All Core Devs - Execution #59

Transcript

[Music] [Music] [Music] [Music]

[Music] [Music] [Music] [Music] [Music]

[Applause] [Music] [Applause] [Music] [Music]

[Music] [Music] [Music]

[Music] [Music] [Music]

[Music] [Music] [Music]

[Music] [Music] [Music] [Music] [Music]

[Applause] [Music] [Applause] [Music] [Music]

[Music] [Music] [Music]

[Music] [Music] [Music]

[Music] [Music] [Music]

[Music] [Music] [Music] [Music] [Music]

[Applause] [Music] [Applause] [Music] [Music]

[Music] [Music] [Music]

[Music] [Music] there was a delay before it kicks in

dirt is cool and we are live so welcome everyone to court of meeting 59 I'm Tim not Hudson I'll be facilitating because Hudson's in Australia so just be going through the agenda shirt and the zoom chat for anyone who wants to see and the first item there was just reviewing the previous decisions made in action items so I'll go through them if anyone has any comments feel free to please jump in but I had to look before and a lot of the people associated with

them are not on the skull so I think it will be pretty quick so from the last call there were a couple of decisions made first one which is fifty eight point one is to stop the prog pal carbon vote the blocked

Tim Beiko: number for that is seven million five hundred and four thousand and if people want to vote on the carbon votes they need to leave their ether in whatever address they used to vote until that block and then there'll be a snapshot taken and then people can which rather either ether so does anyone have anything to add there

yeah I think it was a tenth of April only with that so

ah yeah are we actually passed block 7 million 5 that's there on the website like you need to do it by yeah ok so we are actually passed that block and so the vote should be closed now yeah um thanks thanks Pooja um okay next one was just around the updates of 58 8.2 58.3 that we're not gonna go over every single client or research group and if people have updates they can provide them on the agenda and we'll make sure to leave space for people to ask questions about

those um then moving to like the actions from last time the first one fifty eight point one was around the cat herder is having you look at E I p12 sort of updated to reflect the process

Tim Beiko: I know there's been some stuff happening I've personally spent a bit of time there trying to describe the process and and and sort of gather all the various conversations that have happened around this I don't know if anyone else has anything to add

okay next one was just about reviewing the road maps in ethereal magicians ID to decide whether or not we want to go

with smaller or larger hard Forks I know this is going to be discussed in Berlin next week for the eighth one point X and Istanbul meetings but if anyone wants to sort of comment on that it was just read I think they're gonna be using that as input for those discussions so quick question on that basically I've heard Alexi in particular who's who you know I think he said like let's do three months hard Forks on this call our people open to hard Forks as short as

three months I like object on that I think it's too too quick right for turn around the problem with the three-month hard Forks is that if essentially as a client developer difficult your only job is to implement the artworks and do them than three much is fine but usually clients require a lot of maintenance so if we start doing three month hard forest and we essentially take all the time away from general maintenance and performance improvements objects I mean I think it will

Martin Holst Swende: be a bad thing if we say very hard work every three months but if there are some simple thing we want to

roll out which is implemented quickly and like reprise or something or feature which everyone feels this is finished this we can roll this out something have to be longer than three months in those particular cases

well I'm a team I mention that currently we've been working on the next release next time at our next major release for four months now which means that if we wanted to do hard for him between ten we would need to maintain two versions unreleased version and already released version isn't the idea with the faster hard work schedule that you do much less per version

Martin Holst Swende: yes

well yeah but again I'm saying that the complexity comes with current maintenance so for the past four months we were working on gas and nothing of those four months it's hard for creative stuff but still we can't release because we have so much things to finish right but the idea would not be to schedule a hard work every three months but as see our is feature X finished the exists test cases and it's implemented in all the clients if so then we can heart for pretty soon it like it's like let me ask this to anybody who's kind of better

