IPFS JS Core Dev Team Weekly Sync, 17 Dec 2018

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: ⚡️ⒿⓈ Core Dev Team Weekly Sync 🙌🏽 December 17, 2018

Description

https://github.com/ipfs/team-mgmt/issues/650#issuecomment-447865137

A

According all, right hi, everybody welcome to the jazz core dev team weekly, sync, it's the 17th of December we're running out of time on this year. So let's do this quickly we're 5 minutes late. All hands ran over a little bit, but that's okay, we're still in good spirits. That coexisted a amazing demo on your hands. So you should go and watch that if you didn't see it the first time around cool, the notes are in the crypt pad. If you haven't already, please put your name here: 10 DS list jacob is not coming.

A

You know what this means guys can I get a volunteer for taking notes, raise your hand or type me in the chat all right Alex. Thank you, you're, a sweet, I'm cool! Okay! That's that's that! So what's next! Next we do what we normally do and we do it round of what you did last week, what you're going to be doing this week and stuff that you're blocked off. So let's get that over and done with, who is first, unlist is Jacob who's. Not here. Okay, so looks like he has finished. Wake finish, thought fixed.

A

Limpy to be not started yet: oh, yes, so there's an arrow with the p2p, throwing an error saying it's not started which scene so he fixed. He seems to have improved connections trekking for Limpy to be switched, which is great, he's been doing. Okay, our planning he's not blocked. He is finishing off some little poppy demon work next and the PHP crypto exploring support for the p2p keychain.

A

If you have any questions about that and please get in contact with Jacob after the call next on the list is Volker, would you care to share your update to profess hi.

B

I still work on the ILD, API stuff and so I currently work on spreading it into ipfs itself, studying with me to me/cfs importer and exporter, and it's surprisingly easier than I thought in the IP API seems to be to work quite well since yeah they were still working on that and yep. It's all.

A

That's any questions for Hoka I.

A

Have a question the so what what what does I mean? Have you implemented the API in all of the different formats or the hata, where, where.

B

Are you released so I've done two things, so one thing is I've, courted the big protocol buffers format to the new host spec, because it's the hardest one to do to see if it works and in parallel I also add the Lee there's. Also an API spec update for is IP LD and now my current plan is to first get that chase.

B

I only done so, it will still work with the old formats and then we can upgrade the whole chain basically but I plan to do it in January, so not not before end of the year and the are they sent I take it they're not released, you know, I make sure busy works and then I will double it up, and it also basically is the combination also with the pacing iterator and awesome endeavor. Basically that also the whole thing else, which is two promises.

A

Yes, continue, please assign me is like a reviewer when you do these I'd really like to okay. Yes,.

B

I will yeah so I will I will do it once I have something but I so clearly. So the plan is, as the read suggested that this week just to see if everything works and then yeah publish them before I need to change all of the accident she can assign.

C

Me as a review yeah.

B

A

Go okay! Thank you. Okay, any other questions. Okay,.

A

All right, I'm, moving on it's faster.

D

So last week, I updated the request with fixes that we did in GS and decide to have interoperability in the episode protein I already asked today review to Steven hoping to finish this during this quarter. Yet then in the DHT awesome, endeavor I, fixed Alan's reviews. However, when I enabled all the all the tests of the test fit with the DHT neighborhood I got a lot of flaky tests all over the codebase and I'm I've been mostly of last week and still this week fixing flight tests that appeared to like break now.

D

We know I've found most to guaran every time working finding my chain, but the CI still complains I I'm still working on it, then in the libera Pearman climates I finished tests for the basic commands and I also started the H documents. I have the Giacometti link. If you want to check it, I also close with jacob chauhan, okay, ours for just the PGP and I presented before the IPTV DHT test simulator, so I'm not blocked on anything right now, more or less in the HD stuff, but I think myself and Dolan will handle it.

D

So for this week, I want to finish the retrospective and Auclair scoring finished. The HD awesome endeavour stuff fix a car DHT bug the data reported that, basically, that we are having way too many peers connected and something is wrong, like Oh, add valuable input on that and I will look at it afterwards during the lipid fear.

D

Human client I want to finish the DHD comments to create per request. I will also be helping set Azure this week with the workshop in the university over and finally I want to also start the basic interrupt tests after finishing the PR for the demon client and that's all for me, any questions.

A

Cool thank much doesn't look like there's any questions. Next on the list is me okay, so it done when we're done so I've been implementing the new ad from methods in JSI BFS were in. You kill things for the HTTP client, but have been promoted to the route namespace and and are now in the interface like my first course. So they need to be in friend that I added one for ad from stream and ad from FS and from FSI foot merged there, but it's not actually managed yet. So that's a lie.

