All Core Devs - Execution #241

Transcript

Pooja Ranjan: I think so. Hello, everyone. Just wanted to confirm. Akash would be helping with the streaming today. So let us know whenever you are ready.

Nixo: Awesome. Thank you, Ak

Morning, everyone. We'll give it a few minutes before we get started.

Justin Florentine: Good morning, Ethereum.

Nixo: Okay, I think we can get started. Akash, are you ready to go?

Akash | ECH: Yeah, we are live.

Nixo: Okay, great. Welcome everybody to All Core Devs… Execution number 2241. So it's July 16th. We have a super short agenda today, so I, I think that we can get through it pretty fast today. The 1st things first.st Stefan, are you here? We'll start with Glamsterdam devnet updates.

Stefan Starflinger: Yeah, sure. So we launched 7 at 7 on Tuesday, and we had the Gloas fork on Tuesday evening. European time. And it was a little bit chaotic in the beginning. We created the network with half a million exited validators to increase the state so that we can better test the new SSZ changes. And it showed. But clients were quick to fix everything up, and then we ended up onboarding nearly all of the clients by now. So right now, I'm in the process of adding Grandine and Erigon. I think we'll be ready tomorrow. We have, 3,000 validators, with, 1K ETH, 1,024 ETH each. And I should have increased kind of the deposit. Queue speed, because, depositing such large validators takes a long time.

And so I had to reshuffle the validators a little bit, which took some time, but I think everyone should be weighted equally now on DevNet 7 with 100 validators each. So… And right now we're running, Normal spamming scenarios and fuzzing scenarios as well. There were quite a few bugs found on the network. I'm not gonna go over all of them. But I think most of them were pretty small and could be fixed quite quickly. Which is very nice. And I think right now, we have a very good participation rate of around 80-90%, which is great on DevNet 7.

Nixo: Great. Is there going to be a DevNet 8?

Stefan Starflinger: Yeah, pretty sure that that would make sense to have another DevNet. Since it was quite unstable still, I think 1.8 would make sense, and after that, going for the testnets. I think we also still have the final repricing numbers that are open. from the execution side, and there is some cleanup work in the process regarding the IPs with minor changes. to, I guess, wording and just overall polishing.

Nixo: Okay, great. And we are… we are still work… Waiting on those, repricing benchmarks from Devnet 7. Correct.

Stefan Starflinger: I think regarding the benchmarks, there has been some progress, but not as fast as, I guess, the DevNets have been progressing. So they're still a little bit behind. They're currently working towards, we're currently working towards DevNet 6 for benchmarking. And I think it should be ready quite soon.

Nixo: Gotcha. Okay, cool. Do we have any other… any other Glamsterdam stuff? Anything that clients need to look at, pay attention to, work on. Okay, we're getting close. Reminder that once we get testnet updates, or once we start moving on to testnet, there will still be the 30-day, security, the audit review period, and then 2 weeks before each testnet, and then a 30-day, 30-day notice period for the rest of the ecosystem to upgrade, so, still a minimum of 3 months.

Stefan Starflinger: One thing I'd, think might be worth mentioning is that, DevNet 7, is for, And app developers to try out and see the impact for them against the DevNet. So feel free to use the faucet, create contracts, and reach out if there's anything you need help with.

Nixo: Awesome. Is there a resource that they can go to, to, to figure out how to set that up?

Stefan Starflinger: Yeah, I can share the It Panda Apps Network link.

Nixo: Okay. Nico, I don't know what you mean by the temporary public test.

Nico Flaig: Oh, I thought last, on, like, a week ago or something, we talked about something similar to Mekong, that we had for Elektra. Oh, I remember. It was, like, for 3 months, we had, like, this testnet up, or maybe not even 3 months. It was shorter, but it allowed, like, apps and so on to test.

Nixo: So is that the this version of the is that Devnet 7.

Stefan Starflinger: As far as I know, I think I agreed with Maria also that DevNet 7 should be the one that developers should try out. Before we go to the testnets.

Nico Flaig: Okay, but… but what do you mean, like, testnets like Sepolia, or…

Stefan Starflinger: Yeah, yeah.

Nico Flaig: Yeah, because, for example, for Fusaka, we kind of broke a lot of setups, because people were not aware of, like, supernode changes and, like, layer 2s, right? I was just… just thinking if… if it makes sense to have a more public network like Mekong. I mean, there's nothing that prevents them from using a definite for that. I'm just not sure that's the venue we want to use.