informed because I mean this is we didn't really kind of frame exactly what Alexi said other than hard Forks every three months is the idea that like a window opens every three months for the possibility of hard for so it was me that was sort of putting words in Alexis mouth in the sense that the context was and just thinking about again he's got this multi hard for plan and he's gonna be anxious if certain dependent features don't take it in if we have more like nine months sized ones so wouldn't it be cool if they were shorter I wanted to just jump in Joseph's story for stepping on you there a bit because there has

been a suggestion to basically have exactly what you're saying hard fork windows more like Greenland times that we know that don't conflict with Christmas or or whatever and I think what other people are saying including Martin is sort of like let's ship stuff when it's ready so I'm not saying things need to be every three months but I wanted to check the temperature of any one super freaked out about more frequent and it sounds like three is though like lower end of that that people are like wow that's really tight and then six people seem to like yeah that's no problem can't we go shorter so

my including the the the areas of the year that are Greenland I have to look up there's a either an EEP filed or something a discussion around this we're somewhere between three and six months shipping smaller features on a regular basis and coordinating among client death so that they tune their process to support that is okay as long as it doesn't impact client maintenance is that a fair enough phrasing for now I think we're not tied to I mean large hard folks every three months so yes more like opportunity windows and things are finished to me

there's a couple of things that we probably like we still haven't proven that we can do a six month march forth like that's that's one step but there's a couple of things that we probably need to automate to be able to do that really well like a lot of the time that goes into hard work is not just committing the code it's everything that goes around it so the more we can automate of that the better

Greg Colvin: I don't think most teams have the right you know person to handle a lot of this stuff so developers wind up setting up cast nets running test cases doing testing you know jobs that

that there's actually people who are good at and enjoy doing you know and I don't know what it takes to just get those people

on two teams yeah so that rolling hard Forks doesn't prevent development from happening well I can speak for us it's money we don't know money to hire them yeah

Greg Colvin: I think there's a few rock and roll songs about that

okay I would mention sorry I will mention that one source of

inspiration would be distros like fedora so basically what happens they have they have this very fixed schedule of when better packages must be in when you know when there's a QA period and so on when testing results must be in and and and that determines what goes into each release and practically what happens is that other community projects started lining to that so you get this cascade of related projects like you know ignore or whatever it might be KD that coordinate their releases just by the mere fact that a schedule exists and it

doesn't have to be perfect and and they sometimes over should it by by you know a couple of weeks because some something turns up but it's easy and predictable to have it there so that would be something to look at like what kind of experience they I've had with with these kind of release schedules and I think it's useful as as I think Joseph said it that this is a window where were you or we can do a hard fork and then there's nothing nothing is there no more since they don't yeah exactly so

Boris Mann: I think definitely something that we're gonna discuss next

week and to Fred's point I think everybody sort of agreed that we essentially aim for Istanbul and use Istanbul as kind of a test case to go faster although in this case it technically is nine months what I would also say is that in many ways we should start planning the next hard fork sooner rather than later to just train ourselves to get the cadence of shipping more often and then ideally Istanbul as getting onto this cadence will let us shake out any of the things that we need to scale and automate and collaborate on

and I would say maybe just a Gregg's point into Fred's point you know which of those things can we do across client because I think that's one of the really big challenges especially as we add more clients if everyone is individually building test harnesses and everything else like that this is just the obvious point where you know how do we work together

you

Tim Beiko: so maybe for next step here given there's gonna be discussion about this next week in Berlin if we can just try and make sure that the output of those

discussions get put back into the 8th Magicians shreds that was that that that was linked in the agenda

I said just people this way people who aren't that the meetings can can get that I don't know if anyone else has anything to add on subject of the hard fork length okay moving on so I would took me a while to find me but I would just want to to report I think at the Berlin Mena way will present an idea now we could do safer automated art forks with like n version programming it would be a massive engineering effort upfront to

actually tolerate that so no realistically I don't know if we can make it happen but it'll be interesting and we can talk more about it after that is that he is presented that cool anyone else have anything on the subject okay next action item was for Vitalik the format 8 1559 because I don't think it has been properly formatted under like the template I just had another look and it's still not the case so I guess we can just leave that one there I'm not sure if anyone has comment on

this or okay next action item was the shutdown profound carbon votes that's been done after that basically

