacted to "This sounds like we'..." with ➕

Reacted to "This sounds like we'..." with ➕

Heeft gereageerd op "This sounds like w..." met ➕

Replying to "the CL/EL devnets sh..." soon tbd

Reacted to "This sounds like we'..." with ➕

Reacted to "This sounds like we'..." with ➕

Replying to "This sounds like w..." Have a PR for the create geth split, but worth discussing whether the behaviour is correct/incorrect: https://github.com/ethereum/execution-specs/pull/2639

Replying to "the CL/EL devnets sh..." While both are breaking? Makes not much sense right now

Reacted to "soon tbd" with 👀

"One thing I realized yesterday is that we are refunding the state gas for access lists into the stateGasBudget instead of the regularGasBudget. I don't think we have a test for it."

Replying to "From Marius earlie..." Also in this PR: https://github.com/ethereum/execution-specs/pull/2639 Also have access lists tests

Reacted to "Also in this PR: htt..." with 👍

Heeft gereageerd op "Also in this PR: h..." met 👍

https://github.com/ethereum/EIPs/pull/11475

Reacted to "Also in this PR: h..." with 👍

Reacted to "Also in this PR: h..." with 👍

Want to say yes they should have been caught by tests but finding every edge case is very hard based on the scope of 8037!

Reacted to "Want to say yes they..." with ➕

Reacted to "Want to say yes they..." with ➕

Reacted to "https://github.com..." with 🔥

Reacted to "Want to say yes they..." with ➕

Replying to "Want to say yes they..." its a lot

https://github.com/ethereum/execution-apis/pull/760

Replying to "Want to say yes they..." 8037 has really put our testing framework through a stress test

Replying to "Want to say yes they..." Yeah agree, super hard to figure out all potential cases

Reacted to "Want to say yes th..." with ➕

Replying to "Want to say yes they..." I meant it more like, now that we know the issue, we can create a test for it such that it's easier for clients to fix and validate their fix

Reacted to "I meant it more li..." with 🙏

Reacted to "I meant it more like..." with 🙏

https://github.com/ethereum/execution-apis/pull/770

Reacted to Have a PR for the cr... with "🔥"

Reacted to Want to say yes they... with "➕"

Replying to "This sounds like we'..."

//github.com/misilva73/evm-gas-repricings/blob/main/reports/eip-8037/spec_review_state_gas_accounting.md

Reacted to @spencer is this iss... with "🔥"

Reacted to "I meant it more li..." with 💯

Replying to "From Marius earlier ..." @Maria Silva we should probably add access list issue to https://github.com/misilva73/evm-gas-repricings/blob/main/reports/eip-8037/spec_review_state_gas_accounting.md

safe block is currently useless

as justified

@Stefan Starflinger should we discuss bal-devnet-4 vs gloas-devnet-0?

justified is some random gasper thing which shouldn't be exposed anyway in this context

Replying to "@Stefan Starflinger ..." do you expect any changes in bals ?

Replying to "@Stefan Starflinger ..." Those should be less concerning as it does not include additional mechanisms but are “just”constant change. I lean more on bal-devnet4 but don’t mind the merged devnet.

nope

resumed

Reacted to "resumed" with 👍

tests are ready

Heeft gereageerd op "tests are ready" met 🔥

Reacted to "tests are ready" with 🔥

Reacted to "tests are ready" with 🔥

Isnt epbs-devnet dead as of now?

Reacted to "tests are ready" with 🔥

yeah but we are working on it

Reacted to tests are ready with "🔥"

Replying to "Isnt epbs-devnet dea..." How can you merge it by end of april?

