Node.js Technical Steering Committee, 29 Aug 2018

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Node.js Technical Steering Committee meeting - Aug 29 2018

Description

https://github.com/nodejs/TSC/issues/590

A

So welcome to the note, G s, technical steering committee meeting for August 29th 2018, we're gonna, follow our standard agenda. So the first thing on the agenda is announcements and I. Think we've got some invited representatives from the foundation to give us a quick update on the jeaious interactive.

B

Hi, this is a we have a Tracy and Lera and Greg on the call from the no foundation, / linux foundation, side and maybe Laura. You can start with some event updates and then we can move from there. Yeah.

C

Sure we are um we're tracking well for registration, weird about 40% higher than we were last year. At this point, any cross promotion would be greatly appreciated and drive, excitement about the event and from you all in terms of promoting that you're attending and are speaking at the event. I know that we're working on some Twitter cards for speakers, so there should be being sent out with the new rebrand within the next week or so so we're just working on those. And then we have some really I.

C

Don't have ton of updates from us, but we kind of featured some of the add-on events. We are doing I can't remember if we mentioned this, but we're doing a job fair that sponsors who are exhibiting. It will be day two so on Thursday morning for those who are looking to move careers or you know going from independent adjoining company, it's an open kind of job for about the attendees and then an opportunity for the sponsors to do any recruitment if they have any job racks open.

C

So that is a feature, and that is added from last year um and I. Think that and then we have the community corner and that we talked about last week as well, and so we're just getting that all settled, but that will be in kind of a common space for people to stop by and get updates and I. Think that is it Tracy or Greg. Do you have anything to add.

B

D

B

One thing is, there is still opportunity as well if your company is actually recruiting to sponsor Jace interactive, so you can always email myself ladylee at nodejs org and we can pealed it to the right people. In addition to that, just to elaborate on the node community corner I did send out. I herded an issue for this, and we have a few folks who are going to be helping program that contents. This is a this.

B

Is a community focused corner, so Jim Matteo, waleed, Michael, Tierney, Aguirre, Joe and Joey have all volunteered to be a part of that. So myself and Greg will be talking to everybody at this Friday at 8 a.m. Pacific about the community corner.

B

So if anybody has any questions about that related stuff super happy to chat through it. Does anybody have any questions about collab summit or community corner or specifically JSI related stuff.

A

Great thanks, yeah.

D

A

I'm just gonna add the one thing is: if there, in terms of the community corner, if there's anything that you know, people think should be or would be a good addition. Even if you can't necessarily you know, contribute it yourself would be good to reach out to one of the people that Tracy mentioned in terms of making that as a suggestion, because we do want to curate it. You know so that it reflects what we think you know the teasing comb comb think should actually be covered in that community corner yeah.

B

Actually, an another thing, good thing to bring up and I I believe Michael, you and I were talking about. This was the getting the working groups more involved right. There has been so many amazing working groups, I think we're talking about this charity working group giving an update I'm any other. So if you know any working group folks who might be interested or who you know want to give ten to fifteen minute update or maybe even hosts like office hours or anything like that, all of that is great opportunity at the community corner.

D

Cool thanks, I guess: I only have a couple really really quick ones and then I know that there's a full agenda so um don't want to take a lot of everyone's time. So the first is, um you know, I know we all know this and it we're all trained to because we've been calling it JSI for so long, but it is renamed. So this is node and jeaious interactive, it's a little longer, but I think it's a it's a better name, because you know from the beginning.

D

This has been a an event being put on by both the node and the jas foundations. So now the name reflects that.

D

So it's you know it's totally normal and I find myself saying JSI all the time, but if we can remember that the new name is no damn jeaious interactive, that's a good thing and then I do want to let you all know that I'm going to be sending out today or tomorrow, the monthly newsletter from the node foundation and we'll be highlighting in that a new, a new panel that has just been added it's a keynote panel and it's going to it's really awesome. So it's already up on the agenda.