Tim Beiko: the action item was to continue the discussions on prog POW and whether or not we want to go forward if the technical audit is not funded and I think the goal was to try and bring those discussions outside of Cortez call into the links that were put in the in the notes so unless anyone has anything sort of new noteworthy or urgent I think it probably makes sense to keep the discussion there

agreed and and there will be talk about

prom palace law at the the Berlin meetings next week so I just post and I just posted all the sort of each magicians Lynch regarding Prague now and in the zoom chat okay next two action items are similar basically just not providing updates during the call with providing them in the comments and letting people ask questions about them for both clients and research and then there was a final action item from Alexa around creating a beacon chain finality gadget working group like size not on the call but I don't know if anyone else has comments on that okay so moving on to the second agenda

item which was the roadmap so I feel like we kind of touched on this already but I don't know if anyone else has anything they want to add just about roadmaps in general or Istanbul specifically you yeah I just wanted to talk about it a little bit like on for this road map for Istanbul to be agree to go by the timeline like October 16 is our target date Hall is there a possibility that we will not be going by that so yeah one thing that has been

discussed which I've been a proponent for is to do if we do profile do it as a separate standalone fork okay in that is like oh we have like eight a IDs proposed so far so if we want to start to take proper separately we still do have 70 IPS proposed so if we consider those seven together we are are we hoping to be this four by October so can we step back why would prompt ah be pulled out what because it's done

earlier done later well initially the proposal was because it was to be done earlier and then if it's also considered a controversial Eve that would be an additional reason

Pooja Ranjan: well I'm not saying that it is being pulled out of that that's not my say a stiffened non holders I just have shared the list of the VIPs that has been proposed so far so if we can discuss that how many have been like well considered

puja

Brooklyn Zelenka: this is not the right time for this so they're just

proposed so the deadline is in a roughly a month from now to see which of them are actually more than just proposed but right now the schedule in previous court have calls we've agreed upon and so people have a month to purport prose more epi Peas get them ready if they want to get them ready and as far as I can tell from previous court have calls yes we are locked on the schedule as it had been previously agreed

thank you yeah I guess oh go ahead oh sorry yeah

I'm just gonna ask

Martin Holst Swende: about these proposals I assume that now is not the time to discuss them is that correct

this specific eat yeah the specific Gibbs our feelings about them and so that was the

Tim Beiko: next agenda item actually yeah

Danno Ferrin: before we go there just a plug good test cases belong in each so we can reduce the burden on the developing the test cases

later even now we can keep this as a reference

and we will keep on collecting more VIPs and we'll share that you guys as we go and we'll discuss it after a month or so or maybe burning meeting my little buzz more inside and yeah so I guess talking about oops

Tim Beiko: there were two posted in the comments of this call but I know there's also a whole lot more that that puja just mentioned and so perhaps I see we have bryant and pooja who proposed the two eeap's in the comments on the call so maybe it makes sense for them to both

talk about their eat and then after that for anyone to just comment on any of the other proposed eats for Istanbul um so Bryant I think you actually were first in the in the comments so if you want to go ahead and talk about eat 13:44

sure thing so I actually wasn't the author of this one but about like a week and a half ago I was playing around with some seven one two stuff and I was like why isn't there a chain ID opcode and so I went to go look and there was a proposal for one so I've been helping kind of format it in a nicer way and get it accepted and kind of gather feedback

it's pretty simple for the most part there have been a little bit of a discussion around like what kind of data type do we represent this as I think in the I think in the opcode it makes sense just to use the full word size and do a 256-bit word otherwise there's not really too much to it but I wanted to gather other feedback and see what else we would need to do to get this included first of all are you having a question how much chicken you can guys hear me so I was up