Replying to "Isnt epbs-devnet dea..." fix it by end of this week 😄

Reacted to "tests are ready" with 🔥

Agree with dragan, re having a devnet-4 with dynamic cpsb for 8037

Reacted to "Agree with dragan, r..." with 👍🏼

devnet-4 = devnet-3 + dynamic 8037 spec fixes + stability

Reacted to "Agree with dragan, r..." with 👍🏼

Reacted to "Agree with dragan, r..." with 👍🏼

Reacted to Agree with dragan, r... with "👍🏼"

Reacted to "devnet-4 = devnet-..." with ☝️

Reacted to "devnet-4 = devnet-3 ..." with ☝️

Reacted to "Agree with dragan,..." with 👍🏼

Reacted to "Agree with dragan, r..." with 👍🏼

Reacted to "devnet-4 = devnet-3 ..." with ☝️

Heeft gereageerd op "Agree with dragan,..." met 👍🏼

Reacted to "devnet-4 = devnet-3 ..." with ☝️

Replying to "devnet-4 = devnet-3 ..." yes

Replying to "devnet-4 = devnet-3 ..." next week wednesday?

Reacted to "next week wednesda..." with 👍

Reacted to "next week wednesday?" with 👍

Reacted to "next week wednesday?" with 👍

why not ERC, why not before Hegota: https://github.com/ethereum/ERCs/pull/1658 ?

We have had several AA ERCs. It is time for native AA.

you didn't have one written by core dev

Headliner = we would delay Hegota if its not done

read before judgibng, please!

Sorry, I read NFT controlled smart account and didn’t read further. Are we seriously considering that? :-). I will read and comment there. But, I think its unfair to dismiss other ERC proposals just because they aren’t core developer

No more half measures

IS this title deeds?

Full AA or bust

Full AA. No more ERCs.

Reacted to Full AA. No more ERC... with "👍"

all AA designs are half-measures. always have been

Reacted to "all AA designs are h..." with ❤️

Reacted to "all AA designs are..." with 👍

Reacted to "Full AA. No more ERCs." with 👍

Reacted to all AA designs are h... with "👍"

Reacted to "all AA designs are h..." with 👍

Reacted to "Full AA. No more ERCs." with 👍

Reacted to "Good luck Nixo! ❤️" with ❤️

Wallet deeds seem very cool

@Pedro have you read the proposal, or just jumping to conclusions?

Work the usecase from the usecase, not from the tech

Replying to "@Pedro have you read..." Anything that can be done as an ERC only is not real AA

Native is just an adjective. Pick the right tool

If "Account Abstraction" as a headliner is still controversial, does Post Quantum Infrastructure make sense as a "topic headliner"? Especially given the recent news?

There is discontent around 8141 - it does make descisions that can't be backed out. Reth is opposed to it.

Reacted to "If "Account Abstract..." with ➕

Post Quatum doesn't need Frames

Reacted to "Post Quatum doesn'..." with ➕

Reacted to "Post Quatum doesn't ..." with ➕

Reacted to "Post Quatum doesn't ..." with ➕

Replying to "all AA designs are h..." when native key delegation

Reacted to "Post Quatum doesn't ..." with ➕

Replying to "There is discontent ..." Of course there are decisions that cannot be backed out… because its an EIP that requires EL changes

Replying to "all AA designs are h..." right after address space expansion

Reacted to "Of course there are ..." with 😄

Reacted to "right after address ..." with 👀

Awesome pictures

Replying to "all AA designs are h..." don’t need address space expansion, you can do it like 7702

Reacted to "Awesome pictures" with ➕

Reacted to "Awesome pictures" with ➕

Replying to "Awesome pictures" You are awesome

Reacted to "Awesome pictures" with ➕

Reacted to "Awesome pictures" with ➕

Reacted to Awesome pictures with "➕"

Replying to "Awesome pictures" NotebookLM

Reacted to "Awesome pictures" with ➕

Hella cool

Reacted to "Awesome pictures" with ➕

Reacted to "You are awesome" with ❤️

Reacted to "don’t need address s..." with 👆

This has existed for a while, it’s nothing new, and it doesn’t solve anything about what we do we current EOA

Isn’t this similar to ERC6551?