D

You can check it out, but we've we've just added it. So this is going to be just just enterprises talking about how they're, using nodejs at massive scale. So we'll have you know I'm from Netflix we have Jessica Chan from Pinterest. We have Mike hisk s-- from Lowe's the the home improvement store, and then we have ad and I'm forgetting his last name zippy. Maybe you can help me from Turner casting so pretty psyched about that and that's I think everything.

D

So if anyone has any questions, you know how to reach us, but thanks for for allowing us to join the call.

A

Okay, thank you for the update any other announcements that people have have to share.

E

Franz is waiting to be promoted. Okay,.

A

Done, okay, so I guess moving on to the standard agenda, then the first one on the agenda is under note: GS note stream what WG ad, what WG streams oxide sorry I miss. No, that's the first one yeah ad. What do we do? Streams, 23, 22, 3, 5, 2, so.

F

In the issue on this, Mateo has a request that it be delayed since there's a very long response replied by a Timothy. It cannot be addressed and involves on my reply, which is also quite long, which I'm sure yeah I would.

E

Like to if I hope, we do that, we postpone this, but actually what I, what I would I would suggest. We do is remove it from the agenda. We've talked about it, there's stuff going on and if it needs to come back, someone can riad the agenda label. I, don't want to tell Otto appear every week, so I'm gonna go ahead and do that right now, unless somebody objects I'm plus.

A

G

A

Just taking the notes in the issue.

A

E

If someone wants to volunteer to take notes, so Michael doesn't have to that would be awesome.

G

Yeah say your keyboard sense: a little quieter there Mike, maybe.

A

I'm, just not pounding it quite quite as hard.

G

I do have my headset.

A

A

Okay, so next one FS feature detection for recursive mcdeere. This is 22 302 now I do think eyes looks similar, 22 302, there's a question two miles on the issue asking what was agenda to the issue with the suggestions of France Franciscan suggested it was better being discussed in the issue itself.

G

In this particular, one I think the reason it was added to the agenda was because not specific to this individual issue, but there needs to be a more general conversation about how we want to do feature detection for these types of changes. So this in the particulars change it was made. It was an option that was added to the function, so there's no way of easily doing feature detection. You know, like what versions of note is this thing enabled.

G

In there's it's more of a general question of how do we want to do this kind of stuff overall so that it's not just a one-off for this one individual feature? Now we can have that conversation in a github disco. You know in a github issue as long as the discussions actually had right so I think it was only put on the agenda just to queue up, be the conversation.

E

And this really can't land until a decision is made right because oh yeah yeah so or rather it's landed and master can't land in a release until a decision is made, because if we change it, you know then.

G

Yeah then, once it lands were pretty much stuck with this one and until we can get there, deprecations like this, so yeah.

E

Though, how do we, how do we get to where we're having a conversation and making a decision is? Should this be a separate issue, or should we do it in that issue, or should we set up a meeting and, what's the what's, the path to like arrive at a decision on this I think.

G

A separate discussion issue would be where we certain can start collecting some ideas. If we, if we are not working towards a resolution, then having a meeting supposed to be round this Monday good.

G

That's what I would recommend it I mean I'm happy for there's. No other ideas.

E

F

So well what I'll.

E

Do what I'll do if no, if no one objects, I'll, take the CSC agenda label off of that item and add blocked and I'll open a new issue summarizing what I think it is? We need to decide and I'll ping, TSE and James and Michael and miles, and anybody else can correct me if I, if I've gotten it wrong and the reason.

A

Right and are you so you're kind of Valen? What I was gonna say is we need somebody who's gonna push it forward. Is you volunteering to do that or no.

E

I'm volunteering to open the issue, but I'm not gonna, be the champion of this one I think miles might be a good person for that, but he's not here. So it's easy for me to volunteer him. Okay,.

A

Maybe maybe maybe my champion so maybe just in the issue you open, you can ask him post that question because I think we kind of need somebody who's, gonna champion it right. I.

A

Mean the alternative, is we just decide we're doing nothing at this point right.

A

Okay, so I've got actions that rich is gonna volunteer. To create this, a pretty shoe. Maybe miles can be the champion anything else. We want to discuss on that one before we move on.

A

Okay, so two two two one eight is the next one on the list. Although I see that there was some discussion and I think that one was it actually removed, let's just double check that.