Alex Beregszaszi: for the last two weeks haven't really kept a good eye on everything but at ECC we discussed the option and that perhaps the heart Fork meta a IP should have a proposed list and I think Ryan has proposed to see IP for the Hartford meta is there any consensus yet at what point where do we go for that whether we merge that pr2 to include this as

the proposed VIP or was in the discussion around them that was actually the one that I did Alex based on our discussions so Brian

has some stuff around the audience from you but I think this is the one you mean is the pull request that updates to 3 3 I don't think there's been any update so in brief I'll just I'll quickly look up the PR so I can find up the number but essentially it's using the hard fork meta where people do pull requests to the hard for men are representing the next hard fork as a way that we can actively keep track and move through proposed approved and so on in one place so you can very easily see which yet keys are in flight to potentially make it into a hard fork nothing else changes

in the sense that proposers will still need to come on chord AB calls and say hey how do people feel about eBay X and putting it in the assemble and go through that process the outcome of it just would be actively tracked in that so I think the action item is a follow up from this is just to look at the PR and I'll just quickly look that up now I think that's kind of what you meant right Alex no actually but as a quick response to the I think last time whoever commented on your PR we're happy with that except the distinct mentioning a time timeline and I think we agreed if you remove that

single change then we should merge that PR but we can discuss this offline Francisco what I actually meant by the question there was this PR opened yesterday by Bryant to include VIP one three four four this chain I de IP in the Hartford meta and that has been closed but I'm wondering if it should be merged or it shouldn't or what's the process now it is pre quest-19 three four on the eyepiece repo pull request one nine three five looks like he's reopening it so that was an issue okay

yeah so I was trying to follow I guess Hoosiers process which she also wants to talk about in this call so if just following her process that's what the suggestion was and then I talked about it in 193 for like maybe this doesn't make sense maybe open an issue instead so it's just figuring it out that best okay so this is exactly the same thing is you're wanting to make a pull request on the meta as we change each two three three meta process yeah I think it's just all of us trying to do the same thing at the same time I mean this is a change to the Istanbul matter not the dot two three three following the 233 updated process to

make it into the estimable meta yeah let's think we just went in a metal loop there but yes

Alex Beregszaszi: sorry I may have sidetracked a conversation we can we can discuss this offline

because

Martin Holst Swende: I have a question about 1344 okay the chain ID so essentially instead of storing the chain ID that you expect in your smart contract you just use opcode is that the only thing

that it was yeah it's basically a hidden and so it has a few benefits so having it as an opcode means that if the hard work does

occur then code on one chain won't get replayed on another chain cuz I don't know I mean it would be the same thing I will never change the chain it is that way well not like a natural contentious heart fort where there's two communities

Martin Holst Swende: right

but so like they would change their chain ID and then the OP code would output a different number do we really need an up code or having a a contract at a well known address for

each chain would that be a sufficient alternative I would think an OP code would be most optimal because though the way I came about thinking about this is for like seven one two signing procedures which are like very content very context-specific towards like what application you're doing offline signing for so this this would help in that like you would be able to do it on a chain and then like you know if split happens you would make sure that there's no replay attacks against the offline signed messages that you seven one two so

Danno Ferrin: that was a quest

also I brought up in the youth magicians thread and the one of the concerns was that the contract would be more expensive than an opcode by necessity of the call and querying the chain ideas an incredibly cheap operation

yeah we actually specified it as the base price so to guess I did a small implementation in Trinity it's about eleven lines of code and one way that we could go about doing it by taking the chain ID from the transaction context and leveraging that I wasn't sure if there was another way of doing it so not as familiar with how

the clients were built out but that seemed like a an obvious way of doing it you yeah I mean it's kind of make sense to have the opcode I don't know yeah because good like currently the the basically the suggestion is is to kind of manually manage it when you deploy your contract and if you're deploying so like let's say you're doing stuff on Propst in in order to test it or Corley

or whatever and then those if you forget to update the parameter on the maintenance act now of send the other replay to action so that could be an issue yeah so it's a bit of human error to to it yeah I kind of agree that usually the more introspection or self-reflection you can do the fancier stuff you can make so I think any functionality that adds that lacks the contracts in respect their own changes is a beneficial so I have a question about another one

