Node.js Core Technical Committee, 12 Oct 2016

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: 2016-10-12 Node.js Foundation Core Technical Committee (CTC) Meeting

Description

https://github.com/nodejs/node/issues/9020

A

Technical committee meeting for October 12 2016 and we are going to start with a stand up and I believe the first person this week will be Nikita.

B

Actually, nothing for me that is worth mention. I was busy for the last week and time catching up now. Next in colon.

C

So I was reviewing issues in poor press. The usual opened a couple for festival. That's I left off Evan next.

D

Yeah I open a few PRS and you pull request, reviews and read a Chrome extension and they generating review metadata easier. Next is Jeremiah.

E

I've been doing miscellaneous thousand issues, I've been working a bit with the chakra core team to get wrote, IPS modules, implementation using chakra core I'm, also still in the process of doing the v6 point. Oh eight point Oh release, which is just waiting on tests. The finish I think. um Maybe there was something else, but I, don't remember it. Michael Dawson I.

F

Just reviewed miscellaneous reviews and lands got back to working on adding the nightly code coverage build a little bit of work on the a bi-stable API proves concept and keeping up the data on general issues. Next is Brian White.

G

Okay, so I've been working on improving performance for some of the various and coatings for other for doing string, encodings decoding and coding, mostly on that just the normal, reviewing, pr's and and coming on issues. Next is Ellie.

H

Yeah I've been traveling last week, so I'm done much and I'm on the phone, so I don't know who's next.

F

I

I'm reliving it really some issues and PR already clifton. The next city is a place. The four west.

C

J

Giggy, give you mean me.

I

Yes, sorry I get a process pronounce this your name.

J

No no problem. Last week, I've been catching up with whatever happened or the last week, and now I I was working on the world peace which I submitted, distributing fears and doing comments. That's it.

A

Next up is me rich and I've done some issue in PR review. I did some been having some discussions and open the issue or two around governance, one of which will be discussing today and been doing some work for the prep for the outreach, II, mentoring, stuff that Traci and the out Richie foundation or whatever they're called, have put together working with some applicants and stuff like that and that's everybody unless somebody joined while we were doing that it doesn't look like it.

A

So, let's move on to I guess we go to the previous meeting first huh and make sure all that business is finished or scheduled for this meeting. Last week we talked about the supported platform list that has been supplied by the build group and I. Think it's pretty much approved. If you have anything to comment or add to that either go to issue 8, 9, 22 or say something right now then there's a consider, deprecating, int, L I, don't remember what the result of that was. Was that sent back to github? Or does anybody remember.

A

Okay, I'm looking.

A

The issues closed, so it looks like whatever it is. It's been resolved. I want to look at last week's minutes to determine what the resolution was.

A

There's a the multiple, listen events fail silently, and that was issue a 419. That's.

F

Back on the agenda for this week, right.

A

And I think that's back on, because it's December major that we want to land on v7. If I'm not mistaken, I guess we'll find out when we get there.

J

A

And then the and then scheduling meetings which was resolved because we're having a meeting at a very different time right now so try and try this out for one month and then we're going to review and see see how it works.

A

Okay, then, there's the agenda. For this week we have one two, three four items: there's expanding the use of the CT CSU tracker. There's adding this there's the support of platforms list that we talked about already. There's the multiple listen to them feel that we talked about recently last week and there's also a for discussion issue. Consider folding the tsc into the CTC. Are there any other items that need to be added to the agenda? I think some people might have mentioned some things in the issue: tracker that maybe could be added I.

A

Think, let's see Myles suggest that the download page layout um we can. We can see how we feel about that when we get there. ah The let's see here. That's that's! That's no DJ! No GF / nodejs work! Number 95, 34 people listening um okay. So if it's okay with everybody, let's go to the first item on the agenda, which is node, pull request, 8 9 45, which is.

A

Proposal to use the issue tracker more or and sort of two for decision making, rather than meetings, and it's also sort to codify what the process is and last time I looked six members of the CTC give it a thumbs up. Jeremiah had a question about whether 72 hours was long enough for the minimum time. An issue should be left, open and I. Don't think anybody else had commented. So it's pretty short, the did the main the main.

