Everything EOS EOS Leap Upgrade Group, 11 May 2022

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: EOS MANDEL Meeting #14 (May 11th 2022)

Description

This week the EOS Network Foundation covers topics on platform support, Docker documentation, testing Jungle 4 on MANDEL, and Functionality dependencies. They also go over the Docker repository, long term support on Ubuntu and reports of Jungle forking.

Summary - 🟢 - In progress action items are on track:
All EOS Mandel Meetings : https://github.com/EOS-Nation/mandel/issues

00:00:00 Intro
00:00:05 Recap from last week
00:00:55 Platform support
00:02:00 Docker documentation and testing
00:03:09 Docker repository
00:04:00 Docker packages
00:05:10 Web IDE
00:07:17 Docker images
00:07:52 Pinned compiler for Mandel
00:10:45 Long term support on Ubuntu
00:11:44 Binary packages
00:13:05 Functionality dependencies
00:14:53 Mandel code strategy
00:16:57 Testing Mandel on Jungle 4
00:20:17 Reports of Jungle forking
00:22:28 Outro
00:22:43 End

#55
Docker image testing and documentation scope has expanded to become a full-blown dev environment, which also includes a web IDE. Development is done and testing and documentation is almost complete.
Platform support survey has been distributed by EOS Support and we should have results to share next week.

#41
Scope for 3.1 has been pretty much finalized and all features are expected to be ready for testing by late May or early June.
API+ Transaction Lifecycle features being developed by OCI are expected to be completed in the coming days.

#44
EOS Nation is continuing to test transaction lifecycle features from 3.1 pre-release on Jungle 4 as they become available
EOSphere is now running Mandel 3.05 RC1 on Jungle4 for their public API and seed peer. BP coming soon.
A question from ENF was raised about forks and LIB falling behind that has been reported by developers testing EVM on Jungle4. @matthewdarwin has confirmed that this is not due to any issue with the Mandel software and not unexpected for Jungle4 given the testing happening on unstable releases. Jungle4 is meant to be the testnet for breaking stuff. ENF may want to consider funding a stable testnet for dApp developers, along the lines of the B1 testnet, which is no longer stable.

Participants (17)
Host: Daniel Keyes | EOS Nation
swatanabe
Matt Witherspoon
Max_Cho EdenOnEOS Member-South Korea
Jeff Werner
Yves La Rose
Genie Kim - EOSeoul# NEOPLY
Ross (EOSphere)
Kevin Heifner
Vlad
Ted Cahall
Bucky Kittinger
Lovejoy
Johnsonb - Object Computing# Inc.
cc32d9
junhank | NodeONE
Dario | EOSsupport.io

Timestamps
00:00:00

Everything EOS is the official media channel of the EOS Network Foundation (ENF). The EOS Network Foundation coordinates financial and non-financial support to encourage the growth and development of the EOS Network. We’re harnessing the power of decentralization to chart a coordinated future for the EOS Network as a force for positive global change.

// EOS Network Foundation //
https://eosn.foundation
https://twitter.com/EosNFoundation
https://www.linkedin.com/company/eos-network-foundation
https://t.me/EOSNetworkFoundation
https://discord.gg/eos-network
info@eosn.foundation
#EOS #EOSIO #ENF

A

All right welcome everybody. It's may 11th it's our 14th week meeting a quick recap. From last week uh we we were making progress on the docker image, uh testing and documentation.

A

We've made progress on the survey that got finalized. Last week. I got an update there to share um we're continuing to make progress on prioritizing finalizing. What's in and what's out for mandel and um more progress on testing uh so we'll check in on continued progress on all these items. Is there any other anything else to discuss today to check, in maybe older items or new items, to add.

A

All right, let's start with platform support.

A

We've eosupport teams sent out the surveys this week.

A

I think on monday, or perhaps on this weekend I got an update from randall um looks like yeah, 100 or so emails sent to dapps 15 to decentralize exchanges, 169 to centralized exchanges and they've also been spreading it in across 76 different telegram channels for various dapps and indexes, and centralized exchanges and twitter facebook reddit, so they've been spreading it far and wide and um no results yet to report in terms of responses we'll be checking in on friday to get a account of responses, and so we should have some some insights hopefully to share by our next call.

A

um How are we doing on the docker, documentation and testing?

A

uh Yes, pretty.

