ePBS Breakout #14
ePBS Breakout · 20 Dec 2024 · 8:18
ePBS Breakout #14, held on 20 December 2024. Enshrined proposer-builder separation, or ePBS, would build the split between block proposers and builders into the protocol itself. These breakout calls cover the design, specs and client progress.
Topics
From ePBS Breakout, hosted by Ethereum Foundation.
Transcript
this meeting is going to be short welcome to the breakout room I think this is number 14 uh there's nothing in the agenda just uh there hasn't been any progress on the on the EIP because well essentially because we're waiting just to have at least two teams to start like uh contrasting what the mistakes we have in prison with any other team uh we I promised that I was going to add on the consensus spec the changes that franches brought into for choice I haven't done this yet I got sidetracked by something else but uh but I I promise I'll do this soon uh so that's on my agenda I need to let just uh add this simplification on
on for choice to the to the C spec um besides that uh we've had already agreed on what would be the minimum uh changes for interrup I think PRM already has this uh except this this problem that um thatan actually mentioned I think we still have that bug if there's any missing block with with throws uh I think so where where the block comes in and the payload is missing and it had to include some withdrawals I think prison would just fail to produce any blocks I can fix this quickly as soon as uh we get other clients to to start testing all right uh anyone else I see that
Stefan is here perhaps you can give an update on te yeah so for Te uh yeah we've been a bit busy with the release and de five rep so haven't done that much progress um I did a little bit of progress basically rebased on top of our devet 5 uh and uh yeah I plan to work a little bit during the holidays so yeah hopefully after the New Year's uh yeah I have a better update bigger sounds nice anyone else I see calb is here is working or Kyra is here I don't know if
you guys were working on Inus uh so I was like on a vacation until now I'm going to start like from next week uh hopefully we'll have some progress like by the first week of January all right so that might be it uh so do you guys think uh St who's closer to to finishing this do you guys think that we can have uh sort of like an interrupt by say like the third or fourth week of January is it feasible yeah that sounds feasible yeah GNA have a meeting before that or
uh yeah so uh I I mean it's we are now by the end of December so I think the next one is going to be like um supposed to be like the first week of January so by then we might already have an idea whether or not we will have uh an intro by the end of January um so I think it's probably good to schedule the first meeting by uh Friday the 7 uh well I don't know if it's the 7 but that that first Friday of January I'll I'll I'll open the the breakout 15 for that day so that we arrange if we're going to have it but let's try to Target them by the end of January so two meetings for now uh not a def net but at least an
inup like we can locally run with more than one client and and local local block production this is the obvious question but um we are gonna rebase this on top of the death net five spec right just def5 has like bunch of breaking stuff so and that's okay we can do it but that's fine but uh uh are we going to have like def five on top of the velop yeah defit five I think for prism it's almost on top of develop we're missing one PR so once that PR is merged so that's the one that I'm supposed to finish reviewing no no no there is another one on top of that but I think for interrupt it actually doesn't really
matter but you would be nice to do that because there's no like block structure testation structure changes you in a way actually I suspect that since all blocks are going to be empty anyways and there's not going to be anything so I I suspect uh well right we might have partial withdrawals which is a problem right the other difference is also single attestation so def5 the attestation structure is different they basically propose single attestation so if you want to interrupt with other client if you gossip attestation other client will not understand your attestation unless it's def five yeah yeah okay cool yeah we I mean yeah we probably should do that that five but we
probably just going to be a little bit more work there on top of re basically we we we need a rebase but Stefan said that te is already working on top of a definite five is it yeah yeah I just re this week this is one of the tasks I did so we have a def five ready client so I already re the perfect yeah ahead let's try to Target than that that's going to be the final elector anyways uh what else do we need anything else no are there any questions all right so let's continue working then we can call it a day and out of today we get that we are
targeting the end of January for for interrupt holidays thank you oh just just one question though um I don't know if if people are so we are uh starting I don't actually remember what what was our finalization we we did not start on um on epvs I think we cannot actually have Genesis on epvs on prism so and that's probably a hard problem to solve for us so if if we could have also the fork on the inup that would be good otherwise we're not going to be able to join is that right Terence right I think
that's yeah yeah but we can I think we're doing Genesis in Capella if I'm not wrong or we can do Genesis in the na as well or Electra but we need one more Fork we need we need one fork to get epbs at the minimum but I suspect for every client it's the same thing we don't really have a spec on what is the Genesis epbs State we don't we don't really have the spec for that so I think every client is the same okay good so then then then that's then that we're good shape we can probably Target like a Genesis in Electra or Genesis in the neet but but yeah so the the the finalizing single client intro we did was with um starting I think it was starting on the net yeah because I think
we twice right right right but I think that is the issue of the strip basically and then we can just follow what uh eanda up is doing I think eanda up can start from um uh Electro or then then we're probably fine okay uh and Kev your question uh there there no changes on the L at all uh because in this in this version we are like having payments on the Bon chain there are like much more complicated versions of epbs where the El can have changes but I'm trying to avoid those okay okay guys let's continue working
let's C it a day it was a 10-minute meeting uh but it was good enough we get at least some hope to have in Dr by the end of January have a happy holidays everyone thank you bye have ay bye