Maybe we should CFI something like frames too

8221 requires centralized transaction relayers or users to have separate gas accounts to use public memoool and FOCIL. solving those are a non negotiable for AA in my opinion

Reacted to 8221 requires centra... with "👍"

Reacted to "8221 requires centra..." with 👍

Reacted to 8221 requires centra... with "👍"

Reacted to "8221 requires centra..." with 👍

Reacted to "8221 requires centralized transaction relayers or users to have separate gas accounts to use public memoool and FOCIL. solving those are a non negotiable for AA in my opinion" with 👍

Why are we talking about more ERCs? It's insane. An ERC cannot solve the EOA problem.

I literally built this in 2019~2020

Reacted to "I literally built this in 2019~2020" with 👍

Reacted to "I literally built th..." with 😄

Reacted to "8221 requires centralized transaction relayers or users to have separate gas accounts to use public memoool and FOCIL. solving those are a non negotiable for AA in my opinion" with 👍

Reacted to "Why are we talking about more ERCs? It's insane. An ERC cannot solve the EOA problem." with 👍

Heeft gereageerd op "8221 requires cent..." met 👍

Reacted to "Why are we talking a..." with 👍

Reacted to "I literally built th..." with 👍

Reacted to "8221 requires centra..." with 👍

Reacted to "Why are we talking about more ERCs? It's insane. An ERC cannot solve the EOA problem." with 👍

Reacted to "I literally built th..." with 😄

Honestly there should be a rule that ERCs should not be discussed in ACDE

Replying to "I literally built this in 2019~2020" very forward thinking 😉

Reacted to "Why are we talking about more ERCs? It's insane. An ERC cannot solve the EOA problem." with 👍

The idea is not new. It has been built before (by non core developers)

Reacted to Honestly there shoul... with "👍"

Reacted to I literally built th... with "👍"

Reacted to I literally built th... with "😄"

Reacted to "Honestly there should be a rule that ERCs should not be discussed in ACDE" with 😂

Reacted to "very forward thinkin..." with 😁

Reacted to Honestly there shoul... with "👍"

Replying to "Why are we talking a..." not to lose FOCIL/public mempool or statelessness

lol at ben saying “i’m not prepared” and having a beautiful 14 slide presentation queued up

Reacted to "lol at ben saying “i’m not prepared” and having a beautiful 14 slide presentation queued up" with 😂

Reacted to "lol at ben saying “i..." with 😂

Reacted to "lol at ben saying ..." with 😂

Reacted to "lol at ben saying ..." with 😂

Reacted to "Honestly there shoul..." with 👍

Replying to "I literally built this in 2019~2020" I was late to the party. Fabian did that with 725 before me

Reacted to "lol at ben saying “i’m not prepared” and having a beautiful 14 slide presentation queued up" with 😂

Reacted to "lol at ben saying “i..." with 😂

Reacted to lol at ben saying “i... with "😂"

Reacted to "Honestly there shoul..." with 😂

Replying to "Why are we talking a..." you loose one of them with frames

Replying to "Why are we talking a..." currently

Reacted to "lol at ben saying “i..." with 😂

Reacted to "lol at ben saying “i..." with 😂

Replying to "I literally built this in 2019~2020" Just without the nft registry part

Reacted to "lol at ben saying “i..." with 😂

Reacted to "lol at ben saying “i..." with 😂

Replying to "Why are we talking a..." You lose them here, because you will be forced to use a centralised relayer.

Reacted to "I literally built th..." with 😄

Reacted to "lol at ben saying “i’m not prepared” and having a beautiful 14 slide presentation queued up" with 😂

Reacted to "lol at ben saying “i..." with 😂

“a third party with eth” not decentralized

Reacted to "Why are we talking a..." with 👍

Reacted to "8221 requires centralized transaction relayers or users to have separate gas accounts to use public memoool and FOCIL. solving those are a non negotiable for AA in my opinion" with 👍

Reacted to "“a third party with ..." with 👍

Reacted to "“a third party with ..." with 👍

Reacted to “a third party with ... with "👍"

Reacted to "Honestly there shoul..." with 👍

Replying to "“a third party wi..." Couldn't it be you with Eth?

