RPC Standards #16

Transcript

Loading developers...

Oh, [music] [music] heat. Yeah. Okay.

Heat. Heat. [music] >> [music]

>> Heat up here. >> Heat. Heat.

RPC Call Introduction

Thank you. Good afternoon. Um, welcome to RPC call standard call 16. Today's date, December 8th, 2025. Um we don't have much on the agenda but I think K you uh seems is not here the first which is um Jason APC error course analization since he's not here to update us on that front I think we can go over to the exhibition API versioning by K. Sure, that sounds good. Um, let's see. I can drop a link in the chat.

>> Let me shut that. Oh, we got it. Okay. Um, so basically what I want to do and what I think would help client developers is to um have versioning information in execution APIs. And this is mostly for um the hive specs so that they can start passing the hive um the RPC compat hive tests um so they're not trying to hit a moving target because basically right now anytime something gets merged um a hive test breaks somewhere along the way. Um so I have some comments here. I have addressed the feedback. Um, I tagged Felix to see if he had any other um,

feedback. Um, I'll probably give it a couple more days. Um, and then, uh, merge it in and then get something up in Hive where we can run against the both the version, the tagged version and then also like, um, like an unstable version so that clients can see what's coming down the line. Um, but if there are any client developers here who have opinions or anything like that or feedback, let me know. Um, or feel free to comment on the PR as well.

Um, I don't have any other updates there besides that. So, that's it from my end. Okay. Um, that's good. I wanted to also point out something about um but I was hoping Sims and um this other guy would be here because I think they're the ones majorly in charge of open ROPC to talk about um proper documentation and also the website on its own because I feel like we there are some things that not in the website already and I would actually like for us to talk about how we will make um

um the website much more proper and then to align very well with also with um the open arc but I I'm not seeing them here although I tag them on discord so maybe we can keep that out for the next call what do you think >> yeah that sounds good to me I know Zane um is working, Zayn with Open RPC is working on uh making the site a lot nicer. Um and should have um something coming in on the 24th, I think. Um they have a grant deadline from the EF, so they should have

something around then um to share as well. So Oh, okay. I wasn't aware of any upcoming um enhancements, but that's okay. Um I think that's all we have on our agenda. So if you have anything you want us to discuss um about the ROC or the repo or any open Ps also please the floor is yours. Oh, okay. I guess we don't have any

other thing. Then I will also like to point out some open PRs, especially the ones we have that have been still for a very long time. Uh, I will leave that to you, Kira. It's there like any possible feedback for this. >> Um, do you have ones that you're thinking about specifically? I want to um I'd really like to get this versioning in first because um >> you know once we start adding stuff it'll start breaking the RPC compat tests again.

>> Um so that's kind of what I have been waiting on but um but if there are um specific ones that you want me to take a look at, I can definitely do that. >> I'm trying to drop them on the chat. >> Okay, perfect. And okay for the in since um Sims is not here but if anyone has any uh points about the JSON ROPC error code standardation um that is ongoing like any improvements or suggestion this is the PR we can leave our comments there.

Do you know um has anything changed in the JSON IPC error codes PR meter? >> No, not nothing has been changed yet. There's it's still like the last time we left it. So I'm saying because our last discussion we were supposed to like move over to the ACD call to get more client views on it but I did not hear from Sims. So, I was expecting if we if I'm able to get back or get in contact with him before this week, we can move it over to AC/DT next week Monday, we can add it to the agenda so that any improvement or for um suggestion, we can

also align it properly. >> Cool. Yeah, I think um I think it'd be good to get it on the ACDT call next week even though um it looks like Simpson has um address some of the comments already. So, I would say let's get it on the call on the agenda next Monday and then hopefully we can either get more feedback or or um move towards merging it. >> Yes. And there was I think um I saw a comment about one of the error codes. I can't remember where. I think it's on discord.

And now I can access my discord. Um yeah about um it gets log error. Um, let me just drop the links the trade. So, is this something we need to take a look at or we need to add to the agenda also? >> Hello.

