IPFS JS Core Dev Team Weekly Sync, 22 Oct 2018

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: ⚡️ⒿⓈ Core Dev Team Weekly Sync 🙌🏽 October 22, 2018

Description

Another exciting update from the JS core developers for IPFS \o/

https://github.com/ipfs/pm/issues/650#issuecomment-431834680

A

Alright, hello, everybody welcome to the J s core dev team weekly, think 22nd of October 2018. The note taker is Jacob Thank You Jacob.

A

How we doing everyone.

A

Okay, enthusiastic as always, I love it.

A

Cool right, what do we do? Normally? Okay, we've got a note-taker attendees pick your name only attendees list. If you're here, if you've got a weekly update, then put it in the notes. I've put the notes in the in the chat panel cool. Let's go I've got an update. Last week was the go. Ipfs team were in town, so I spent a little bit time with them, and also we managed to organize the first ipfs London meet up in two years.

A

I think I was one like previously two years ago, but it clearly didn't come to anything. So we started up again a kick started by the go team who get some really good talks and insights, and it was just awesome and everybody at the end of it, was like when it's the next one and at the time I was like I. Don't know, I literally just stop just finish this one. So so that's really encouraging very, very happy with that.

A

But what else happened this week and so I spend a lot of time reviewing Vasco's IP, NS, pub/sub PRS, there's like some of them quite big, and here yeah, they're, they're, all reviewed I think they need some work and then I need to review, but that that was a big hurdle to get over.

A

What else happened? Oh so, in last week, I released ipfs API is version 25, with a fix to for allowing chrome to upload files that were bigger than 150 megabytes, so that's cool and then almost immediately after I did that Olly told me that it didn't work with Jase ipfs and we didn't spot it immediately because it was only about. You only seem to apparent when you were running ipfs api and talking to a ipfs demon.

A

So this was there's this weird weird thing there, but it meant that we couldn't upload big files to to a JFS demon and it was because happy actually has a max bytes setting on its roots, which is set by default to two megabytes and weirdly in node. This doesn't like this. This. We don't see this happen like it doesn't. It doesn't say that the payloads too too big, but in the browser it does and I couldn't figure out exactly why.

A

My best guess is that we know in the browns of the stream HTTP actually buffers off everything before it sends it in one chunk and I'm wondering if there's some sort of thing in nodejs, which chunks up your your input and sends it sends it in chunks that are probably smaller than one Meg, and maybe that max bytes option is not max bytes for the entire payload, but per chunk from a stream, so I'm not entirely sure, but anyway, I tracked. It down, figured it out and there's a pull request.

A

I'm just waiting for ci to pass, but to get husband down today. So I haven't made much progress with getting it to rerun on fields and stuff like that and then so what else haven't M j SI p FS is 0. 33 is coming out soon with a new web. Ui are really sort of the least candidate, so people can try it out before they buy and it seems to work pretty well. There is a like there's a couple of things that just need to go in into the release and including a well.

A

Hopefully we can solve the issue where the web UI by default, doesn't point to a j-type EFS, the JFS api port. So we'll see if we fix that, before the actual release, I'm currently not blocked on anything I will hopefully release Jes IVFs 0:33 this week and.

A

Wait hang on I've had someone Victor's changed something in my notes. So, basically, because of like the Meetup because of releasing Jes IVFs, they were fairly free. I didn't get a lot of time to work on cid base again I'm, hoping that I might get time to work on it this week, cool!

A

That's me: does anyone have any quick questions? I.

B

Just have excitement for all the updates.

B

A

Take that cool who's who's next who's, next on the list; okay, so wait and so ask from Victor I guess that's just a note from me. I will check that out later good, okay, so Jacob, would you like to share with us your update beans, yeah.

C

Last week was primarily prepping for the lip PDP. Oh 24 release I have a released candidate out there. Now that I'll be using to test the ipfs cuz the test Suites.

C

For that to make sure we don't have any regression issues that includes delegate routing state machine and enabling relay by defaults and then Lapidus, which has already been released about 41 that has full switch support in it and then last week started looking at the interplanetary test lab all of the issues that have been around for the past year or so so getting up to speed on that as part of the test bed stuff that we'll be doing for loopy TP this year, or this quarter rather so I'll be getting working on a test bed proposal this week, based on all the the prior art that has already been done and then continue working with the go team on that and then work on the Lippe to be demon start that process for the Interop and testbed work to get those.