of these if it's okay to move on and that would basically be EEP 6:15 a my concern about that is that it's humongous and will make for really really large changes into in the consensus engine and to me it's really unclear what the benefits are because I don't think it's EVM speed that will be the benefits so if anyone would like to champion that I'd really like to hear it I'm sure Greg would love to jump in on this too but so speed we can absolutely do this is a first staging for improving Jicks or ahead of time compilation allows us

better language support because mapping better to for example LLVM and doing better static analysis for security so it's the first step for a whole bunch of things you know it's yeah it's a number of opcodes currently we could potentially break it up but if they really do work together as a cohesive whole this has been discussed for a bit on eath Magicians and strictly on the github thread I'm not sure if much more on this call is needed or if we just need to you know hash it out in in those

places but I'm sure I did take part of those discussions and I wrote the same thing there that for the first hard fork this would be not positive speed effects on the EVM because we need validation in runtime and it would be much more expensive validation than the jumped assess and analysis done today not done after until we have a second hard fork and we can do deployment time checks then we can skip the validation check in run time but even at that point I still think that improving the actual how we manage our

we'll only have a marginal effect on the VM speed because the main bottlenecks for a DBM is or the disk IO and that so yeah that's my main concern about C it's a lot of a lot of work but I don't think it'll lead to much better EVM it might be better for the external tools like if you do not bursary analysis all the security properties of the contract yes if the probe help you yeah

Greg Colvin: I don't see any immediate performance improvement out of this one the validation is is very fast in the C++ client it happens

at deploy time and so it's it's really not an issue it is it is a better EVM from the point of view of external tools yeah that's that's where we really see people saying yes please the K analysis was very much harder because it had to deal with pattern matching looking for what a solidity subroutine call looks like rather than just an ordinary sub instruction

like every other chip in the world has so

Boris Mann: I think the point is that this whole thing is in part forward

compatibility to get us tools for a variety of people building on top of the EVM in building more features that link us up more directly to what our future execution engine is in wasum so in some way as part of this has to be how much of this do we do in the EVM sooner rather than later especially if we do hard Forks more often versus just waiting until some future time and then this shift for anyone writing contracts today will be much larger because of a deficit in features of

our Assam compatibility was another design

constraint on you of how hard it is to implement it with several hundred lines of C++ and it took me about two months to implement half kind now there's we haven't been through last call on it and as Martin knows there's still there's another AIP for version control that needs to go in so that it's

possible for VMs to know which which code am i running old-style code wasm code news file EDM code you know we need to get that in place anyway yeah

Martin Holst Swende: with that in mind is this heap actually still a proposal

or byzantium first of all

Greg Colvin: yeah it's been it's been sitting here for two years now

right but I mean there needs to be a virtually and it requires the versioning as well yes yeah so so

Greg Colvin: we've been working on that in parallel

and you'll just have to work it in parallel so they're both ready

okay

Martin Holst Swende: yeah I guess today is the time to make any kind of decision on these things just discuss it

Greg Colvin: yeah I guess may 17 stack our deadline but we've been working together on that one so we I think we have a straightforward notion of what that needs to look like

okay yeah I have nothing further about that on today at this monitor yes sir Pass

Greg Colvin: has been making changes in parallel to the discussion so his his P ours is is ready to go

Fredrik Harrysson: something that I

just wanted to bring up and hear people's thoughts on when it comes to this stuff is at this time for thing Alexei had a pretty long speech about not making changes that are in for the true survival of the chain not make improvements but only I'll fix this or whatever you want to call it but I like I said the same thing many times like we shouldn't be changing things anymore we should just try to keep the chain alive while we work on e2 and the new stuff goes in there how much do we sway outside of this because this yeah if he leads into like a large chain of improvements that go into like several

years before we're seeing massive benefits from

Greg Colvin: 2021 is two years from now startups like die if they don't get new technology out every year