A

Sorry, but it's very close. It's really close. The CID based CID base option is now in and unmerged ready to go. So that's good. The the changes I did to the repo for Co ID version agnostic, getting so putting in CID v-0 and getting at getting it back with a CID v1.

A

They were done and there was some fallout from that that needed to be fixed up in the in upstream in Jos ipfs, so that was that was all done and merged and then I did a whole bunch of other PRS that are less important and less interesting and I did some reviewing of the DHE PR and project roadmap, and currently the only thing blocking the chair.

A

So if has Siri fateful release, is that PR that I mentioned that it hasn't quite been merged yet and from our FS as well as the DHT one, but I'd rather get the DHT one I, don't want to rush this and get it in really quickly. So, let's get it. Let's get it to a good standard before we release it.

A

But what I recommend, if you're at all interested in having a DHT nsj IVFs, is to switch into that branch and what your peers go up like mad, because we've got an interoperable DHT and it's super brigh and exciting see you should go an Avenue, it's really cool and it's so close. It's so close cool next up I need to do quarter. 1 2019 planning with you fine people, so we are running out of time.

A

If you could, please put your ideas for what you're going to be doing for quarter, one on the PR that I've made on the team management repo and fill in a retrospective I've been doing a retrospective today and thinking about what I'd like to do. I just need to know that down and we'll get that done, and then I'm presenting that on Wednesday, and that is me any questions.

C

I'm so vaguely on track 434 to come out before Christmas.

A

Yes and no, like I, said it's currently blocked by the DHD being ready to go in. That's that's, probably the only one that's gonna that is a can affect it and I'm Eva. Well, if it, if it went in tomorrow, I could maybe get it released into on time, but yeah it's looking unlikely. To be honest, it might be early January, the bead.

A

E

Can we do to help it happen? Isaac perhaps Pascal as I can more convenient, say.

D

You know so basically now I have all the tests, like nine and ten times passing in my computer, but in the CI, mostly in Linux and Mac OS I have like since I have flaky tests, like that's that usually test, but sometimes do not pass through timeouts or even more random. Stuff and I also have lots of JavaScript a peep out of memory and I'm kind of trying to fix them.

D

But as I have to wait like one hour to have a test running I and then only in the CI I'm, not sure how you can help and instead or my guess would be just try to run the tests in your computers. And if you have problems report to me or try to fix them.

E

So like it seems like it's like a resource problem because, like you just have a more powerful laptop, like you, don't see those things happening but like you can totally simulate a less powerful laptop.

E

Otherwise, it's like reducing the amount of memory that goes into nodes or even I, just creating a VM. That's like Seenu is a similar mix, Linux environment in CI. That will make it more fast, reproduced.

E

And in yeah I like not trying to distract everyone but like I, think it would be pretty cool if we managed to cut well if people have the time to kind of like help here, because it would be like one of those moments where that BFS team and I build a team I see that there is a requirement of a peer to peer and clearly not enough resources to deal with the demand. And so, if we can I just like help that, against the finish line, be awesome. If not, it does fine.

E

I saw a hand from Alan or Alex like okay Alex go ahead. Yeah.

C

I mean it's not over still quite a lot of work to be done on the DHT like. Is it worth, maybe pulling it from the release and putting it into 30? Fine, that's.

E

Also very point.

A

Yeah, if people are in agreement with that, then yeah we can. We can get Jess for ipfs 0:34 out the door tomorrow. If you like it, wait like yeah. Do you know that would be fine? It won't have this big bang for buck, as I said. Yes, it does at the moment.

E

Still a lot of exciting stuff there yeah.

A

Absolutely absolutely I am I, am fine either way.

E

Yeah and I think, like part of the recent work, was also it's easier, not been able to HD so, like you can still say, oh like if T is from alpha to Bella like just use this. Why? Because it's easier now to enable it in other users, can I help debug things if they want. Okay, can we take a vote yeah? That sounds good once 0:34?

E

Let's do two around so I captioning option B so like who wants 0:34 really this week without HD raise your head like so at four five by wants option B, which is wait for next year to like, like and again like wait for next year, doesn't mean like we have these ready on the first week of January. There's still work to do, but.

F

I can only really see.

E

Say again: oh I'm, sorry I vote it wrong. I mean like wait for next year for the whole thing.

E

E

Like so one side we have sheepish 0:34, which has a bunch of new features about super robust and reliable and solid and like the API, is fixes and so on, but I keep the 0:34 without the DHT component right, okay, new thing this week. The other option is do not release anything this week and and, like just say, vow it all together for next year the performance improvements if I can pay. My PFS is using all that stuff right. So, okay, just just get it out.

E

Like all the nice things we already have solid and robust this week, let's see some hands: okay, 1 2, 3, 4 5, something with mine, hey I, got off she'll, be like. Let's wait three weeks, we actually don't know to wait to do at the DT so that it can go out. Oh I, don't see so it's option B! You can vote now. Okay, I guess like we have five versus zero with some.