B

Much everything's done. uh We got support for windows, 10, 11, um mac, os mac, os m1 um you blend to arch, I mean any kind of linux you kind of want to throw at it runs perfectly fine. um The only thing I've stayed up last night trying to work through was. I wanted to get support for sort of web app stuff that we had in the web ide and it didn't quite get there. It's not it's not that big of a deal. So it's not a it's, not a huge breaker, but the documentation is done.

B

uh I forked it over from the repo that I'd had in my own personal one, so the the documentation- that's there's not quite up to date, so I got to fix that after this caller go lay down for a little bit, but at some point today I'll fix that.

A

Great, it sounds like we're: we're yeah pretty much there.

C

So where does the docker uh stuff live? You.

B

Yeah, so so yeah there's another thing: it's currently I'm just calling the tool and the thing that it is uh doom. We can change the name to whatever we want to the images.

B

So so I have a pr that's going up for eos um and eos vm that effectively gets rid of or doesn't get rid of, but it allows for the m1 builds and then uh what's the other thing there, um the it also has pin, builds and uh package management because I needed it for a package construction because I needed it for this, um so so, whatever matt wants to do with that, um you know he's full reign to go in and tear it out or do whatever he wants to.

B

It's just needed it in there, so those two pr's will be in the packages themselves. So the way the docker image is set up is that basically it's pulling in um packages that are currently in a repo that I'm calling enf binaries just for right now until we do the main release, um so the the actual setting up the machine should be pretty pretty fast.

B

So the repo is enough binaries, uh so the the repo that you would want to go to would be just d-u-n-e.

B

A

And that's from the enf org yeah is that is that ready to like? Should we be including that any of the documentation at this time, or or still some work to do maybe to wrap up kind of packaging it up and- um and I guess deciding whether that is the final destination or whether it's going to be something else right.

B

Yeah, I would say I would say karen, let eve figure out if he wants to change the name of it to anything that we want um and then kind of go from there.

A

So it's kind of like a full suite testing deployment package kind of, and that includes web ide. So I know last week we were talking about web id. I I don't think that was originally in scope, for this right is that so.

C

A

Something you you added based on the on the feedback in the last call, which is awesome.

B

Yeah, so it also uh added so so it wraps over uh clio, kiosd, nodeos cdt, um and that was a bear in itself to kind of pull over cdt and then get all the weird pointing to eso stuff fixed with that. um So it it she, it actually does a lot uh it wraps over a fair amount so that the end user, you know just can just basically use it as a dev environment and not really worry about.

B

uh You know oops. Where did I put this node at? Where did those files go and that kind of thing.

A

Cool, so would you what is this more than a docker image? It's it's like a oh yeah, yes, test, suite kind of uh best environment.

B

Yeah, it's a full-blown sort of dev environment is what I would call it great, because the biggest thing I really wanted out of it was not just having docker images and some some sort of thin layer over it, because it docker itself is a nightmare.

B

I've found this out, um so the biggest thing I really wanted out of it was was for the the actual environment itself to feel like you were working with regular clios and regular nodeos, and you know not have to sort of worry about docker stuff at all.

A

Great awesome, great um so we'll keep checking back in next week to see where we're at in terms of the final packaging it up and and all that sort of stuff. We'll awesome news will uh include in um the broader communication we're preparing.

A

uh Cool anything else to add to on that topic.

B

I guess the docker images themselves. We can also still end up um using the actual images for that for our own ci, cd and and that kind of thing, so there's sort of a multi-level of utility. Out of this thing,.

A

Cool yeah, so that I guess the images are an element you could go directly into the images or you can abstract it and use your more full suite is a is a higher level abstraction great awesome.

A

um I don't think we check back on this. One here notice. We uh is stan on the line confirmed the pin compiler for 3.0 will work for mandel or recommend upgrade to alternative version.

D

um You know I didn't touch this one somehow I missed it from my channel cool.

A

Sure does uh this kid have an issue.

D

Yeah yeah yeah the whole issue. This one.

B

D

B

Out from so uh so, what's what's this particular one because for the stuff that I had to do, I ended up having to re-implement, pin compiler, because I because I needed to get that to work for the the docker image stuff. So would that suffice for this or do we do? We want something else.

D

We need something else. I mean it just needs to be: uh reproducible, yeah, compiling procedure right so yeah. That's that's what it's doing.

