FOCIL Breakout #34

Transcript

Justin Traglia: Hey, Jihoon.

Jihoon Song: Hi, Justin. GM.

Justin Traglia: How's it going?

Jihoon Song: Doing great. Was, half OOO today, so I haven't… checked your, message, but I'm gonna take a look into it after this breakout.

Justin Traglia: Okay. Yeah, it's fine, doesn't really matter. Nothing is, critical.

Jihoon Song: Got it.

Justin Traglia: Did you see the new, FOCIL channel on Discord?

Jihoon Song: Yeah. even the EIP channel, right? Yeah… Do we need this channel, by the way?

Justin Traglia: I don't really think so. There is a FOCIL channel, or sorry, inclusion list. channel. In the same way that there's an ePBS channel, but… I don't know, I mean, it is kind of nice having a singular place for future… Chats.

Jihoon Song: Yeah. I don't know. I can't ask.

Justin Traglia: I think we should actually keep it. That's my opinion.

Jihoon Song: Okay.

Justin Traglia: I don't know.

Jihoon Song: then how is it different from InClosures Channel?

Justin Traglia: It's probably not. It might, it might… Actually, I don't know. No, it's just a different name.

Jihoon Song: Yeah.

Mehdi Aouadi: Maybe we could just… I don't know if it's possible on Discord to move around channels, just move the inclusion list and rename it, because we have the chat history there, and it's quite… Important, in my opinion.

Jihoon Song: Agreed.

Justin Traglia: Yeah. I know we can do that, we did that for the Payload Builders channel. We moved an old one to it.

Jihoon Song: then… Are we gonna also move, like, BAL and 8 PBS?

Justin Traglia: That's a great question, I don't know.

Jihoon Song: Okay. Yeah. We can figure it out. Yeah, the call, I mean, this breakout call, We scheduled one hour earlier than before because of some conflict with other calls. And I think we were gonna stay in the same time in the future. And maybe this is the first time we are just to a new time, so yeah, I would suggest to wait just a few more minutes.

Alright, I think… We can start. Alright, so… Hello everyone, welcome to FOCIL Breakout 34. It's been a month since our last breakout. Today's agenda is ballpark Recap, client development, and spec and testing. So let's start with Svalbard. As everyone would already know, we made a local devnet between Lodestar and Besu. Besu was ready before the interop. Probably because they designated a person working on Hegota. And Nico and his agent built those third prototype during the interrupt. Slightly after the intro, we also onboarded ethrex, And…

Eitan also started working on FOCIL, but unfortunately, the PBS didn't let him go. So, for the next couple of months. What we were gonna do is focusing on stabilization, in my opinion. So we could use this time to find bugs and unknown issues in the spec. And add more tests. So, for example, Nico raised an issue about Encodileist enforcement during non-finality. And during the FOCIL breakout session in the interop, we reached to the conclusion that it's desirable to turn it off, because nodes will be disconnecting with each other, and nodes are busy. EL is syncing, CL is processing test stations from different chains. And if we decide to include frame transaction in Hegota, than… we're gonna change Enclosure list validation logic, so currently we check if Any omitted encryption-based transaction.

Can be appended at the end of the payload, and this will be changed to check at a certain index at the payload, at the end. This is to prevent builders from having to execute verified frames over and over again, so with 60 mil gas limit and 10K gas per frame transaction per enclosure list. A builder might have to execute 100ml gas in the worst case. But again, to clarify, This is when we decide to ship for any transaction, so… this might or might not happen in the future, I mean, this spec change. So, to sum up, we achieved one by one local DevNet and agreed on some spec changes. Such as non-finality and the compatibility with frame transaction.

And this is a recap for the small bird interop. Does anyone have any questions or comments? Otherwise… Let's move on to client development. So we skipped our last breakout right after the Svalbard introp, because people were flying back home. But since then, we also made a good progress on the EL side, Bachelor and ethrex already, but currently ELs Are having some issues with some BAL, so we are waiting for BAL to be stabilized. On the sales side, Teku, Low Star, And lighthouse are synced ready. Can we have more detailed updates per client?