Brooklyn Zelenka: yeah I I have to echo that one it's you know each to has a sort of undetermined release States it's making changes like this improve what we have today make its forward compatible so that it works with stuff later so that's smart contract developers have a smooth experience keeping things in essentially

maintenance mode you know there are other options for smart contract developments out there we need to keep cerium one dot X alive and that means continuing to move

Greg Colvin: remember and D to team knows it that the main chains is going to keep on operating indefinitely if there's no value on the main chain there's nothing to stake

that's not entirely true like the the intention is to move the value from the main chain to like you would stake it on the beacon chain but it doesn't matter

like I'm still curious yeah the piece what the sentiment here was around that you guys proposing the EPI it's pretty obvious that you are included for the future we don't know when my p2 is coming there's no firm date yeah so with regard to each one versus if - I think I mean if - is this silver bullet that's supposed to solve everything except whenever it arrives and I think as maintainer stuff if if one it's really really important to keep it alive until these two actually arrives and whether

if one will die a horrible death whether it will be devalued or something that's kind of up for the future so I I wouldn't I don't think it's a correct approach to already start burying eath one while we have absolutely no idea what the future contains or holes yeah one forward as much as we can and then if something better comes along then it's fine we know that that is a possibility but let's not kill it before that happens I don't think that was Fred alluded to and I can remove being skeptical about implementing such a large change into

the old engine which basically takes a couple of hard Forks before finally settles and even if we don't have to that though we might have also passed from somewhere lurking around the corner so I don't yeah I'm skeptical about it I wasn't really referring that okay now we should upgrade aetherium into zero 3.0 all of a sudden so I'm fine with discussing all the changes and betting what is what should be done what shouldn't be done I was rather aiming that we shouldn't hold back possible upgrade there isn't just our thinking a

Fredrik Harrysson: roadmap based on what these two may or may not be

yeah and

Boris Mann: I think a lot of this is exactly the purpose of why trying to have an in-person meeting that talks about some of these things I'm glad that this is coming up today already like six one five aside I think EP 1x and especially from what I heard at ECC there seems to be pantip demand for all sorts of things again slow and considered in the sense of making sure that there's a stable healthy Network but part of that is having features that other people building on top of it can

make advantage of I would point to so it looks like there's gonna be four new pre compiles that people want to propose for Istanbul and each new pre-compiled unlocks more cryptographic primitives that gives our application layer people a ton of stuff to build on or helps them improve or reduce the gas cost of various existing ones and that seems to be what people are still excited about seeing because we have a system that is live and people can ship on today

you

you [Music]

Tim Beiko: um unless anyone else has anything to add on that Danno had a question in the chat I think it's just worth getting parody on so for example when do we want each to go through last call that they have to have gone through that call by May 17th or do they need to move into last call by May 17th

Greg Colvin: I intend to be through last call by May 17th I don't know what the requirement

is

in the chat I think it's been said that the requirement is unclear so uh yeah I don't know if anyone else has an opinion on unlike where the each what what should the sort of last call point B it should definitely be before the for it hard Forks I would think you wouldn't need to be in last call if you're in May 17th especially if there's discussion it goes back to major changes review can you get on the next hard fork train but when it comes to starting to find about open I think that's when the last call would need to have started for sure because it's like what a two-week window or you could at last call and people are invited to so I can RSS feeds people

isn't super that's the only time you look eat some people so that's my thinking on it

Boris Mann: for reference the last hard Fork everything was still in draft and hadn't been moved through anything so there was a tweak to the Eve's process that theoretically means sticking to these dates and so on so that's why some of the uncertainty exists because we're trying to improve the process for the first time I think people need to actually have them proposed ideally between so the next five weeks we have people actively doing PRS against the

hard fork meta saying that they intend to pursue getting the sins they assemble and then they have to pass the other checkpoints like major client implementations and security and so on to actually get it in Joseph's EEP one got updated recently and this is literally the new process and exactly how something should be proposed is not described in there but I think that's part of saying that proposed means you've done a PR against the hard fork metaphor the upcoming fork

do you think a security review should be

required before that point I thought that was up you know maybe leading into the actual you know forking off date there's a lot of time between now huh I think the plan is that we'll do security reviews on the ones that are accepted for hard work that makes sense to me I did we may also get added to the yellow paper or whichever paper we want to use so security of you would be parallel to client implementation does that what I'm hearing