C

And then there is a PR out I pinged Pedro again for for pure ID, for support for different key types. So then I have a PR that I can post once. That's all done for ipfs to actually add it to the docks. In case anybody wants to run ipfs with different keys.

C

A

Really quick question: the p2p switch is that in yippee to be 0 24, yes,.

C

It will come out with that as well.

A

Good that won't, because that's a minor version so that we're not gonna pick that up automatically in the current version of yes, I. Think that's right!.

C

So I'll do a I'm testing right now against that and then once that goes I. Imagine we'll probably do that in a either a hotfix or minor version released for Jas I. Think if s.

B

A

B

People are picking up on liquids appear like the projects like stop getting it. These are the tension, and so other projects are jumping trying it out running our examples and people do I come to Jessica to appear as a source or like way to understand how we p2p works, because I wrote it.

B

Orioles are pretty good actually, and so, if we are not happy with API at all and I say this, because I now see like new dial methods appearing for the same stuff and so on, and so it would be great to like you just get everyone to think very carefully about like that top level.

B

Api I know that Allen has been doing the top level API for Jessup, EFS and so taking some ideas, like you, did an amazing presentation both at swap week and also at the meetup, which, actually, by the way, the presentation at our lives got Jesus super-excited to them to gossip, serve on Jess, so wash : jelly like checking away visa and getting mean you're gonna need more excited so that it chips that, like they.

B

And so yeah being very careful about that I think now it's the well but yeah like I know. Just like that in mind. It's all I guess really yeah.

C

I think going through some of the Damon stuff and then the testbed stuff will help like uncover things of. Oh. This would be way easier if we just did it this way, and maybe that's something. I can also talk with Alan about for that integration. Piece with JSF EFS. The see will make it could make that easier. I know, we've talked a little bit about future stuff like generators and things like that, but just general API.

B

The bureau and specifically, is really they ease to plug in your modules. One of the things- and this is you know like very raw like it- is something very obvious, but still very raw in terms of like the commutation and breath like it's still.

B

Something I am like reading a lot of notes down, but but it is about running contests for multiple of our modules, and these will go from things like pitch, swap but I, getting a big hop contest going on and where people can not just swap swap it swap and try different strategies, the key modules, pubs, etc, and so always having like a very clear way and I. Think the peer-to-peer is a little bit better than just because right now in the sense, because it's very easy to like getting a good HD.

B

But right now, for example, it's not easy to plug a different pops up, like the pops up, is very like attached to plots up so yeah in mind, because it might be a big thing of 2018 to add this multiple moments, where we invite the whole community to like, develop like and buy these modules. Like think about, like dependency injection a little bit because, like we want to be able to say, hey a subset module gets access to you like the connected connection manager.

B

Gok gets access to like the dialing sonically sitting like a knowing protocols, get it gets access to anything that tracks reputation. All the connections and Ike weed those inputs now as to figure out a strategy to like develop minimum spanning tree. It's something.

D

B

We invite, like the whole research community to hack with us, and to really compare it with peer to peer in outside PFS into this platform. For innovation, like we want to see like the next bleeding edge way to do the HTS and pubs and boutique a stand and beat swap or pocket change to be on top of IP offense, and we could appear so yeah. Another thing to have in mind expect monotonic singing.

A

Cool okay, shall we move on? Who is next who's? Next Cole? Could we have your update, T yeah.

E

Hello, so I came from only in the middle of last week and I started while I went through all of these reviews for all the parents right over, so I think I fixed always fail so Alan. When you have time to go again, it'd be my guess, then Allah knows also found a bug in status or pub/sub, which I already fixed I also finished.

E

The spectral puzzle for a pianist was in the previous week as I didn't came to the single, then I went to the DHT table as well. There was a problem with only 50 records interoperability.

E

Finally, I had the P art for its March and released, and then I created the arts for all other modules that were using Liberty record. Then I also created the PR for fixing age. That's only reported about the local option being a global option, not yes, ffs and, in the meantime, I also reviewed several Jacob be ours. For the liquids beer last latest release blocked I'm, just locked in an issue for interfaces at the store which is blocking another PR for Jessica fest.

