All Core Devs - Consensus #148

Transcript

[Music] Okay, great.

Alex Stokes: Cool. Welcome back everyone. I hope you had nice holidays and we will get started now with consist call 148. Uh the agenda here is in the chat. It's issue 1218 on the PM repo. And yeah, I think the main thing today is to discuss DevNet 5 with petra. Uh there's like a few things to touch on there and then a few other things after that. So let's go ahead and get started.

So first up, yeah, DevNet 5. Uh we had gotten I think the specs mostly together before the holidays. Uh people have been working on them since then. So thank you for that. And yeah, just to close out the specs, I wanted to ask if there are any open questions here. Um, I don't know, Harry, I see you're here. Are there any DevNet things to touch on before we dive into the specifics?

Parithosh Jayanthi: Uh, yeah. So, there's just this one open PR. Um, I've listed it in the chat, but otherwise, I think there's nothing um there's nothing on the devet side of things.

uh we have a specific hive instance and I guess um updates since the last ACD there have been execution spec test release as well as consensus spec test releases. So those both are pinned in the spec sheet and we have a hive instance now that runs the execution spec tests and most clients are doing quite well. Um I think Netherm also just commented earlier today that they've already updated and addressed some of the failures. So the next round of hive should pick that up. Um besides this, we're also um doing the execution spec test inside Aerta. And I think Philillips found a few things that he's just clarifying with the testing team and we'll bring it up um probably soon

in chat. All right,

Alex Stokes: great. Uh right, so yeah, this was one open thing that Perry linked here uh 4077. I think we got a little carried away with our versioning. And so this PR basically just rolls back a V2 endpoint to a V1 uh with the blob sidecars uh sort of RPC set of endpoints. Pretty straightforward change. Looks like a number of people already approved it. Anyone here opposed to this change? I don't think there's much to oppose.

Otherwise, we can go ahead and merge this in after the call and get that into DevNet 5. Let's see. Okay, I think that was it on the CL side of things within five specs. I looked at the spec tracker that we have. There was one more EL specific thing just because it was open. I figured we could talk about it now. Uh is this PR? Let's see. This is 9144 to the EIP's repo. An update to 2935 system contract. Uh does anyone have any more context here? Maybe like client this is your PR.

It looks like there are a number of changes but I assume these are good to go.

Uh from what I understand this has already been released on DevNet 5. um like optimistically released on DevNet 5, but yeah um the update like the the PR itself needs to needs some work but it should be it should be merged at some point.

Alex Stokes: Okay, great. Uh can we merge in the next few days just so everything's ready for net 5?

Well, you have to ask Matt, but apart from that, yeah.

Alex Stokes: Okay. Uh okay. I could follow up with him. So cool. And then in terms of specs, was there anything else anyone had or wanted to discuss? If not, uh then we'll move to Par's question here. Where are the CLS on web workflows? So yeah, this also kind of gives us up to the next point which is implementation uh sort of progress and yeah would anyone like to give any updates there? So one we're a client with 5 and then two specifically with

the builder specs and implementation of ME support. Our clients mostly ready to five.

Sean Anderson: Uh so for Lighthouse, we're mostly ready with DevNet 5. We've got like a couple small things to fix, but Lighthouse is like working with itself on uh test nets. Um and the builder spec is a work in progress for us.

So

Terence Tsao: I think Prism has been passing death crytosis, but it's on a net 5 branch.

That's not ideal. We want to run on the latest basically the main branch and then yeah like P said we have a PR that is out already that's that once it's merged you can run latest develop besides that we have a pending builder PR that will start putting more on that will review and merge in the next hopefully in the next few days. So besides that we want to be ready for the builder path as well.

On

Enrico Del Fante: the tech side we we are the five ready master. Um we don't have yet support for SSD builder um flow but the

builder flow should be fine as well but we will double check on that. So good to Um so for load star we are also good to go. I think the only thing we need to do is now revert the v2 uh reg grasp endpoints I guess um because we would start querying those after electra. So and also we support the

build of flow and ssc as well. Great.

Let's see for grand team.

Saulius Grigaitis: Uh the update is that likely everything is ready except the SSD builder flow. So we are quite similar to the other teams.

