All Core Devs - Consensus #83

Transcript

[Music] [Music] so [Music]

[Music] [Music] [Applause]

[Music] so

[Music] so [Music] [Music]

[Music] [Music] [Music] so

[Music]

Danny Ryan: hello welcome to slayer call number 83.

i'm fixing something real quick and then we'll get started okay it is fixed this is issue 489 on the pm repo sorry i'm dealing with one tiny technical issue sorry about that we are now good to go

issue 489 shared in the chat um we will spend as much time as necessary on kiln on the merge and then go into our normal discussions uh thank you for joining let's see i don't have anything under the kiln agenda but there's plenty to talk about the i think most pertinent thing is the launch of the kiln the public the not devnet the test nut kiln um i think perry and or mikhail probably have the most up-to-date information would you mind sharing it with us

hey

Parithosh Jayanthi: everyone um

yeah so we have the kin configs that went up yesterday um the test net itself is about a hundred and four hundred and five thousand validators and a couple of client teams um are participating so they've already been sent to mnemonics and would be validating with them if anyone else wants to join and doesn't yet have validators and we can just make deposits for you guys uh all the tooling links have been posted and we are slowly working on a guide for each and every client um if you guys do make releases with the conflict built in then please let us know so we can update the debits

uh ttd is said to be i think like 100 million or something like that so based on last week's run that should last us up till mid of next week um but it's almost impossible to know if someone from the public runs about it let's run the miner and choose through the dtd but we're hoping that it will last for

at least a week

Danny Ryan: awesome so is the proof of work and the proof-of-stake genesis had they already occurred at this point ah yeah good point to mention uh proof-of-work genesis has occurred proof-of-stake genesis has not happened yet so the beacon chain hasn't been launched yet the beacon chain would go live tomorrow around this time um

yeah that's that's the main difference and the merge fork epoch is i think uh epoch 150 so about half a day or a day

later

Danny Ryan: cool and i believe the eath staker folks are trying to do a party call but now they have to feel the pain of work and scheduling so they'll probably release some details soon and i do believe on the ef blog we'll have a blog post go out either today or tomorrow about details here so keep your eyes peeled if you are listening and you want to participate and learn more okay uh did anybody have any questions

about kiln questions

let's do a

Ben Edgington: quick recap on on why it was um put back by 10 days or so because this is not what we agreed at the last meeting and i'm very mindful of little delays so they killed us on out there there are two things that we're attempting to do

Danny Ryan: one is to have the technicals ready and to vet things and to run stuff for long enough and to understand what's going on the other is to engage with the wider community merge devnet 5 handled the former kept things moving on the former so i didn't there's no delay there whereas the latter was simply not ready for public engagement and so although the technical stream continued to move forward at the pace that was whatever pace developers could do

we couldn't really enter into that secondary stream but it was decided that the secondary stream which engages with users this month wouldn't actually delay the release of test nets in april the only thing that would really do that would be if you push that way back or if the technical stream was the delay so we kind of and wanted to ensure that the public version of this was ready um that's that was the discussion at least had last week

got it okay thanks make sense

Tim Beiko: yeah i mean and and part of part of that we could debate whether that actually

shifts the timelines but that was the uh that was a discussion and as of now we do not plan on delaying the public testings in april yeah and i think one thing i'll add on on what danny just said is uh you know what's really important with kiln is getting as much kind of real world usage um but now it's really important to try and reach out to folks and get them to uh deploy applications test and tooling make sure that it's not only like deploying an actual smart contract on mainnet but it's uh or unkilled but it's you know deploying your contract through some pipeline uh you know with your with your ci tools making sure that your

front end works making sure that things like truffle and hard hat and that stuff works um so that we don't have any kind of bad surprises when we when we fork the actual test