A

The main part of the diff is at the bottom, where it's just an added paragraph on some bullet points and I believe they're, mostly things that we've talked about in in the scheduling issue. So it should hopefully be familiar if we can get consensus on that, or even a majority on that today. uh That would be great I'm. A little I I wanted on the CTC agenda, because I'm a little bit reluctant to change the way decision-making is done without putting that proposal through the current decision-making process.

A

A

If anybody have any comments on it at all.

D

I think the CTC review label would be beneficial because there's so many issues on the issue tracker. Now it's really hard to keep up on what we actually need to be reviewing.

D

But I don't know, I don't think it's directly related here, so we could definitely on that or whatever is that.

E

Is the review of a label like something that we need to review like a somber major that doesn't need to come to a meeting? Is that the idea.

D

From reading it, I think that that's part of it well, no, it's it's stuff that doesn't have to come to the meeting exactly yeah.

A

I think you pretty much. Have it right, Jeremiah that I think James I believe proposed it as a way to label issues that the CTC needs to reach consensus on, but that, hopefully we don't have to bring it to a meeting that, hopefully we can just do it in the issue. Tracker right.

E

A

I still propose I mean I. Think of the CCC review label would be is helpful and we should add that, but I still propose opening a separate issue in the CTC issue: tracker for two reasons: one so that people can watch ctc. Members can specifically watch that repo, because it's relatively low traffic and pay closer attention to it and also so that it's really really clear how long the issue has been open, whereas like, if you add or remove a label, it kind of you know can get confusing.

E

So it is there something with like we need to decide in this meeting about this well.

A

I think it would be useful to to begin this process. Yeah I was hoping. I was hoping to reach consensus in this meeting that we should proceed with this.

F

Right, that's basically just pushing more issues into this region to ctc repos opposed to the general always.

E

Say I guess this is a governance change, so it requires a vote.

A

I yeah, you know, I I, didn't look through. The I was trying to find somewhere in the documents where we're in our current governance docks. Where would indicate that this requires a boat and I couldn't find anything, but my gut says that we really ought to have a majority in favor that we ought to you know so so there's currently six.

A

So if we can get four more than will have the the ironclad, you know fifty percent plus one arm, but yeah yeah I, think I think I think not having a vote doesn't pass the smell test for me. I think we should. We should have you know ten identifiable individuals who said: let's do this.

F

So do we have for people who haven't already commented in the meeting here and.

F

E

I will note I, don't I, don't necessarily think it's really clear, but I think that should probably not apply to governance changes.

A

A

So add something so you you would like to see something at it. That says that, if we're changing our bylaws or whatever that we that we, that that that that needs to that, that needs to go to a meeting or that or that needs to have a vote. At least.

E

Yeah, it's not I, just don't I, don't think I'll wait visit. Maybe it's just because I'm not reading the entire document.

J

This might be slightly out of tough outer topic. um We have a section called consensus seeking process. I have to say that is kind of mistake, abell, because it is linked to a sec. It is linked to a section where it talks about the pull requests and I mean. Let me rip that out in key in the case of pull request proposed by an existing collaborator, an additional operator is required for signup.

J

Consensus should be sought in short, if additional collaborators participate, and there is disagreement, if occasion so just next to this, we have linked consensus, leaking process and consensus seeking passage talks only about CTC members.

J

We may have to change that a little bit yeah.

A

J

Very clear to me: yeah.

A

I agree that the whole thing about that that whole part of the doc that it, which I.

I

A

Like two sentences in a link to the consensus seeking process but I agree that that aspect of the document is could use a rewrite III do I do think that might be outside the scope of what what we're trying to accomplish right here right now. I agree that it's confusing, though, and need some work.

A

A

Alright, so it sounds like we don't really have a whole lot of opinions, one way or the other going on here on this right now, but I would encourage everybody who has not weighed in on this topic to get over to the issue tracker because and and and give it a read and weigh in on it, because it's kind, it's kind of important. It's it's a fundamental change to the way we make to the way we we we try to resolve issues that get elevated to the CTC and I.

A

Think it's a good change, but um it's kind of a big one so lets you know your your opinion is wanted with that. Unless somebody has something to add, we'll move on I'll give a few seconds of silence in case someone something they want to say, quick.