Reacted to "Couldn't it be you w..." with 😅

Reacted to "Why are we talking about more ERCs? It's insane. An ERC cannot solve the EOA problem." with 👍

Replying to "“a third party with eth” not decentralized" Exactly. This is a very hard thing to decentralized. So far 4337 is the solution to decentralizing this

Reacted to "Honestly there should be a rule that ERCs should not be discussed in ACDE" with 👍

Reacted to "lol at ben saying “i’m not prepared” and having a beautiful 14 slide presentation queued up" with 😂

Replying to "“a third party with eth” not decentralized" no because smart accounts can’t originate tx without native AA. 4337 works but you need a bundler. which means you need a middleman for FOCIL

Replying to "“a third party with eth” not decentralized" Saying « you don’t need 4337 » just show you don’t understand what the problem is

Reacted to "Saying « you don’t need 4337 » just show you don’t understand what the problem is" with 👍

Reacted to "Saying « you don’t n..." with 👍

Reacted to "“a third party with eth” not decentralized" with 👍

Reacted to “a third party with ... with "👍"

Replying to "“a third party with ..." I believe this has in mind a lot more than just AA. It tries to reason further with other protocol-parts in mind. The goal isn’t to just do AA. But do it on a way compatible with the rest of features we plan to ship.

q: will txpool contain signatures for gasless txs? (ig no)

Reacted to "lol at ben saying “i..." with 😂

Replying to "“a third party wi..." So the core of AA to you, @lightclientis that a contract can initiate a transaction?

Reacted to "I believe this has i..." with 👍

Reacted to "I believe this has i..." with 💯

Reacted to "lol at ben saying “i..." with 😂

AA is the headliner

Replying to "“a third party with eth” not decentralized" yes. we have good mechanisms for the execution abstraction portion of AA

Replying to "lol at ben saying “i..." He has magic powers

Reacted to "lol at ben saying “i..." with 😂

We had multiple votes from the majority of client teams against frame transactions as a headliner

Reacted to "lol at ben saying “i..." with 😂

Replying to "“a third party wi..." Thanks.

Reacted to Honestly there shoul... with "👍"

Reacted to "We had multiple vote..." with 👍

Reacted to "We had multiple vote..." with 👍

Who are “the AA folks” who “will be taking this decision"? Is there a separate meeting where this decisions are taken?

Reacted to "lol at ben saying ..." with 😂

pedro@walletconnect.com

this breakout will be tracked at https://github.com/ethereum/pm/issues ?

Reacted to "this breakout will b..." with 👍

Reacted to "this breakout will b..." with 👍

I think “AA” was CFI’d

3 teams were in favour, no?

Reacted to "3 teams were in favo..." with 👀

Reacted to "3 teams were in favour, no?" with 👍

Reacted to "3 teams were in fa..." with 👀

Reacted to "3 teams were in favo..." with 👍

Replying to "3 teams were in favo..." yes

It is not 1 team.

Gotta count the numbers beforehand giulio

Replying to "Who are “the AA folk..." yes, having a public breakout call to resolve a “best path forward” for an AA proposal. if we can find consensus among a majority of clients, it makes sense to discuss making it a headliner. AA folks = authors

Reacted to "3 teams were in favour, no?" with 👍

Replying to "this breakout will b..." yes

Reacted to "yes" with ❤️

Heeft gereageerd op "3 teams were in fa..." met 👍

Reacted to "Gotta count the numbers beforehand giulio" with 😂

Definitely more than 1 team was in favor of Frame Transactions and 2 transactions were neutral

Łukasz Rozmej: My problem with frames and all native AA proposal I read I can't combine: native AA statelessness/partial stateful FOCIL/public mempool My concerns arn't political, my concers are purely technical.