Martin Holst Swende: yeah

yeah maybe I'd miss her what Boris was talking about

I heard that all this had to be done before the proposal date no but there is the thing there there's a separate proposal to do that if that each EEP should have a section called something like security consideration for the EEP author takes well outlines some security implications from its own thinking but also from the discussions that have happened around the EEP and compile it down in today which is so that is separate from actual having someone externally doing a security audit of the yeah so

Tim Beiko: in that case does it make sense to say that on May 17th what we expect

is emerged PR that's compliant with the two to three PR so basically merging your specific deep and the hard fork meta EEP and that includes some security considerations in the EEP but not a proper security audit nor does it include sort of nor does it imply that the EEP is being accepted but just then it's been merged into the meta hard for e does that make sense

I think the security reviews thing was considered separately as a modification of that okay the IP but yeah I think that the real question is you know last caller

accepted before they 70 no

Boris Mann: has to be in a draft PR I'm sorry a draft EEP and in the hard fork meta is I believe the line in the sand because all you are in is proposed at that stage there's no guarantees that you're making it in

you being the person saying that they're gonna work on attempting to get it in our signaling that you're gonna work together so Draft by May 17th and proposed and by proposed we mean emerged PR into the meta EEP yes yes

okay so just does anyone disagree with with this should've just post the steps somewhere because I'm not sure yeah I think yeah let's let's definitely have that in the summary but from what I understand it is having a draft EEP that is included in the mellah EEP like the Istanbul Mela apart fort and Boris's is PR that that was posted in the chat a bit earlier I describes how how to do that yeah I have another question I don't know if

I'm code blue minutes on but the there's a proposal to do pre-compile for elliptic curve linear combinations and from what I can tell I can't even read what the gas cost calculation is yeah maybe I should just bring it up on the forum it's done yeah I don't think go is on

Tim Beiko: um other question in the chat from Joseph when would the three client implementations we do for Istanbul so I assume basically the the client

implementation deadline right now is is the 19th of July but I'm not sure is that what we want to use

Boris Mann: unless somebody fights against the schedule as it is the Istanbul harm for planning dates are set unless core

tabs are going against things that they currently accepted okay so let's keep the July 19 and then there was puja you also had an e pin the comments that she wanted to talk about each 1929 I think yeah I guess Brooklyn just mentioned

that it is you know kind of duplicate to the process that has been like requested for two 33.8 processes there it's the P R 1852 so I was just trying to add on things there and like we had been discussing this and it Aiden gathered a forum about how to streamline the process so it is not nothing as specific to this Istanbul only the if that I happen I have requested is for stating the process that we need we may follow to streamline the things for any

upcoming upgrades and the one thing I have received in comments from gas make Johnson in the etherium magician forum that in my proposed proposed process I have requested to submit everybody to P R of P R at the EIP github which may create some kind of confusion that the EIP editors are the one who are taking decisions on which one should go or not go so it is it may be better that if we can take it somewhere as other than the EIP github I mean this is what he has

proposed so I I further propose to change it to issue section that's what I guess brand dead for his proposed EIP but if sending a IP do the IP github may be a problem then we made probably take it to entitled also github and receive and collect his EA B's over there I mean that is my suggestion and if anybody would want to sand that merging at EIP does not

Greg Colvin: indicate that the editors approve of it it only indicates

that it meets minimum standards for form so it's perfectly fine to merge them there and point the discussion over to magicians or anywhere else you want to discuss it