A

I would prefer to get something out, I think I.

E

Think specially because of I can't even type capacity scale, so I can Albert in here. Alex has been working on. I totally want to see this performance improve. Is there like those are huge cool.

A

Okay, who is decided thank you for your time and alright cool. So, let's move on quickly cuz. We just wasted a bunch of time, not wasted productive voting, but we got sidetracked anyway. Who is next I? Can brain Alex.

C

C

Change the XFS importer to use an in-memory IPL d. This is great because earnings of the tester now we're not Foster to run and they all run in the browser which previously they didn't, which is great so easily yes, so the ones I race, though, are the massive directories and they need a bit of performance enhancement before they'll. Actually work reliably on Jenkins I started updating the ipfs HTTP project, which is kind of the next generation.

C

Imagine it's pretty ipfs for the base is trying to update that to work was kind of the new changes to the object, API and all that kind of stuff and I got sick, which is really boring and so I had you know day out last week, I wasn't working on their scoring and and the retro stuff, and next up is okar planning, finish off the HTTP thing, and then I really want to look at the repo migration cool now probably meet the end of the week in questions.

C

Can't see any hands cool.

A

Okay, cool, thank you Alex! Just around and now it's gone: okay, Hugo whoa you're, a pig.

G

Hi guys so I've been working on the bundle size, stuff, I've, worn, Pakistani, Asia about making karma and barbell builds. All default. I also did prototype for the azure dev ops, CI, which doesn't look very promising because it doesn't allow for a single job restart but yeah. Let's see also did already did a rework on the error proposal. It's wrapping I'll be applying the poly quest today or tomorrow and yeah.

G

The rest is all about the pokey questing every every repo possible, reducing the model size httpclient, it's done only p2p is also done, I'm finishing up the JSF FS, so it's almost almost done and the results are pretty pretty great but yeah. You will see the final ones today or tomorrow with some nice pictures, so you can feel the reduction I've made I'm blocked in a couple of PRS.

G

The main one is the AG one, but I still need to figure out, what's happening with the some karma package that I'm getting stalled on or are detected by karma on drinking's and after that, I can start merging and bubbling up all the poker quests and for this week of finish, the window. Size okie are the error. Codes. Okay are hopefully, and today or tomorrow the rattle the retrospective and for the rests. It will be planning. Okie are planning any questions.

A

A

Material is not attending this one just quickly, so there's a bunch of poll requests and issues there that have been done and he's plucked on a few there as well. I would check them out in a bit more analysis. Next for Matteo and improving ASIC management and reduced pressure on the event. Loop and GCD, cool crypto and compact be a lift. I'm cool and the other Alex is also not attending today, and this is arrived note, okay, cool and he's done. First go benchmark, that's cool for the benchmarks run, run clinic I.

A

Guess that just means it. It runs now. That's great, fixed scripts and fixing scripts, okay and he's blocked on storing stuff on ipfs. Okay, I will also look at that at some point and Ron you are here. Would you like to give us your update.

H

Yes, so yeah so I'm integrating what I did ok so last week, mostly fixed, did a lot of fixing on just a runner running it on the dashboard in the on the bare metal machine. So last week, I think the dashboard was broken, so that's been fixed. We've also added a lot more tags to it now, so we have a further before the whole daily run. We have a UID. We can actually map against you know so not just a branch or a version, but just even each run can be tagged and then also yeah.

H

We have the branch in version, so you can. You know, drill down to find your specific criteria that you're looking for in the dashboard. So that's that's, looking a lot better and it's running and we added IPS version to the info box. Db tag so mostly just a lot of fixes around the dashboard itself and also did so working with integrating the go benchmark. So it's very close. The PR is out I just there's a couple of fixes.

H

I need to do to that, but that will be probably landed within a couple hours from today and then also again to be integrating the browser benchmarks, which will be at least one and then a couple more bug fixes which potato's been jotting down. So that's me for this week, so any questions.

A

Alright, it looks like you got off Scott. Thank you all right. We are at the end of the list of people. Do we have we have a few minutes left? Is anyone her? Oh, we had so other note. So some people are on holiday because I don't know, there's something going on I, guess and and then so I put this down no call on 24th of December or the 31st of December, because it was also on the All Hands call. The people agreed that we won't be around for that ones.

A

Up good, yes, again will be like I, don't know drunk or something and okay. If do we have any other like questions or comments or anything about anything at all before we sign off for this year,.

A

No okay I would like to add: it's been very nice working with you all for this past year and I very much look forward to doing it all over again. Next year. Yes, yeah.

A

Cool okay: ours do do that thing because I need to present them and it's awkward when I don't have anything to present so cool all right. Thank you. All for attending and I'll.

F

See you next year have a fun holiday time all right, bye,.