Stefan Starflinger: I don't have a super strong opinion. So far, the talks were that we would use DevNet 7. I'm happy to use DevNet 7. I don't think it will be too unstable from my perspective.

Nixo: Nico, is there any other reason why a DevNet isn't appropriate for that?

Nico Flaig: I don't see a real reason, I guess. I mean, is there like a faucet and people can get the ETH and so on?

Stefan Starflinger: Yeah, everything like that is set up.

Nico Flaig: Yeah, and also what Dustin says, like, people just ignore this, it's not really… Like, we have zero actual builders on the DevNet right now, so… I mean, there's just a lot of surface with glass and clumps that I mean.

Dustin Brody: I actually want to add a little, I mean. To that as well, which is to say. even if we can get people to not ignore it, I'm not sure that's a great So, because NetNet pins a very specific spec, version in place, and, and right now, like, it has been useful to be able to, sort of expeditiously remove older spec version running, DevNets. Like, DevNet previously, DevNet, say, previously Five. For example, we didn't have to keep it running because it was just for the core devs to test. their own implementations. And if we now say, but technically anybody can actually connect to them, that also creates a bit of a social expectation that it will keep running and keep being maintained, and I think that may be more costly than it's worth.

Stefan Starflinger: I mean, that's kind of the long-running thing, that's what testnets are for, and if we create a testnet that is short-lived, then we have the same problem as.

Dustin Brody: Right. True. So what I would say is converge to a near, at least very, very near final spec first. like, and then… so that we don't have that divergence and kind of backporting, and… and not… and worse than that, even… so that's one way of framing it, it's like, well, that's more work for Core Dev, who cares? But it also means you're not testing what' Which is worse, and it's worse for everybody.

Nixo: So is there any reason that DevNet 8 can't be that middle ground?

Stefan Starflinger: I am worried that with DevNet 8, we're so close to the testnets already that it would make more sense to go to the testnets. So in my opinion, it would make sense to give clients or developers, I mean, app developers time now before we move them to the testnets.

Nixo: Okay. in that case, let's just, try to publicize DevNet 7 as a very short-lived, temporary place where app developers can test if they don't want to test on Sepolia or Hoodi, and break… break things on there, because I think that was the concern last time, is that they have things running, and that, testing on there broke things. So, if they want to avoid that, they can use DevNet 7. We'll just have to make a concerted effort to actually, reach out to app developers and make sure that they're aware of this temporary testnet. Devnet, testnet, thing. Okay, if that, is everything else for Glamsterdam? Anything else? Then, I will mention that the community has,

been having input on the mascot. There was a vote on ETH Magicians. It appeared that the flamingo won, but, there was some consternation about botting and some shared IP addresses, so there's a runoff vote between Flamingo and, pandas. or polar bears, please… please go vote, so that there's not drama about this, because this just sounds like the… Yes, the most important topic. This sounds like the the just an absolute, We should not spend a bunch of time fighting over this. So please go vote the, here, I'll drop the, I'll drop the link here. Oh. Great, thank you. And if nothing else for Glamsterdam, we'll move on to Hegota. Carlos, do you want to, Carlos is going to propose EIP 8268 storage routes in block access lists. Carlos, are you here?

Not. Yeah, I… Marius is also not here. I had put that at the end of the agenda, but I guess that makes sense to include so I… So Marius had an option or an addition to the agenda. He would like to take five minutes to discuss REST SSE, in particular, the missing non-engine methods that are exposed on the engine port. Is there anybody who's not me, because I don't… because I can't speak to this, who could speak to this? If not, I will leave it at the end of the agenda. Maybe Marius will show up in the next couple of minutes, maybe not. I think he's at an in-person event.

he basically flagged, our other clients, if other clients need the same four methods, are there any additional methods that are needed or useful, and can we retire the ETH1 deposit path, Yes, thank you for putting those there. So, if clients could take a look at that, otherwise, I will have… Yeah. Marius Ping, who is necessary on that. Okay. Then, in the meantime, I will, go back to Hegota. So, just a reminder that, the non-headliner closing date is August 6th. That is an ACDC week. So for execution layer proposals, just PR your EIP against the Hegota Meta EIP, that's 8081.

by August 6, and if you have proposed it, if you have PR'd it into the Meta by August 6, feel free to propose it on the next call, which is 7 days later. And then that'll be considered in time. And I just wanted to remind people that the status of the Hegota headliners is that FOCIL is the headliner. There was a question of Frame transaction, or native account abstraction. And I just wanted to remind people, because people forget, like, what the outcome of things were. We CFI'd frame transaction as a placeholder for native account abstraction, assuming that the people who are working on those different implementations of native account abstraction would work on converging upon some sort of consensus for native account abstraction.