5 years spent, so broader community need to spend how much time to agree on them? 1 month?

Reacted to "My problem with fr..." with ➕

Reacted to "My problem with fram..." with 👍

Reacted to "My problem with fram..." with ➕

Reacted to "My problem with fram..." with 👍

https://github.com/ethereum/EIPs/pull/11481

Reacted to "My problem with fram..." with ➕

https://github.com/ethereum/EIPs/pull/11481" with 👍

https://github.com/ethereum/EIPs/pull/11481" with ❤️

https://github.com/ethereum/EIPs/pull/11481" with ❤️

Reacted to "btw i did open a PR ..." with ❤️

Reacted to "btw i did open a PR ..." with ❤️

Reacted to "btw i did open a PR ..." with ❤️

Replying to "3 teams were in favo..." Who is the rhird? I forgto ethrex

AA can’t be allowed to block PQ

Wonder if this discussion would be more useful in the breakout

Reacted to "AA can’t be allowed ..." with 👍

Reacted to "AA can’t be allowed to block PQ" with 👍

Adding a precompile don’t solve the fact that we need a tx.origin, and that tx.origin is an ECDSA. Without alternative to ECDSA for creating protocole transactions there is nothing ERC/precompile can do

Reacted to "Adding a precompile ..." with 👆

Replying to "3 teams were in favo..." AFAIK, Besu were ok with it. They just didnt want it to be a headliner.

Reacted to "AA can’t be allowed ..." with 👍

Replying to "Adding a precompile ..." Which is what the scheme txs did

Reacted to "Adding a precompile ..." with 👍

Replying to "Adding a precompile ..." But AA != PQ. You can solve the second without the first

Replying to "Adding a precompile ..." And ofc the second with the first

Reacted to "But AA != PQ. You c..." with 👍

Reacted to "But AA != PQ. You..." with 👍

Replying to "3 teams were in favo..." No they were not

Reacted to "But AA != PQ. You c..." with 👍

Replying to "3 teams were in favo..." Its 2 then

Reacted to Adding a precompile ... with "👍"

Replying to "3 teams were in favo..." I asked this question to the Besu team explicitly in the last ACDE and they responded that they wont block if 8141.

Replying to "3 teams were in favo..." Thats not support

Replying to "Adding a precompile don’t solve the fact that we need a tx.origin, and that tx.origin is an ECDSA. Without alternative to ECDSA for creating protocole transactions there is nothing ERC/precompile can do" AA is not « one thing », it’s many things!

you can't minimize and focus an intentionally abstract feature

Replying to "3 teams were in favo..." Neither is it against. We are getting into semantics.

Replying to "3 teams were in favo..." I spoke to them too and they largely prefer schemed, both can be true

Replying to "Adding a precompile don’t solve the fact that we need a tx.origin, and that tx.origin is an ECDSA. Without alternative to ECDSA for creating protocole transactions there is nothing ERC/precompile can do" Gas payment by a third party for smart contracts: 4337

now, if you want to focus on PQ, that can be minimized and focused. AA cannot.

Replying to "My problem with fram..." is there any proposal on how to make frame txs compatible with statelessness?

Reacted to "IS this title deeds?" with 👍

Reacted to "now, if you want t..." with 👍

Replying to "Adding a precompile don’t solve the fact that we need a tx.origin, and that tx.origin is an ECDSA. Without alternative to ECDSA for creating protocole transactions there is nothing ERC/precompile can do" Allowing EOA to behave like smart contracts (including 4337): 7702

Replying to "My problem with fram..." I don't know of one

Reacted to "But AA != PQ. You..." with 👍

Reacted to "I don't know of one" with 🥲

Replying to "My problem with fram..." And it is inheritely impossible, you need state to validate transaction

I think the core of requiring execution and full state access before inclusion is the thing about frames that makes a lot of other (non-frame transaction) things difficult.