A

Yeah, it looks like it was removed from the agenda, so we can move on to the next one.

H

Two one eight five, seven, where we need to compare them and figure out with which approach me once you go, so that is just the next one coming up right.

I

A

Well from agenda right, you've already got the comment: okay, so moving on to the next one, then you'd expose promised events to one eight five, seven.

A

Take a look unless that's.

E

Me but I am like working on that other issue, that I said I'd open just now: okay,.

H

E

Struggling with the new github template, where I have to decide whether I need help with node or found a problem with node or what we need to come back to that one. If we could, if we could come back to this one, that would be great sure.

A

We'll do so moving on to the next one, two one 766 password hashing, API, I, think and I'm just looking in the original issue, because I think chakra commented on that too one eight five, seven.

A

Although, if I got the right number, there now sorry 2 1 7, 3, 2, 1, 7, 6 X.

A

If I could type correctly, it might actually show up right. So an update for truckers. So there's no updates. He hopes to have an update by next week, so I think unless anybody else has something to add on add to that one we will move on and the next one actually is to 135 1.

A

So basically he says his comment was that he has filed two to five eight four, so this is basically related to the DEP runtime deprecated buffer constructor everywhere by default. So he's filed 2 to 5 84 and would appreciate comments there.

A

Any other comments, questions thing to discuss on that. One.

A

No okay, um so the next one is it'll expose promise events to one eight five, seven again, just looking up whoo I did that unless anybody else jumps in this. This is the one that you were just yeah.

E

That's also me: okay,.

A

So two hundred five seven is that so that's really okay, just part of the same one. So.

E

Yeah so I can I can talk about that now, if you want okay, sure and both both that with this one in the previous one.

E

um So the implementations are very similar, well have a lot of similarities but ultimately- and we might want to loop in Benjamin, groom, almond and Gus- maybe on this conversation, but ultimately the decision that and Ruben to perhaps the question that Benjamin Benjamin Benjamin's is still the question down to this, which is, should node attempt to be as useful as possible and then ask people to opt out of features like warnings about you know, promises that have been resolved after they've been rejected and things like that or should note attempt to be as configurable as possible to make the least assumptions about user code and people should opt into instrumentation features, and so I'm not sure if this is something that would be.

E

That would be relegated to the Diagnostics working group if that's even appropriate, or what how to proceed on this. But that's that's the question right. So.

G

At this point, we have two different api's that are roughly roughly doing the same thing.

G

There are variances between them and it really just comes down to kind of what you know similar to the feature detection thing: we have to kind of figure out how we want to do it and then what we want to continue to support over time. You know and then decide which one of these approaches we want to go with or reconcile them into a third common approach.

G

Right, I, don't know if we're going to be able to get clear resolution on that now without bringing you know, Gus and and Reuben and Benjamin together into a call. So what I would like to suggest is that we actually schedule a call for this topic and we and we try to get it hammered out now. The discussion between these individuals has not always been the most productive, because I think that it's just there's just been a lot of contention in the conversation.

G

So we want to you know. We should probably make sure that it that we have some good control. Its conversation facilitation happening, make.

E

Sure that we're able to make progress so I think they've done pretty good lately in and I think they're. Actually, there's there's there's much more agreement between them than difference at this point. I could be mistaken, but it seems like a lot of the contention has died down, which is nice.

E

And and Benjamin is probably a really good person to facilitate such a conversation. um I do ii-I've had a hard time nailing down whether a decision is wanted on these two PRS or whether a broader decision is wanted about the approach we should take. I think they want a broader decision about.

E

You know you know, do you know, do we want to have people have to opt out of these types of these types of things, or do we want people to have to opt into them and basically make a decision, and that will be the node way. You know, which I think might make sense, but I'm not sure I will.

G

Say that as early as this morning, I was asked to decide one way or the other, and it let you know, I I think. As far as the public conversation is concerned, there's not as much contention but I still think that there's there's still either this or that thing happening that we need to whispers, and you know so I definitely think we still need to have a conversation.

G

A

Mean it's: it's yeah.

