eth_multicall Meeting, Mar 10, 2025

eth_multicall and eth_simulate Meetings · 10 Mar 2025 · 6:52

A meeting of the developers working on eth_multicall and its successor eth_simulate, held on 10 March 2025. Client and tooling teams compare implementations and work on the specification for simulating transactions. This meeting covers a client implementation update and the spec for eth_simulateV2.

Topics

From eth_multicall and eth_simulate Meetings, hosted by Ethereum Foundation.

Transcript

great I propose we get started Kari is not going to be here this week he's on vacation oh um yeah sorry I'm in a bit of a nosy environment um but I can get started basically so what I did is uh we wrote up this document on the estimate gas Alternatives that we have and I communicated that to uh two teams um and I'm still waiting a response there so not much of an update apart from that uh somebody raised an issue on a repo um they would like to

add the ability to overwrite the beacon route and withdrawals for a given block did they indicate why they want to do that uh yeah I can just post the issue and that's basically the idea is that um if simulat can simulate the kind of consumate the chain um and it is just missing these things and once we have it they could use it in their testing setup uh to test uh an execution Cent or a consensus client I

think the idea is it's very close to being something that will allow you to basically run a full Fork of main net in memory and all it needs is that and then it would be there I think yes that's a good um I'm not terribly opposed to it these are just two parameters on the execution block block right yep yeah I'm not against it if there were two parameters that we just kind of missed I think it's fine be override over should we add them to

B2 I think these are things that we can add to V1 already they're not breaking in any sense they just be additive or purely additive mhm if all the clients that have implemented this so far are okay with it I'm okay with that to be one we should get make sure we get Buy in from um I think there's two other clients that have implemented right besides NE mind gu I'm guessing they'll be okay with it but we should just make sure so a make sure they're aware of it that there's a change and make sure they have no problem adding it in the one thing I'd want to avoid is having like some clients have this and some clients not and when you want to use this RPC method you have to kind of probe to see what's

support and what's not but if all the major clients have it implemented then as long as you're on a fairly recent version which as of the next Fork everybody will be then I think it's fine if we just update the V1 there two clients where re I think implemented it right and B is it Basu okay yeah so we should just are they in our telegram channel so by what name we know yeah U they are in the

telegram and we didn't hear back from bisu um it seemed like they were passing quite some tests and it was because of the parent has issuing G they could not pass some of the test so I'm wondering now that has been fixed where they are ads yeah I do not know I will see if I can get a hold of so um G Gabriel is from B and I'll ask them in telegram if they're okay with this propos this is just adding sorry what are the two

Fields you can route and with WS okay assum we can get buy from them I'm okay with adding override support that NE Minds um test sweet going you guys get everything past uh for me I was mostly looking through the dress end points so almost implemented but I on the test side uh still I'm facing the issue where parent has is not matching to the previous block has is that still the case

what um I'm asking if it's still the case because it was an issue and it's been fixed and I believe the tests have been regenerated so it would be nice if you can rerun and see if um on the fixed sorry the re I against see it and I I did rerun it oh and they're still failing yeah so we have a second second bug in the same area

presumably but I aftering points I'll look into this okay be great if we can get a third client working on the test just so that way we have a at least a two of three so when we have these failing these mismatch failures we can have a educated guess who's got the problem right now it's just never mind Basu is just up in the air the bugs in one of them so hopefully Bas or re can get Hive tests to the same point we have anything else guess not okay talk to you guys

next week then