Alex Stokes: And then let's see. I think Nimbus, I'm not sure if anyone's here, but in any case, uh yeah, good to hear that there's been good progress on deadnet 5.

Uh I will call out this other comment Justin put in the chats, uh beta zero specs release today, so keep an eye out for that. But yeah, sounds like things are moving along pretty well. So if there's anything else to add there, um yeah, I think the next item

Parithosh Jayanthi: I think on the CL side they're mostly good because they're passing ktosis tests as well as CL spec tests. um would be great to hear from the ELS if they're around cuz I think most of the high failures still need to be addressed as well as um like client's point about the 7702 transaction pool if we want to have that invet 5 or not but both of those

are clearly E topics

Alex Stokes: right uh any want to speak to that I suppose we want to have the 702 transaction pool sorted before mainet So that kind of implies it's time to do it now. Uh possibly no EL's on the call who want to speak up.

So, in that case, uh, we could check in on next ACD and get a better sense of timelines, unless someone wants to do something else. Uh, but it seems like that's kind of where we're at. Oh, yeah, there is a testing call on Monday. So, yeah, I mean, yeah, we'll we'll keep checking, I think, every touch point we have. Yeah, by AC2 would be great. Uh, okay. There's some stuff in the chat here.

Okay. I mean, yeah, sounds like everyone's busy and trying to get this as soon as possible. So, hopefully in the next week or two, we can have DNF5. Uh, then from there, I do think it's helpful to talk about timing for test nets for Spectra and then mainet. Um, sort of just like a sketch of one possibility here is, you know, if we have DevNet 5, let me just pull the calendar, you know, the next week or two, that kind of tees us up to start doing test nets in February.

If we have no issues and can move quickly, I think it's possible to do Sapoleon and Hleski in February. Then from there, that kind of tees us up for mainet in March. Anyone feel good or bad about that timeline? Tim asked about devet six. I mean, I would say we do devet six if we need it. So that kind of depends if we see devet 5 uh having issues.

So optimistically there are no issues and we can kind of move forward like I said otherwise yeah we'll have to squeeze in devet six somewhere. no comments. So, I assume everyone's okay with that. At least in theory. Anything else to discuss around vector coordination around dev nuts, test nets and then to mainet.

Okay, Tim has a good question here. What do we think about test net ordering? Would someone prefer Sapoleia Hleski versus the other uh sequence? people are different. I would suggest doing Sapoleia just because uh it's less chaotic and then Hleski. Yeah, there's another comment here with Justin saying the same thing. One thing is if we do that's kind of like a stable app facing test net and I think some people have been asking about testing things like 7702 as soon as they

can. So uh that would be nice. Uh there's a question about ephemery here part of the official deploy schedule. Um yeah I'm not as close to operating ephemery. Is anyone here closer to that who can speak to that? Uh yeah, we can think about uh doing it for FM2, but uh we need client stable releases for that first that include uh all the pectra changes.

Gotcha. Uh should we like block the test nets on ephemery do we think or it's not that important?

Tim Beiko: I think if we have the sepoleia releases, we can probably use those for ephemery for people to start testing and then um it depends as well if we want to do a single client release for Sephogia and Holki. But if we don't then we can use the Sapoleia release for ephemery even if we want to wait to see how that goes before we do Holki and then update it if there's anything wrong.

Alex Stokes: Yeah, that works for me. Which by the way, I do think that's a good point. Like if we can cut releases for Hleski and Sabo at the same time, that will help us pipeline quite a bit and move faster. So something to keep on everyone's radar. Okay. Anything else here? There's a question about higher validator counts on the DevNets.

Yeah, Perry suggesting DevNet 6 with higher validators.

Parithosh Jayanthi: Yeah, I think I can just bump the validator count for DevNet 5. And for six, we can discuss having something with a lot more validators and potentially just doing it as a shadow fork or something so there's also some state to play with.

Alex Stokes: Yeah. Would you want to do that before or you'd be okay doing them at the same time?

Parithosh Jayanthi: I think same time's probably okay as well, but um I just focus for now on DevNet 5 being out the door and if that's done, then everything else can happen in parallel.