>> A good question. [laughter] I'm not sure. Um, I can take a look. I sort of think uh Felix might have merged this one, but I can double check and see if a response is needed or if we need to um put it on one of the agendas. >> Okay. Like I was thinking this should be also be part of the addition the error codes we were working on. >> Yeah, I think it should. >> Okay.

No. >> Um, that is from my end. I don't know if any other thing. >> Nothing for my end. >> Okay. So um we als nikita I'm sorry if I pronounced your name wrongly. Do you have any any imputes for this call before we call

it a day? >> I guess I can say hello. Um I'm Chase. I just joined the EF and my responsibilities are around node infrastructure and RPC stuff. So, this is my first time joining the call, so I'm kind of just listening in, getting a feel for what you guys are doing before I try to have any opinions on anything. >> Okay, thank you for joining us. Um, thank you, Katya. Okay, so that's basically all for today. I guess we'll have to come again next week. Next two weeks, rather.

New Participant Introduction: Chase joins EF, focuses on node infra

Oh, okay. Sorry, you can unmute. >> Yeah. Uh, hi guys. I'm from Nethermind team. Um, I'm working on RPC compatibility now. And yeah, uh, I wanted just to discuss few uh, G internal issues that we face in Nethermind, but I probably need to talk to someone from GU team. Um okay so the idea of this call is to convey any information or to discuss anything pertaining to RPC.

Um we can have a broader discussion and then we we might if there is no time left we can still continue on the discord the Jason channel. >> Yeah I will raise on discord. I talked to Felix about it but um on >> Okay. Can you make an intro about it? Can you like >> Yes. So um it's a issue with transaction types. So inside G ignores them and never mind doesn't. So we check transaction types on RPC calls and like uh act according to transaction type.

Yeah, that's the issue. >> Okay. Um, Kira, do you have anything to say to this? >> Okay. Sorry. Can you summarize that one more time? I didn't understand. >> Yes. So, for example, when you do is call guess and put transaction type uh into transaction uh guess will ignore it. just remove from JSON entirely. But in Nethermind, we will use this transaction type and execute transaction as transaction of this type.

>> Gotcha. Does that come out in the uh JSON RPC like what comes back or >> Yeah. Yeah. Yeah. So for for example like if you put like transaction type two and blob hashes guess will treat this as blob transaction and never mind it will take this 1559 AP transaction. Yeah I think um unfortunately yeah I think the channel would be the best place to bring that up um with the guest devs. One of my longer term goals is to get um G off of the as not the spec and start using the

execution specs repo instead. Um but that's I don't that won't happen very soon. Um but hopefully within the next year. Yeah. Okay. That's actually something else I wanted to discuss because I think that's where most of the issues come out from because uh if guest get is missing something and or maybe he's not complying to something the whole test kind of flaky at the end of the day. I I need to share okay I will share a notion page on discord. I did made a draft about um such cases alongside um Robocop. I think that was the first issues we looked at

at the beginning of this year. So we I'm not sure how we're going to go about it, but I just wanted to ask is the spec um is there a test for this um transaction on ROPC test gen? >> Uh no, not really. May maybe you could write the test that works with Nethermind and then GU would have to fix it.

>> Yeah, but um I'm not sure that we should do like Nethermind. So maybe G is right. Maybe no one's right. I don't know. >> Yeah. Okay, we should answer that question first. >> Yeah. And guess the tricky part is that GU is the spec. >> Okay. Well, you say GU is the spec, but like that's not true in so much that the specs are written after GU, >> right? But they uh return the response values for any given request. There's a tool that we when you put in a JSON IPC request, it goes through GF and so whatever comes

back is the GF response which then becomes All right. Sure. But that's an implementation detail, not get is the spec. >> Sure. >> Hey, I think we need to consord because I don't think there's any client get client de here. Um, that's think that's all for today. See you guys next two weeks probably.

>> What's up? [music] >> Thank you.