E

A

Can see if it's a this or that like asking for the broader context, it still means there's a fundamental disagreement that that would help inform right. Yep I mean it is it is you know one approach: is you try and basically figure out what we think people will will need and make that default? The other one is like people have to discover it themselves.

A

There are pros and cons to both for sure, so so the next step that sounds like they're recommended for next meeting. You know what we need is somebody who's. That sounds reasonable to me. Do we need somebody to can facilitate and set up that meeting I.

E

Tried raising my hand and I felt completely sorry we should we should. We should raise hands when we're talking, I think I, don't it sounds like only the three of us have anything to say, because there were some pauses in there, but just in case somebody else has something to say. We should use the raise hand feature before talking.

E

Aye aye, aye, aye I would want to know if this makes. If this is something the diagnostics working group would is, that is this something that falls into their or domain? Or is this not something you know? What you know like is is is a warning about. A promise you know is, is instrumentation and all that sort of thing you know is that you know is that is that something that the that they that they would I got distracted there, that they would have opinions on, and that would be relevant to diagnostics in general.

E

Michael has his hand literally I.

A

Can't actually hand in the tool because I'm the but but the the point I'd make is like I think actually, if we just want the what's the best answer in this case, you know, involving the diagnostic working group, makes sense to me if we want the. What is the general approach to note in terms of options? Gonna be, that's actually I think a bit of a broader thing, because it doesn't necessarily apply only to diagnostic options.

A

It apply to all sorts of options right and it's which of those two questions we think we want to answer, would determine that. In my mind,.

G

I'd like to you know, Jeremiah you yeah I, said you put in the chat here: I like good, informative defaults.

G

That's correct, but I wanted to see if you could get a little more clarification on what that means. In this particular case, I.

F

Mean in this case, it probably means, if they're saying that I think the user might want to know about in a development scenario, particularly where they may be a first case user. Then we should then the program should say something rather than being fired. I.

G

Think that's the kind of feedback we need and just kind of lay that out on these API, these API choices and which one of these is actually not actually gets us to those kind. You know to that goal all right.

G

Because so far me, yes, the conversation has gotten better I think, but if it's not a fantastic job of facilitating that and breaking it down to okay, what are the actual decision e be made, but we still have these open PRS and we still need just decision. Is it gonna be one or the other or something else right and I? Don't think we're getting closer to that particular discussion? I! Don't.

E

I, don't think we need to figure out the PR issue, at least not right away. I think I. Think the broader issue is the one that we need to figure out right away, and but it's it's entirely possible that once well anyway, yeah I'm gonna, stop rambling I'm, gonna, I I typed, in what I think we should do in the in the chat, which is have a separate meeting or a portion of the next meeting and invite Benjamin yeah.

A

So I think that's why I agree with that. Unless there's any objections like maybe if you open the issue, there's some discussion there, we put it on the agenda so that we talk about it in you know or or do people think it's big enough? We should just try and schedule another full TSE meeting time.

E

I think we should have my hand raised. We should we should. We should try to make it part of the next meeting and if it ends up just like blowing out the schedule and taking 45 minutes, then you still don't have an answer. Then yes, schedule is a separate meeting. Let's do the small. Let's do that. Let's do the least intrusive thing for us.

E

In fact, first thing: they're gonna do is I'm gonna, open an issue and label it TSC agenda and remove the label from those two PRS and just mark them as blocked on that issue, and if it turns out that the entire TSC is an agreement that you know about the approach we should take, then you know we don't even need to have have a meeting. We have agreement. If everybody agrees with Jeremiah that you know we fall should be to inform the user. Then, let's, let's, let's just do that: yep! Okay, that sounds good. Okay,.

A

Any other discussion comments before we move on.

A

No okay, so the next issue in under note ESC is update the Charter. So this is 569 I! Think rich. You were the originator of that right. Yes,.

E

And the only thing I want to say on this one is that before sending it to the board, I would love to have unanimity. I know, I've wrote in there that I wanted a majority of TSC people to thumbs-up it. I would love to have unanimity and here's. Why? Right now we have right now there are 13 of 18 of us who have who have endorsed it with me.