D

So, okay, I don't know what do I have to do there I mean, maybe just I mean if you have already something for docker which is representable, then we're done.

B

I don't think we need to have many.

D

Advantages here.

B

Yeah, it's just as it's just a small set of scripts, for I mean it's really really really minimal stuff for um for the eos repo itself, to kind of pull down, clangs specific versions and then build the particular uh libraries that we need to build with clang and the ability eos.

D

E

Different from what.

D

We had with this minus I option and build script.

B

Yeah I mean so. The whole thing was, I didn't want to go back and re re-touch. The big crazy.

E

B

E

It's just a sort of very, very tiny.

A

And thank you. Ted for sharing the link to the repo I'll uh include a link here for our.

A

um So it sounds like we're covered dune already solves for this issue, so I'll check it as done.

D

Okay, oh, but I mean, if you have something that works and and can be used on on uh ubuntu without the docker yeah.

B

Yeah, no, it's it's like that part of it has nothing to do with the docker. I just needed to be able to have sort of that um that ability to build with a pin, compiler all right.

A

Beautiful okay, so anything to add on platform support.

D

Okay, there's a new new, stable release, a long-term release in ubuntu 22.04.

D

Uh-Huh, maybe we might have two.

A

D

Yeah 2204. it's a new long-term uh release and long-term support. That means it will be available till at least 2027 or something so it makes sense to target. I think it goes until 30.. I think it goes to 20 30.

F

So it's a goodie.

G

E

Already made the changes.

G

To mandel, to make it compatible with that, um we just yeah. We already have it we're already compatible with building on that new ubuntu.

D

Okay, so that's done yeah yeah when.

E

We start publishing.

G

uh Binary packages right right when we have, we have signing.

D

G

I want to do uh to me the what I've wanted to do for a long time, but I got a lot of pushback prior and blog one for doing it, but really what I want to do is I want the binaries to be more or less platform agnostic. So we don't have this sort of.

G

uh We don't have the sort of situation in the future.

D

Yeah it makes sense, but then you need to delete all the postgres. I think is no longer there right, but there are some weird dependencies.

G

Yeah I mean that's. The thing is that uh the probably the one dependency? That's still a problem in this ubuntu case is like ubuntu 220 uh ships with open ssl 1.1, whereas ubuntu 22 ships with open ssl 3.0, at least by default out of the box and yeah. Those are those are sort of problems that make having a portable binary so to speak, um an issue right now, but I think that's really something that I'd like to see us uh migrate to is have a binary. That would be indifferent to differences like that.

D

In qosd, it was also you usb library.

G

Yeah yeah and I've I've put a. I have a github issue up um asking basically for feedback on removing the functionality that requires those dependencies and replacing the functionality with something that does not require those dependencies. um So yeah, I'm aware.

C

G

And that yeah, that that is another one, that's very problematic, um particularly because one of those dependencies is lgpl. So we can't just like statically link into it or something like that. So uh I mean that's yeah, I'm aware of it and it's within the scope of things that I'm trying to sort out so that we can have your binding yeah.

D

If we split uh command line utilities from daemon in different packaging, then you, you reduce the dependencies and we need actually anyway, we need uh separate client utilities right which are separate from from the node use just for for also for alternative cpu architectures like arm so maybe yeah in the long term. Target would be just to to get clears on qsd away from nodeos packages different and separately, and then you, you, reduce the dependencies. Also.

A

And that sounds like that would not be in scope for mandela right. This is a more of a long-term goal.

D

Think yeah, it's I mean 3.1 would be to to I mean we have we're pushing too much time 3.1 already so.

A

Yeah: okay, cool, um okay, any other thoughts on platform, support.

A

All right, let's move on to code, um any updates on your site, ted or anyone else on how the finalizing of of our scope for 3.1 is going.

C

I think it's going pretty well, um we went we meet every tuesday at uh at one and uh looks like everything's coming in on time. uh Oci is kind of stepping up and cleaning up as much as they can, and you know getting uh you know extra stuff in there that won't blow the deadline or make it too risky.

C

So everything's looking pretty good. uh So it looks like we'll be done, end of may going into jungle, four early june or maybe even late, may.

A

Great, so that's pretty well right on track with what were we predicting here.

A

What do I have it.

A

I don't know I lost it, so we're.