Mehdi Aouadi: I can give an update from the Teku side. So yeah, I published a Docker image. by the end of last week, with a branch which is, which has all our, Gloas changes, merged, so the branch I used has all the Gloas work. So we should have our latest version of ePBS. I ran some interop with Lodestar, which seemed to be working fine. And, after the interrupt, we had some issues related to ePBS, that we've been fixing on Teku. And… since… Yesterday, I started merging those fixes, and I'm close to having a new version of our Docker image with yet more fixes of some issues we had during the interrog. So, yeah, that's from outside.

Eitan Seri-Levi: Nice.

Jihoon Song: Maybe it then?

Eitan Seri-Levi: Yeah, on Lighthouse's side, We have a working implementation, though, I've been seeing that we're orphaning blocks and reorging payloads, maybe, like, one or two of them per epoch, and I think it's an issue with Lighthouse. So I'm trying to debug that. And I had one question, since we're… Moving to the 1.7 alpha… 8 spec, would we just rebase FOCIL on top of that, or how do we plan on doing that kind of stuff?

Jihoon Song: Jason, do you recall any possible changes in 1.7 RFL?

Justin Traglia: Yes, I think you changed the engine API, right? Like, we added a new fill to it. Or something.

Jihoon Song: FCU… FCU V4.

Justin Traglia: I don't recall.

Jihoon Song: Okay, yeah, I think I'll look into it, and then…

Eitan Seri-Levi: Yeah, and there's some, like, ETBS changes that if we included, like, if I included that change right now in FOCIL, I don't think I'd be able to interrupt with the other clients.

Jihoon Song: Okay, since we… don't have any depth yet. I think we are quite flexible, so I would suggest to just… keep targeting the latest release, and once we feel like we are confident that we can launch the first DevNet, then we can breeze our spec to a certain target, and then, yeah. Start from there. Thanks, Justin. So this is the changes. Target gas limit, yeah. So, Yeah, so this is just, changes coming from the PBS, so…

Justin Traglia: Yep, and there could be some underlying, functional differences as well. That's something to keep in mind that wouldn't show up there.

Jihoon Song: Yeah, let me take a look into it, and then… share something, if I found… Tomorrow. Did we include FCU before? Or is it still… ongoing discussion.

Justin Traglia: I really don't recall, the numbers. Don't know. What was the change for that?

Jihoon Song: So the… the… the main… Yeah, yeah, for… I mean, isn't it for the… the… To enable a reorg for the same chain when you want to reorg a payload. Besides that, Our FC version is already good, if I'm not wrong. Yeah.

Justin Traglia: I really don't remember. Someone else would have to answer.

Jihoon Song: Yeah, okay. I can… yeah, let me take a look into it right after this call. And, yeah. So that we can try the first local intro between CL clients. At the latest spec. Yeah, so, thanks for asking, Ethan. Yeah, so… I guess… We don't have anyone from Lodestar and Besu today. Nico is on his vacation. And Besu's not here, but I guess Besu is already in a good shape. So, yeah, do we have any more… Client updates.

Barnabas Busa: Do we expect any more major changes to the FOCIL architecture?

Jihoon Song: No, I don't. It depends. If a… if we decide to include a foreign transaction, then yes, it does change to both CL and EL. If not… I cannot come up with any major spec changes. Alright, I think we can move on. The last agenda is spec and testing. So, as I mentioned earlier, there are a few… minor spec changes into, including not functionality stuff. We're gonna have those changes. Soon TM, but for now, as I said, I cannot come up with any changes to the main part of the spec. still, if we did find any issues or bugs in the spec for, during the, upcoming DevNet, then it could be changed.

for testing? Felix from Steel is helping us on EL testing. EL has basic test vectors and hybrid tests, and… both passen, it regs best all tests. We don't have proper CL testing yet, but CL interrupt works with either BESU or ethrex, only local DevNet, so I wouldn't worry too much about Hegota Interop, and that's all I have for spec-it testing. Do we have any more? Updates, or questions, or comments?

Justin Traglia: Would you mind if I remove the outdated FOCIL testing checklist from the repo? Or do you wanna, like.

Jihoon Song: Mute.

Justin Traglia: Update that.

Jihoon Song: You mean just removing it?