Pooja Ranjan: that is fine being we are referring magician thread where the discussion about that particularly IP is going on but what we need to say here is when we want to indicate that this is our AIB and I would want you to be included in mix heartful where we should submit it my suggestion would be in issue separation so that when it is picked up from there issue can be closed

but when it is in PR that needs merging and when it refers to any meta ID this automatically goes to the author of many IP because every time the person would be mentioning I want to propose this Eid for this particular matter at the end it would be going

that

Greg Colvin: sounds like something to discuss on the EIP channel I don't know

maybe in a Pia so if

Boris Mann: the very specific thing here is is I

think Nick's concern was that eeap's are meant to be standards and if we use hard fork meta it mixes standards with aetherium main net chain but I think that's fuzzy anyway so 1852 I'm about to submit an update to it that should satisfy what Alex said and then use that one to further this discussion

Greg Colvin: okay this is an old discussion I'm glad if we ever see it resolved

I had another point in her proposal I

think kind of overall so in discussions too I think right now does it say that you have to do it through eath magicians or is it like this is a choice you can make no all right and then the second part that I brought up was socializing the IP it was suggested to like bring it up through all the other places people gather and I made the suggestion to change it - these are places people gather you can socialize it through those means or whatever means you seem necessary but the official discussions -

forum is where discussion has to happen in order for it to be kind of considered yes that

Pooja Ranjan: has been updated accordingly yes your suggestion was like really relevant if we would want to discuss it in some other foreign forum to make more people aware of it it's fine but if Dida magician is a suggested forum because we are going to get every discussion listed at one place that is definitely a good start for

any Yeti oh I more mean whatever is listed in discussions - that's where conversation has to be captured to be considered

Pooja Ranjan: yes

and like somebody might propose an economics VIP where like maybe eath

magicians for them doesn't make sense as a place to discuss it that's up to the author to propose an alternative forum but there has to be exactly one listed and that forum is the only place where discussion can kind of be considered official

Pooja Ranjan: um yeah I mean that that mean that we can write it up too but I'm open for suggestions like if everybody agrees that if magician should only be the place where it should be discussed it can be changed accordingly

I think I think it makes sense that

people have said that maybe eath magicians isn't like a catch-all environment for discussions some people view it as too technical so like one half of my suggestion is that a logician tongue Brian

Boris Mann: let's just figure out who cares about this discussion have the discussion and come back with a proposal in the form of an

uneat sure but I guess the last thing I wanted to say is does it make sense to say that discussions two is the only place for a IP conversation be captured

Boris Mann: No

you you okay so I [Music] have anything about any of the other eeap's for Istanbul so Marlin I know you commented on if you already but just want to make sure that we leave the space if there's other questions or comments about the other proposed ones you okay um so next point the agenda is the working group updates I'm not sure how

many of them are actually on this call state fees I know Alexa is not here and he had an update directly on the github agenda so unless anyone has anything to add then he was him is there anyone with an update on he was him on the call I prune flashing so it's 64 V for each V 64 call for proposal hey I'm here just a quick update basically just moving the discussion from eighth magicians into I just pretty much minimum viable lightweight eeap's so that there's something that are

discussed in Berlin and then from there we'll just see what we actually want to adopt and again these are mostly gonna be pretty lightweight changes and would probably land at the same time a sinking proposal is made you cool um next working group is simulation and the last not sure why this guy added as a working group but there was a line for the Istanbul / eat one point X meetup in Berlin next week

Tim Beiko: we've already touched on it a couple times but is

there anything else

anyone wants to share on that okay our next point was the prize well audit update Hudson typically des does so does anyone else have anything to add on the audit okay and then last part is client updates and research updates so like we said there's some there's some of those posted on github so if anyone has either a question or something that they didn't manage to post on github then I feel free to share it I just wanted to can you hear me yes

hello yeah I just wanted to add the parity aetherium just recently added click consensus so full support for girly chestnut isn't that's another two point five beta would nice anywhere else an update or a question on one of the updates ok is there anything else in general anyone wants to bring up okay then I guess that's it I thanks everybody for your time and we'll see

you in the next one of these am thanks for rescuing us today no thanks what about alright you [Music] [Music]

[Music] [Music] [Music] you

[Music] [Applause] [Music] [Applause] [Music] [Music]

[Music] [Music] [Music]

[Music] [Music] [Music]

[Music]

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