E

This fundamentally changes the way that management of the github organization happens. It may not change it in practice, but well, it probably will anyway. The thing that this does, along with the corresponding changes to the community committee charter, is that instead of the TSC being the sole you know having sole authority and responsibility over the github organization, it is shared between the TSC and calm calm. Now this is effectively what happens because of the github management policy, but the TSC can can. Can you know, declare that null and void at any time?

E

With these charter changes, the community community is on a full equal footing and we can't take their authority away from them. Only the board can do that, and once we've gone down this path it will be difficult or impossible to I am in favor of this as are apparently ever you know, everybody else, who's looked at it, but I don't want this to be something that sneaks in and is a surprise I don't want. You know, and in particular I mean I, don't want to name names but they're there.

E

There are one or two people who have not commented or approved it, and I really want those people to have a say if they have a problem with this approach,.

G

Just be clear, this does not change any aspect of the existing. You could have organization management policy. That's over in the oven. Repo just basically makes it more more of an official stamp yeah. Well.

E

It means that that that the shared responsibility between Kham, Kham and TSE comes from the board right now. Tsc has authority over the github org and the TSE has chosen to share that authority with Kham Kham.

E

G

It just says it's a it's official board, recognized yeah.

E

So yeah and well I mean you know it changes.

A

Would be where the authority for the policy comes from.

E

And and and I think that you know like right now. For example, the only there's only two members of calm calm who have the owner role on the github organization and.

E

You know like like if, right now, someone said you know, the calm. Calm argues that that more people on calm, calm, everybody on calm calm, should have that role. The TSC has a lot of leverage to say to say no I think that, like this will reduce leverage in situations like that, my and I don't have a problem with that, but I also don't want that to be something that like happens, you know getting snuck in yeah.

G

So, in terms of that may be probably the one thing that I would like to see or kind of more more rigorously documented process for how someone becomes a member of the calm, calm and then gets that guess that commitment or that bad owner bit.

G

So you know so my that would kind of be. My only ask there because of the potential security issues and ramifications are brown with that I mean we've talked before about taking the owner privilege away from TSC members because of of the security issue and just having more automation in place right, so I think we need to make sure that we have a really good reason for making that a more liberal process. If that, what it's going to be I, hope that makes sense. I'm.

A

Not you know so, but I think that's a separate I mean I. Think the point rich was saying: if you feel strongly enough that the TSC needs to be able to impose that, then you know you might want to object to this right or like cuz. That's going to reduce the ability of the TSE to to you know, influence that decision. Yeah.

G

For me, it's not something that the I would read. I would use it as a point of rejecting it. I would you know, I mean I signed off on it. Yeah.

A

G

It's more that there you know, I think it needs to be a condition of of approving this. Does that make sense, I.

E

Think you need to put that comment in the issue because it's gonna get lost here. Yeah.

A

I think Rich's point: he wants to he's being extra, careful that it doesn't just land if there's people things should happen. First,.

A

And I guess rich! Is it that we need an email to the rest of the people to say please approve because you want unit item II, all.

E

Email people, whether or email TSE or email them privately I'll figure. You.

A

Out one one of the two right yeah, it makes sense.

A

Okay, so any other discussion on this front before we move on.

A

Okay, if not there's the tracking next one is tracking issue for TC onboard million meetings. We don't have miles so I guess! Oh no! We do sorry I'm looking at the wrong place. Yeah! We don't have miles today, so I guess we'll skip that. One.

A

Okay and then the last one is strategic initiatives, so tracking shoot for two three knots get the list.

A

Okay, so James http/2.

G

Hb2 is graduated from experimental and I'm in master we haven't, we haven't, landed those even those computers in Nam and release. Yet so it hasn't gone now stable and a release yet, but that should be coming soon. There's a couple of PRS that need to be to make sure that land with that as a patient, they become super major but other than that. I.

A

Mean you who that's gonna, go back to ten I, guess initially and yeah.

G

That'll go to ten it's stable in ten and at once, all the PRS make their way back to eight will be able to, but in also okay.

A

So there is of the stuff so you're just gonna ask this: there is a plan to back port it into eight as well yeah.