Replying to "Adding a precompile don’t solve the fact that we need a tx.origin, and that tx.origin is an ECDSA. Without alternative to ECDSA for creating protocole transactions there is nothing ERC/precompile can do" Saying « this solves AA » makes no sens. You have to say what problem about AA it solves

Replying to "My problem with fram..." well potentially if you include witnesses/zk-proofs of their validity with them - maybe

And those who disagree with the proposal do not attend the breakout call as well

Replying to "My problem with fram..." Strategy 2 makes it compatible. BUT, if we at some point need more, the mempool participant number will def sense it

Replying to "My problem with fram..." VOPS was designed with native AA in mind: https://ethresear.ch/t/a-pragmatic-path-towards-validity-only-partial-statelessness-vops/22236#p-54075-vops-and-native-account-abstraction-aavops-9 This was written before frames, but frames was designed with VOPS in mind

Reacted to "well potentially if ..." with 👍

Reacted to "VOPS was designed with native AA in mind: https://ethresear.ch/t/a-pragmatic-path-towards-validity-only-partial-statelessness-vops/22236#p-54075-vops-and-native-account-abstraction-aavops-9 This was written before frames, but frames was designed with VOPS in mind" with 👍

Replying to "My problem with fram..." THAT IS NOT TRUE

Reacted to "VOPS was designed wi..." with 😮

Replying to "Adding a precompile don’t solve the fact that we need a tx.origin, and that tx.origin is an ECDSA. Without alternative to ECDSA for creating protocole transactions there is nothing ERC/precompile can do" IMO a problem worth solving is « how to permanently move EOA to smart accounts, getting rid of the ecdsa key ». « How do we make EOA that are not ECDSA » is also an interesting problem to solve (for all the devices that have P256 keys)

Reacted to "THAT IS NOT TRUE" with 👍

Replying to "My problem with fram..." Wait sorry which part?

Reacted to "Saying « this solves..." with 👍

Reacted to "THAT IS NOT TRUE" with 🤔

Replying to "My problem with fram..." I wrote this with some colleagues. We just mentioned that it could work. But was very hard and would be a very minimal AA. Otherwise, if we need arbitrary state, all crumbles

Yeah, one theme of one of the ACDEs re 8141 was people looking for one bullet point or one tl;dr for why 8141 specifically was good, what was it "really" about. Seems to have been an allergy to its general-purpose abstraction angles

Reacted to "IMO a problem wort..." with 👍

specifically a couple of ELs disliked it exactly because it's a general solution

Replying to "My problem with fram..." AA-VOPS works only if we go for Frame with Strategy 2 or Schemed (or any ERC ofc)

Reacted to "I think the core of ..." with 👍

Reacted to "AA can’t be allow..." with 👍

Replying to "My problem with frames and all native AA proposal I read I can't combine: native AA statelessness/partial stateful FOCIL/public mempool My concerns arn't political, my concers are purely technical." why is it so minimal?

this is the second call in a row where we've been holding ad-hoc "votes". not a thing we normally do.

Reacted to "this is the second c..." with 👍

Reacted to "this is the second..." with 👍

Reacted to "this is the second c..." with 👍

Reacted to "this is the second call in a row where we've been holding ad-hoc "votes". not a thing we normally do." with 👍

Replying to "My problem with fram..." “Canonical Paymaster” should answer the question no? 🙂

Reacted to this is the second c... with "👍"

frames are cool. We need to start prototyping it asap and facing tx pool issues. If it will be very bad at it we may fork without it. that's why not headliner

Reacted to "this is the second c..." with 👍

Replying to "My problem with fram..." Let wallets adopt whatever they want as contract. And all the sudden, you’ll be in deep trouble tracking arbitrary state

Replying to "My problem with frames and all native AA proposal I read I can't combine: native AA statelessness/partial stateful FOCIL/public mempool My concerns arn't political, my concers are purely technical." did you look into actual paymaster implementations?