Justin Traglia: Yeah, I would opt to just remove it and come up with a better solution, such as automatically generating it based on tests we already have.

Jihoon Song: Yeah, yeah, sounds great, yeah. It's good to go.

Barnabas Busa: Do we want to launch a public DevNet? Anytime soon. If you do.

Jihoon Song: That depends on…

Barnabas Busa: Stable?

Jihoon Song: DevNet 7, because we cannot launch it, because ELs are disagreeing with each other, because of BAL's issue.

Barnabas Busa: Hopefully that's no longer the issue. 10 minutes of it.

Jihoon Song: Yeah, let's…

Barnabas Busa: DevNet 4, which is based on top of all DevNet 7. But yeah, that was an issue for the previous subnet. So hopefully… hopefully both Evan 7 is pretty stable. And I would say if we can run Glamsterdam devnet, for, like, a week, stable, then we could actually just run a FOCIL.net as well. Which, I guess, at this point, it would not be a FOCIL evidence, but it would be actually a… Hegota W0. Yeah, I've already specified.

Jihoon Song: Yeah.

Barnabas Busa: Well, sir.

Jihoon Song: Yeah, so… We were gonna target… to base off of Glamsterdam devnet 4. For Hegota and Definite Zero on CL and EL, right?

Barnabas Busa: Yeah, ideally, unless something changes in the next week or two. Yeah.

Jihoon Song: Right, so we can't just… make a decision to tentatively base off of Glamsterdam devnet 4 for now, and then… If it stabilized in next week or two, then… We're gonna target lunch the first. I go to DevNet.

Barnabas Busa: Yep.

Jihoon Song: Cool. Great, anything we do want to discuss?

Barnabas Busa: Do we still do this call? Not that I'm here that often, just, Just checking if this is still needed.

Jihoon Song: Yeah, I think…

Barnabas Busa: If…

Jihoon Song: we are in a same situation with the PBS and BAL, so they… Had their breakouts. For the time being, when they needed to discuss some… feature-specific. items, and then they moved on to ACDC. And I think we are in a phase That we might need to discuss some FOCIL-specific stuff. So… I mean…

Justin Traglia: I'm sorry.

Jihoon Song: Yeah, just.

Justin Traglia: I agree. this is the… a good place to discuss FOCIL things until after a glance again. Otherwise, there just won't be time to discuss it at ACDC. Like, I feel like we already don't have enough time there.

Barnabas Busa: Right. Okay.

Jihoon Song: Yeah. I feel like we are still a bit early to… move on to ACDC phase, like, yeah, I think Glamsterdam needs more time for… I mean, need more… focus time on their issues, so, yeah. We can revisit this issue in a, I don't know, a month or two. Okay, cool, Do we have Fabio here? Oh yeah, he just joined. Yeah, Fabio, just, wanted to share our recap that we want to target to base off of Glamsterdam devnet 4, for Hegota dividend at zero. So… Please make sure your best prototype is based on Glamsterdam Demon4.

Fabio Di Fabio: Okay, yeah, we'll, revisit. Yeah, I haven't not touched it since the interrupt, so I think I'm pretty close, I just need to put everything for DevNet 4. Okay.

Jihoon Song: Amazing. Thank you.

Fabio Di Fabio: Is there… is there… target… let's… time for… the DevNet Zero for Hegota.

Jihoon Song: Not really. We are… It's… it depends on… Glamsterdam DevNet4 status, so we're gonna see if Glam DevNet 4 is stabilizing in a week or two, and if it does, then we were gonna target, Hegota DevNet 0 quite soon. So, yeah. Ideally, if everything goes well, then we can expect to have, I call a definite zero in, I don't know, 2 to 3 weeks, or 4 weeks.

Fabio Di Fabio: Okay, nice. Thanks.

Jihoon Song: Awesome. Thank you. Alright, cool. End of May. Yeah, thank you guys for joining and sharing inputs, and… let's keep discussing in the, Discord channel, and… yeah. Thank you, guys.

Fabio Di Fabio: Thank you. Bye.

Mehdi Aouadi: Thank you, bye.

Justin Traglia: Hi, everyone.

Jihoon Song: Bye-bye.

Transcript adapted from ethereum/pm, CC BY-SA 3.0.