G

A little more complicated to some of the fixes and stuff I mean you know. We need to be kind of rebirth just a little bit, but you know with that. That will come so no lag it'll like a little bit, but oh yeah. You know I honestly think now that it's out of experimental we could probably take it off the strategic initiatives list. Okay,.

A

If you want to PR that that sounds reasonable okay modules.

A

I'm, sorry, we don't have miles soul. Skip that one NEP I still focus is on basically we're still putting together the walk she would put together. The next milestone set of milestones. So you can actually go. Take a look at that. If you want to see the specifics, but it's it's gonna be focused on testing evangelization and some things to basically support adoption.

A

We don't have rod for OpenSSL evolution, I, guess, I haven't heard Anna, so I, don't know we don't have Anna for workers. We don't have to wait for error messages so back to James on core promise. Api.

G

Core promised API is pretty much steady state right now, don't want to push too hard on it until we get more validation, FS and DNS promises, or you know, or what they could be right, and that approach not a huge push on that. Yet I'm thinking after 11 goes out the door we'll work on it a little bit more.

C

A

G

It really just comes down to making sure we're doing it right. Yeah.

A

So giving people basically time to comment on the experimental, api's and yep get experience and see and then once we're comfortable move on okay governance, we don't have miles Arad new stream. Api is Jeremiah.

F

Yeah, so using the experimental sort of minimal pull streams, I have a working, TCP, socket sort of experimental doesn't do everything. Obviously they note sockets do but it's up there and it works. You can make HTTP requests over if you want to so that's that's there. I will send a link to that and post a link in the minutes and also every one, of course, relating to our interest in streams is kind of heavily discussing huh what we G streams at this time. Okay,.

A

Okay, okay, v8 currency: do you have Michael, don't think they have Michael's ah so at the moment, I'm moving on so async hook. We don't have a Lee, so we'll skip that one and open winter standards. We don't have miles okay, so that covers the review strategic initiatives and that takes gets us to the end of our scheduled agenda items. Are there any other things people want to add in or discuss I.

E

Just want to say much respect to Franz II, for forgetting for for proof, or you know, attacking the agenda to make sure that it it was. It was as lean and representative and ready to go as it could be. Yep.

A

I will echo that many thanks.

G

I will just you know, make note we are making progress on the process binding to internal binding lots of law. Requests have moved it out now. Part of that was this interim step of this fall through right were things that are looked at. Internal binding will still work with process blending.

G

At some point, we will need to add a deprecation in there to get people to stop using process finding, but it's going to take a long time and we need folks just kind of focused and paying attention on that part of that deprecation process, maybe introducing api's to support some of those use cases for why people are using toss was fighting in the first place and there may need to be some difficult conversations about what we need to add, or we don't think should be added, so just ask folks to be paying attention to that and engaging in those conversations right.

A

Is is there a good way like, as opposed to trying to follow all the different PRS and stuff like that? Is there somewhere that focuses that discussion or.

G

Is a tracking issue for trucking natural progress of those individual PRS? But, to be honest, we won't know all of the uses that may be impacted until we actually start introducing the runtime deprecation and we start seeing those screams rights.

A

Alright, but I guess I'm just like suggesting like this, is it that you know if we can give people that link? That's the link to sort of pay attention to what's going on or yeah.

G

The the tracking issue is probably the closest thing we have right now give.

A

That link Andy I can pick.

G

Up whatever yeah.

A

E

Jeremiah was that an error in the chatter we're trying to interject something.

F

No, that's that's. The link. I was having trouble actually copy pasting on my computer into Jim chat.

I

But you nailed it no.

I

Sorry Richard, you see as.

A

Sometimes it seems to just default the private somehow so yes.

I

Weird: okay, okay,.

A

Add it in two minutes, please five I was gonna, say please, please follow tracking.

A

Okay, any other shoes.

A

Okay, let's take a quick look to see if there's any questions on the stream. Does anybody have that handy, yeah.

F

I solicited for questions, but there haven't been any okay.

A

So I guess we can assume there's no questions this week this week. Okay, so I think we will call that for the meeting today and if everybody can stick around for the short private session I'm just going to stop the stream.