C

A

uh Late may early june, we'll have basically all of 3.1 features ready for testing.

C

Yes and uh matthew, darwin has an early build on uh on a node right now that he's been able to look at some of the trace stuff, and um so he's he's been running. The the kind of the latest builds uh that that have quite a few features in already uh and and not seeing any any issues, at least yet.

A

Yeah the update I got from him was that yeah oci is doing great work on the transaction life cycle features um in particular. Those should be done in the next few days. This looks like and yeah we've already been testing the all the new features as they come, and things are looking good.

C

A

Beautiful anything else to add on code development strategy.

C

A

Right and uh testing I mentioned already: we matthew's been testing the new features as they come um any any updates from anyone else in terms of their success or not on testing.

F

Compile it and our public nodes or a public node, the one only one on jungle, four is now running: three: zero: five rc one.

F

H

And it works, I can query it.

H

So you're running, you said three three dot, one: three: zero: five rc one. uh Let me just.

F

Double check that pretty sure that's what it was uh jungle, four.

H

Yeah three zero five rc1.

F

And that's just on the public and the api and and seed pier and then we'll be moving, we'll be moving the block producer over to that in the next couple days.

A

Okay, public api and cpr: that's a bp.

G

A

G

F

It'll be it'll, be in the next few days, yeah.

A

And so far, so good, no, no, no problems from.

F

I try I try to get well. I tried to get running on 1804 in the beginning, um and I just had I mean I didn't really troubleshoot it. I know that it was uh suggested to use 20s 2004 and I had some some packages. Some of the dependencies wouldn't work and I just was like. ah Let me just go into what was recommended anyway.

A

All right, uh I wonder if this is, we have anyone else. Oh.

C

A

Here we go theosphere showing up on the on the validator running mandel. Yes, would you look at that looks like you're configured correctly.

H

You sound surprised.

F

I

No, not at all hey.

J

Ross, it looks.

I

It looks like you can blockchain good job.

J

Go dig out my uh that.

H

Wasn't patronizing at all.

J

At all it didn't, it didn't mean to be I'm sorry.

I

You know what a bp.json is ross. Apparently you do. I know.

F

I just I just turned it off and on a few times and eventually started working.

I

You hit it on the side a little bit. Oh there. It goes.

A

Yeah, we were also looking to get aloha benchmarks running that we know have. Did we get any success with that.

I

uh I uh reached out, they said they would, but there was no uh timeline or anything, so they still have jungle two on there, which is dead um right. It's just priority and uh it takes. I guess it takes time, so they did confirm that they'll add it in the queue at some point, but no, we don't have any um any timeline for that. Yet uh okay is jungle. So last night we were having an evm uh call and it was brought up that jungle earlier in the week started.

I

Having a lot of forks have those been resolved.

I

uh In general, I know matthew ross, you guys were aware there were a few. Other people um started forking off and I guess lib fell behind. uh Is that or did we well? I guess two questions, one, I'm guessing at this point that it's good. Can you confirm and two do. We know what happened as to why uh jungle started uh losing lib.

A

H

A

Is matthew on the call? uh No it's not so I can check with him offline and and report back.

I

A

So reports that jungle 4 was forking.

I

Yeah jungle yeah, it was losing lib, uh it was creating a lot of micro, forks and just looking to see if we can stabilize it, because that's where we're pointing people to for evm testing, that's where evm is currently deployed on, and so we get a lot of reports from new developers that are outside of the eos ecosystem that are setting up for the first time and their first experience is that eos works all the time. So it's pretty bad. So if we can stabilize jungle, that'd be great.

G

I

G

I

To whether or not is it because of madel, is it just because we've got a new bunch of new bps appeared for the first time on jungle? You know in a long time, just if we could figure out the the cause that'd be great.

A

Great yeah and then yeah, there's value outside of mandel testing to have it stable for the evm stuff. We've got a lot of.

I

Yeah for sure yeah, whether.

A

Or not it really first impressions.

I

Well, it's also whether or not like. Is it something perhaps in mandel. I don't think so. I'm thinking it's peering with dps that haven't appeared a long time, but if we know then it makes it much easier to then explain afterwards.

A

Okay, yeah I'll get an update from matthew on that and report back.

A

um That's all I have on the agenda anything else.

A

All right well, thank you. Everyone have a great week and we'll see you all around.

I