Alex Stokes: Yeah, for sure. Okay, any other questions? If not, uh we can move to the next item here. Uh there was a last call for adding uh IP B6 support into the ENR. Let me grab the PR here.

So, yep, there it is in the chat. Uh 3874 and yeah, there's been a lot of discussion on this over the last couple months and I think it's pretty much ready to go. Um yeah, anyone here want to add anything? pop you here and had the comments any more context

uh yeah this PR is about adding uh three new fields to the ENR uh we add we are adding TCP6 quick and UDP 6 the purpose of adding this T field is to to to allow like the the the node to to have uh uh a different port for IPv6 uh than the the port for IPv4 because if

if we don't have this feel we we we can have only like single part that that is used both by IPv4 and IPv6 but yeah that's it.

Alex Stokes: Cool. Thanks. Um yeah otherwise I think uh this PR is pretty much ready to go. So yeah pop I mean I guess you'd like to merge this soon.

Uh yeah if there is no objection I think yeah we can merge like this soon. Okay

Alex Stokes: so yeah everyone listening take another look if this is relevant to you and yeah looks pretty straightforward.

I don't believe we need to gate this onra or anything. So it's here and can go when it's ready. Uh so yeah please take a look when you have some time.

Uh it

it's actually backward compatible. So we don't have to ship it like on any hard. We can just ship it like as soon as possible because it's backward compatible. Right.

Alex Stokes: Cool. Okay. So I think that's all we had for Petra today. Uh anything else? Any final comments? Otherwise we can move on to the rest of the agenda.

So, next up then, uh we have items under pure uh you know scaling the blobs and there wasn't anything in particular. I just wanted to ask if there are any updates. Uh I know that there's been some work on pure devets kind of in parallel with PCRA. People have been pretty heads on Petra over the last, say, a couple weeks, so there might not be anything to report, but I figured I would ask. I think there was a breakout earlier this week, but I couldn't make it. Let me double check. Yeah, it looks like there was one. I don't know if there's anything to share from that breakout.

Okay, I think people have really been focusing on petra which is good but uh yeah expect a lot more pure dos once we get petra to mainet. So next up then uh there is a reminder here to review this document on hardware requirements. Uh let's see is Kev on the call? He put this here. I don't see him. Uh so yeah, I can share a little bit. Um there's this document here and basically it's just trying to define like a minimum set of hardware requirements. This is really helpful for all kinds of things both like communicating to people who want to

interact with Ethereum say as a validator. Um also when designing the protocol and thinking about different parameters of different designs, what you know makes sense and what would not be okay. So this document's here a number of people been working on it over the past say month or two. Um yeah, super cool to see everyone please take a look and I think the concrete ask is if you have any uh feedback on this document uh to send that to Kev. Otherwise, uh yeah, I think that's all we need to say for now on that.

And then otherwise, uh there was just one last thing here on the agenda for the name of the Gar. So, we have a poll here on ETH magicians. Let me grab the link. Uh that's there. And there are a number of names uh from AB Kotup. I'm actually not sure how you say his handle, but in any case, number of names here. And in terms of a poll, uh if you look at the poll, the leading result right now is LOAS. Uh which I also like. I think it's pretty cool. So yeah. Uh oh, wait. Kev is back. Yeah. Let's jump back then.

Kev, do you have anything to add? Yeah, I guess just chime in if you do, but otherwise, yeah, so uh we have Amsterdam, the Gar Glowass is the leading candidate, and yeah, uh don't really want to open up a huge bike shading combo, but unless someone is super strongly opposed, let's move ahead with Glowass for the the name. And then we have a great Portman two here. Glamsterdam.

Okay. Sounds like people are indifferent or they like gloss. So, let's go with that. And I believe that was it on the agenda for the day. Uh super short and sweet call. Anything anyone else would like to discuss? Otherwise, we'll go ahead and wrap up uh early today. Okay. If not, uh, let's go ahead and resolve these few open questions for the DevNet 5 specs, uh, implementations.

Yeah, keep, you know, polishing the things you have. And, yeah, we'll get to Neet 5, uh, as soon as we can. And with that, I think we'll go ahead and close. So, thanks everyone and I'll see you around.

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