Danny Ryan: cool other kiln or merge related comments or discussions

Mikhail Kalinin: i have a minor update on the 2.1 release so kill me 2.1 it's released now it doesn't have any yeah it's a it's backwards compatible with with two and i'm pretty sure that uh we started

like this persistent test net with v2 uh then clients will be able to upgrade to be 2.1 independently so it has two like changes that requires changes to implementations first one is the quality of service on exchange transition configuration method and another one is the um related to optimistic sync um yeah basically it allows for importing any post merge block optimistically so any blog that is uh the descendant of a transition block can be applied to the choice

optimistically and yeah it's it's considered safe um it's just you know makes it simpler the code of optimistic sync so it doesn't introduce any breaking changes that's like short update from my side

Danny Ryan: thank you and i believe

you're gone i

Marius van der Wijden: i have a small update for for the merch uh the the engine fast stuff so i've written a differential fuzzer to test different implementations of the engine api against each other on the consensus layer and i

updated that to the to the new spec and i'm currently starting to add clients to it right now it's only doing gathering bisous and it immediately finds a finds a difference between gather and bisu and that's probably because of some some reverse sync in in visu but i'm planning to add nethermind and ethereum js uh to it and then as soon as aragon is ready also aragorn um yeah if other folks are interested in in being added to that then they should just contact me

regressions

Danny Ryan: is probably mario's um hive engine api tests i would presume

Marius van der Wijden: yeah exactly and we're also trying to get the other execution layer clients to the to the hive tests so that we can execute the tests that we already have

on them

Danny Ryan: got it on that note any other um testing updates i know that the media and some others are kicking off um getting beacon fuzz up to

bellatrix targets um so i believe by end of the month that should be in the in the posing infrastructure i have like a small discussion point um do we want to fill out the kill milestones tables so i see some clients uh finds updates here but others

are not like filled in

Tim Beiko: i would push for it just because i think there's like thousands of you yeah there's like 7 000 views on that page almost so if people want to quickly update it i

think that would be valuable um or even if if you're too busy to update the actual table but you have a tracking issue or something i don't mind doing it on behalf of clients myself um but just because there's thousands of views on the page i think it'd be good

Danny Ryan: and um i guess just to state it out loud we're running uh marius running the bad transactions or transaction buzzer and

bad blocks on kiln

Marius van der Wijden: so the transaction fuzzer is running intermittently every now and then every time i have like some free cpu cycles i'm running the transaction fuzzer and the bad block generator i updated the bat block generator to the new v2 spec but um i don't think we have a note running that at the moment

Danny Ryan: okay i'll probably be valuable soon

Marius van der Wijden: yeah but that also the bad block

generator only generates bad blocks after the merge so it's not really useful before that

Danny Ryan: gotcha another thing that i think would be valuable for this or other kind of simulation type tests is to have um quite as uh maybe quite as interesting as what's going on in the transaction world in terms of what you can do but probably worth sending some uh some operations midweek good luck estimate we'll plan on doing on the consensus layer side for a release um i guess semi-major version bump or

major version bump i think we did a major version bump uh at altair um or not i need to go back and look anyway we will do a release probably at the end of march when we are making decisions about forking test nuts so whenever we're doing that and kiln has been running for at least a couple weeks we'll get some sort of more frozen official release together um we might see some iterative test vectors come out in the meantime we're doing a pass just to make sure we have the coverage that we need anything else on the merge before we move on

Parithosh Jayanthi: um just one more point it would be creative in trouble didn't recognize that and would would not serve you any content and while we can argue that execution endpoints shouldn't be accessible by dns and in like setups which use kubernetes that is by default what's going to happen so upgrades list and some other stuff where we will solicit people to to do so merge going once going twice so well sorry i actually have something

really quick um yeah uh tj rush uh our resident difficulty bond expert and someone else on each research whose scandal is you'll see you also anyways they put together a dune dashboard uh looking at the amount of blocks that gets produced on my net per week which is basically a good indicator for when the difficulty bomb is going to start to kick in so it looks at like all of the

theorems history

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