E

We are was already reviewed by several people's, but it.

D

Needs to be released.

E

I think when we defeat in the free.

D

Door left permission, so the.

E

Really, fifth time just shake it out and for this week I want to finish the Ibanez over purpose of interrupt and get the HD ready for being enabled by default in jessup infests and with the interrupts.

F

E

F

Questions from.

B

B

So I didn't really get. What are the thing that you are waiting on me? It's.

E

It's a PR for an interface that to store it, so it was already reviewed by Jake event on I think it needs merchants really.

B

Wait when you say, like nobody fault, you have any recent good results on the intro part.

E

Yes, with the lip, your sphere record being fixed, now, I tried it before and it seemed good. But now after the releases, I went to check it again and also try to make some stress tests, as you recommend it before.

B

That's great news.

A

Awesome cool Thank, You Vaska Hugo. Can we have your update? Please.

A

You're, muted and not talking.

E

The microphone is appeared for you.

A

F

Think it works.

G

F

G

G

F

Find I should leave last week.

G

F

Finished the pipeline right now, I was wasn't able.

D

F

Perfect problems, but the only thing missing missing is the test running just because they are committed to be so. The pipelines run faster, so.

D

Just to give you.

F

A quick rundown through this on every look at this one. It's a simple commits, forget all the stuff the shakes deletes. The best will come after this one and then we'll get really released job. This C is a use case. I can.

A

A

F

So saying this is the pipeline so check it out? In this view, tests would come here in the middle, they are perfectly working. Just commented, you get the pre-release. This is a use case that talks with Ellen to get like RC versions published everything gets published to the next time like continuous deployment right and then, whenever you want to just manually, run the relief job to like promote this privilege to a.

D

F

An another use case we run interrupt, interrupt test on schedule. Raisins, it's not the scheduling part is not active, but that's just a simple configuration.

F

The important part here is the multi project relationship I'm, going to show you so basically, this job activates another job on the interrupt people and, as you can see, this is made to fail on purpose just so, you can see how this all works, and through this you can set up lots of stuff similar run, other types of tests and so related to this I think is enough for today we still having conversations with in fátima, guarding circle, CI and Whitlock CI, and all that stuff, hopefully I, will to talk today on time zone.

F

Another stuff. It's also came from lab week. This is like greenkeeper but we'd like a filter to only our packages. So the renovate is basically green paper, but it has the configuration we need. So if we can say okay naval through and then we give it whatever you want. This works with shareable configs like a slings, so we can have one in Asia for all our rifles or we can have a triple with each on configuration and we get nice for requests with the changes this one. It's not the the normal dependency of place.

F

Poly classes is on board. Booty quest related to renovate I was I. Wasn't too happy with this whole configuring, but, like me, so to be, we cow pursue also did want a couple requests to improve the speed of CTL. This, hopefully, will improve all the testing, all the ripples. One of them was swapping sub-command ensue exactly you can see some of the benchmarks here. It's improved a little bit, not that much, but it will open doors, swallow improvements and another one.

F

This is kind of bigger, is to be able to like run in it, because when you start with TTL, you've kind of run for comment and like tomatoes on for shell process with these public baths, you want me to spawn too, because you, when we need to Europe us the whole configuration you need, we don't need to like ranking. It run so coffee place, confit and then just do anything demon. So you can see the numbers are looking good, the and also I'll part with the Jessica.

F

If SEO grading that's for two and release it so.

A

I think that looks really cool, really promising, I'm glad that infra are informed about this and they know what's going on. So that's good I would still like to see Jess IVFs a fork of that running and try it out before we do this, which I want to make sure that this this experience is going to be better than what we have at the moment and but, like it looks, it looks amazing I'm super excited for it yeah any any other questions. Yes,.

D

Of why why this pipeline is implemented on on github, rather than Jenkins,.

F

This was a conversation we had van Marwijk, and this is just proof of concepts like the interesting is kind of looking to circle, CI and other options to improve old, old pipeline and CI situation, and then I went with Michael and we started testing with it. Club, CI and I said: I just did a simple setup for the or pipeline. This would be useful for our use cases not like permanent or anything just like a proof-of-concept, so painful team can really late. If this is any, has any value or not? Okay,.