Replying to "My problem with frames and all native AA proposal I read I can't combine: native AA statelessness/partial stateful FOCIL/public mempool My concerns arn't political, my concers are purely technical." i have looked at them and they are very similar

Replying to "My problem with fram..." They are 3 fields. Yes. Sure. But that doesn’t mean this is what will end up being used in the long run

Reacted to "IMO a problem worth ..." with 👍

Replying to "My problem with fram..." So you promise compatibility “now”. I agree. But future? No guarantees IMO. And if we ship, we’re locked there. And the implications for statelessness and FOCIL will be high

Reacted to "3 teams were in fa..." with 👎

Reacted to "So you promise compa..." with 👍

Reacted to "this is the second c..." with 👍

most aggressive should be finalized epoch ?

Reacted to "most aggressive shou..." with 👍

For erigon, be as aggrwssive as you want

Reacted to "IMO a problem wort..." with 👍

Replying to "most aggressive shou..." weak subjectivity period?

I think most aggressive is weak subjectivity

minimum 82125 epochs in Nethermind

Replying to "minimum 82125 epochs..." where is this num coming from?

@Łukasz Rozmej if the Ethereum public pool does Strategy 2 (local state only + slot restrictions), doesn’t that give us native AA + statelessness + mempool safety?

Replying to "minimum 82125 epochs in Nethermind" as specified in eip-4444

90k is arbitrary

Replying to "minimum 82125 epochs..." so 1 year

Replying to "My problem with fram..." I don’t understand sorry, if strategy 2 is compatible with VOPS, what’s the problem? Accounts that need arbitrary state will not be able to use the mempool and we can specify it as a fundamental tradeoff, but still allow the option rather than removing the tradeoff point from the space

Replying to "minimum 82125 epochs in Nethermind" I believe we just drop pre-merge by default though

Replying to "most aggressive shou..." If it is a finalized epoch, you won't be able to save the blocks as ERA files or in any other way.

Reacted to "I don’t understand s..." with 👍

Replying to "most aggressive shou..." we don’t need all clients to be able to save era files. We are talking about most aggressive, not least aggressive

Replying to "most aggressive shou..." yeah, I agree

Reacted to "we don’t need all cl..." with 👍

wouldn't you have to ask the cl?

Replying to "wouldn't you have ..." for CLs it's ~finality

Replying to "wouldn't you have ..." reorgs don't go beyond that

Replying to "wouldn't you have to..." the cl knows the wsp

One year is too long; a couple of months, let's say three months, is better

Replying to "My problem with fram..." Yeah same question as Luca above. Would be helpful to know the pushbacks against adopting strategy 2 (or a variant of that) in the mempool while keeping the underlying AA protocol flexible

Replying to "wouldn't you have ..." oh right, yeah, that's a CL calculation

Yeah ~18D

Replying to "wouldn't you have to..." seems like a good thing to expose so the EL can ask

Reacted to I don’t understand s... with "👍"

sorry its 4096 epochs not 8k eitehr way its 18.2d for blob expiry

Reacted to "I don’t understand sorry, if strategy 2 is compatible with VOPS, what’s the problem? Accounts that need arbitrary state will not be able to use the mempool and we can specify it as a fundamental tradeoff, but still allow the option rather than removing the tradeoff point from the space" with 👍

Reacted to "Yeah same question as Luca above. Would be helpful to know the pushbacks against adopting strategy 2 (or a variant of that) in the mempool while keeping the underlying AA protocol flexible" with 👍

keep in mind that when blocks grow post glamsterdam those numbers will also grow

I think 2 weeks is fine

Reacted to "keep in mind that when blocks grow post glamsterdam those numbers will also grow" with 👍

+1 for one month (or some round number of epochs around that time)

Reacted to "+1 for one month (or some round number of epochs around that time)" with 👍

Reacted to "+1 for one month (or..." with 👍

Reacted to "keep in mind that wh..." with 👍