So, native account abstraction is CFI'd for it, Hegota, it is not necessarily a headliner, which means that we wouldn't wait for it. We left it open to possibly propose it as a headliner. after we close the non-headlining EIPs, we will have another conversation about that. About the status of. getting consensus around native account abstraction. Okay. I think Carlos still isn't here. He PFI'd, 8268 storages in block access lists. I will ask if he wants to propose that at the next… if he wants to present something at the next, ACDE. Otherwise, I will just, consider this a notification that It is being PFI'd. And then… The last item on the agenda is some housekeeping from Pooja. Pooja, do you want a couple minutes?

Pooja Ranjan: Yes, I would love that. Thank you so much, Nexo, and Hello, everyone! I'm back with some of the housekeeping agenda item. I'm quickly going to share my screen so we can go through them. so this is a list of Glamsterdam eip Which should be promoted to review status as per my understanding. We have shared this in the previous meeting as well. We have received approval from multiple authors and those PRs have been closed. Thank you for approving it and moving the proposal to the review status. But there are still seven EIPs which are still open. Some of them have already received approval, but it is dependent on other eips, and that's why it is not being merged. So if you see your name here, kindly go ahead and take a look at the Pr. And if find appropriate, please do approve it so we can promote the Eip.

Next thing here is about Devnet 7 and the Glamsterdam Sfi stage. So I made an assumption that if proposals are in the Glamsterdam Devnet 7, then it should be moved to Sfi. My bad, and thanks to Jocham, it has been corrected right now. But at this point I would like to quickly check with the testing team and client team. Are they comfortable enough to move all these EIPs to SFI stage, assuming that these EIPs are more stable. To give some context for people to understand how this is being considered, we have tried to, I mean, I have tried to put it together, this list based on my understanding. So if we are in Upgrade DevNet, almost close to Public Testnet, so when we move into Public Testnet, it will be in the last call, so if we are in Upgrade DevNet, I believe all the proposals should be in the review and SFI stage. So if you find this is the right time.

Can we have some confirmation from testing team on that? Or if there is a recommendation, we should still wait. I'm open to the idea.

Nixo: I think the process is now that there are four, milestones to hit for eips to be moved to Sfi and Acdt will determine when those are met, and then just ratify it on Acd or Acd. ACDC or ACDE. So as long as those four milestones are hit for those EAPs, I think it's fine to move those. And those are all laid out in 7, 7, 2, 3.

Pooja Ranjan: Right, but just for clarification purposes, do we need to wait for another ACDT call to get it ratified there, or should we come back on the next ACDE call to, like, confirm that they are ready to move in SFI stage?

Nixo: I don't think we need to, like, ratify it on an ACDT, but I think it's the testing team that should be init.

Pooja Ranjan: Initiating the stage change, you mean? Like, creating the PR?

Nixo: Yeah, as long as you flag with one of the testing teams that all of those four, things have been hit, I think that, it is fine to… Say, FYI, the testing teams have… have… determine that these 4 things have been met, and we're moving them to Sfi.

Pooja Ranjan: Sounds good to me. I will go back and take a look at it and maybe come back to the ECD channel to confirm that with the testing team. Apart from that, there are a couple of open PRs. I think one is open PR. If authors may get a chance, please take a look. I feel like the status and stage confusion should not be there. This PR is still open. Please take a look, add your feedback. If this is not appropriate and something missing, we can take a look at it. And the final one was there is one PR which was there to add BPO folder, but it was closed. I'm not sure if it is

by mistake, or is there a reasoning for that? But I'm happy to accept the feedback on the PR itself. And that's pretty much it from my end. Thank you so much for the opportunity.

Nixo: Yep. Thank you, Pedro. Anything else? If not, this could go down as the shortest ACDE. Possibly, ever. Anybody got anything else? Go Glamsterdam, Hegota, anything else? Perfect. Let's go get Glamsterdam Thank you. Everyone.

Justin Florentine: Excellent work, everyone.

Pooja Ranjan: Thank you.

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