D

And where it, where is this discussion happening so I can so I can jump in because I think so cuz? He is still suffers from the same problems. That was the reason why we moved away from it in the in the FIR. In the first place,.

F

Yeah I also I also be that's why I used it lab see instead of circle, but right now at least last week they spent a lot of time on circle, see I I still, don't know why? Because I still in at the opportunity like to call or something with them, because I think they all long like oh we're at a time zone. So it's like 10 a.m. for Mike or something like that. But it's all it's all been like slag and I'm, not on a nation.

F

We should probably start one when, when these precious comes to become more stable or whatever, what you think I discuss with down in just what it's like to make a little video explain all this, and probably on that issue, that video will talk about everything and get figured for any anyone access validated. This is.

D

I have a second question as well: the blocked it's a couple, ager peers or you blocked by those peers or or the PRS blocked themselves.

F

I'm, specially for the window, science right at least on one PR dissipation stuff- is also blocked by it's, not nothing, really majors, but it's really important one for the experimental. Also. So just couple peers.

B

We do you mind like turning off screen trader.

B

Yeah I think this is pretty cool. I knew a little neighbor I'm really happy seeing that there's progress, and then there is like a prototype like there's. Actually, research I would love to same time. There's a PR or an issue on infrastructure. Repo I'll actually like to see a PR to the guidelines, because it's changes. A lot in terms like I will review PR that we like new releases, where we check for stuff, like a lot of the burden, can be removed from the developer to see I, which is what see I supposed to do.

B

The whole I chain I'm like running, interrupt tests from like one of them dependencies Abed is really great. So if you will open a PR because really guidelines so that we see so we have an opportunity to see all these removing burden from the developer or adding other tasks that we are not seeing yet like. Who is going to maintain how this is it infrastructure? Is it as etc, easy the limitation of the module, for example?

B

And so, if you could like go through that document for requests to the places where you need to like who need to change, I think we will reveal a lot of like what it means to everything. One here and I can be one that is going to be a contributor to the project in the future.

F

Of course, in fact, that I'm, just kind of waiting for the talk I had scheduled with Michael and every night to talk about Circle, C I, and we 5 CI enough today to make our show describing anything in the video. So you can better understand how all this works. Awesome.

A

Cool okay: we are rapidly running out of time who's. Next, on the list, do we have anyone else called Volker? Would you like to go.

A

You are music, yeah,.

H

It didn't work. Okay, all right so like that is pretty quick and I did some breaking changes on IPL d1 merge for in review, and this those will help to make the API better, but also make help with who goes world on the mono science. Then I finally recorded my phosphor G talk. After all those weeks, it's published on YouTube check it out. If you care about data and then yeah I look into like have Viktor Bout and putting into CI issues, and next is for me working on the graph same stuff.

H

Basically, writing down the spec that Michael and I have in mind, which is different from the other graph upcoming graphs expect, but real yeah I will put it down. So we have a thing to discuss about um yeah. That's all I have.

A

I was I'm. Sorry, I should stop typing. When does anyone have any questions? Avoca I was just typing in a video of one presenting like some selector stuff and graphing stuff, which I haven't quite finished watching, yet that you should check it out. Yeah.

H

Which is basically yeah different format, so I watch it. It is different from what invitin I propose. Okay.

A

Cool well, do we have any other any other items? Does anyone have any general questions or problems or want to say anything.

A

Cool, if we, if we're all done here, then we're all done. Oh David.

B

One more thing which okay, if you just said that I'm sorry, because I was just like trying to do like a peer review.

B

From last week, I just.

H

B

30 minutes you can watch a 15 if you do 2x speed, ingressing and what it means for the notes. It is one of their show me what you got sessions.

B

The show how interested I think you device is very hard. My head keeps spinning between my oh, my my my apologies. Okay,.

A

Cool, we are over time now and it's been rad to talk to you. I will see you next week, I'm sure for a another, exciting session of a ipfs cortez talk to each other and tell each other what we're doing last week and this week and next week in the meantime, have fun IPF essing everyone bye.