F

Question is there anywhere else that talks about like quorum rules and anything else that this would be counter to that? We would also have to update so.

A

Quorum only applies to two things: to meet it into two meeting attendance and two votes, and this is on this is for things that do not require votes. This is for things that only require that merely require consensus.

F

Okay and what's the distinction between like what requires a vote versus consensus, mostly.

A

What requires a vote is is something that can that can't reach consensus. ah Okay,.

F

Oh baby, yeah, okay, the.

B

Premier, media separated for that to see the moment when the issue was escalated to the bottom, so that probably means three tags. Ctc reviews, dis agenda, NCTC wall.

A

Probably a good idea.

A

E

I think we can move on okay.

A

I'm happy okay, this sounds like it's reached. Consensus but, like I said, really would be nice to have a majority so and we'll leave it open and on the agenda. uh Moving along. Adding some whips is this yeah adding supported platforms, rod actually removed the CTC agenda label from this and I.

A

Don't know that that we need to discuss it here of people want to take a quick look at the pull request, but I think we talked about it last week and there's a PR open for discussion. If anyone has anything to go, this is um yeah. This is the link is to the issue, but there's a UH gosh, oh I'm, sorry, okay, I'm completely mistaken. This issue still hell no I'm. Looking at the wrong issue: okay, I'm very confused today, I'm sorry uh trying one more time: yeah rod removed the tag. It's a poor request.

A

You can read it and comment on it. There um and there's already been a lot of discussion um in the gold working group about it before it came to the CTC. uh So probably questions and whatnot are probably best left in the issue tracker so that build working group members can answer them, um but if anybody has any discussion or comments, I want to make about it here and now I'll again, you have a few seconds of silence. If someone wants to talk about this or if they feel I framed it in properly so.

F

I agree it. This input was given last week, so just didn't remove from the agenda.

A

Okay, great so, let's say that that issue was already resolved and did not need further discussion. Let's move on to multiple listing events fail silently um and I believe James Tague. This. Let me see who tagged this I'm. Sorry I'ma tagged it, but no James tagged at ccc agenda. I'm! Sorry, what's James and his comment was as a semper major can't get into version 7 without ctc reviewing okay and he said we can try adding it to the ctc agenda. So um does anybody it looks like Jeremiah's active in this?

A

Is anybody want to frame the issue if anybody understands what exactly the CPC needs to consider for this.

E

Well, we did discuss this in length last week. The idea is that calling listen twice on a server currently causes like under the hood problems, and we should probably just throw an error if you do that and I believe that's what this does. So, if you throw an error, sorry, if you listen a second time without first having closed the server it will air you'll have errors I under the hood as it is as it stands today that just won't be as visible in the same way.

E

um So the question is really I'm, just a technical one, if if this covers all the places that a server can get closed by, so I don't- I don't really know, there's much to to talk about unless someone objects to this landing in v7 keep in mind. While this does introduce a new throw, you will have under the hood problems without it.

F

Right, I think it's just yeah. That's the real question to answer here is a.

K

F

At the last minute, should it go in.

K

There's an issue with the pull request. It only sets the faults if it's closed, but it should be possible to listen a second time if the initial buying failed.

E

That sounds like a technical thing that can be resolved, though okay.

F

That does that does I think if it's that there's risk of getting it wrong. You know it sort of says: should you be more caution and wait till the next release versus how to get in now right and.

K

It is difficult to get right. Somebody already attempted this before I posted this quite a while ago as a new contributor issue, and they gave a good attempt and the the issue of replacing honest were handle was fixed. That's but now there's an issue where, if you call listen with the hostname and then call us in again without a hostname, it immediately binds to the latter port and not the previous port, because it the one with the host thing, needs to do an asynchronous resolve on the hostname.

K

So yeah there's a lot of weird little quirks that can cause listen to behave improperly. That.

E

Sounds like it's actually like, like what you're describing sounds like buggy behavior. Yes,.

K

It is so that buggy behavior was introduced because of a previous fixed to not replace underscore handle. My point: is it's easy to screw things up right in an attempt to fix this and I can see from this PR right here? It does have a bug as well, so.

A

Trevor would it would it be safe to interpret your comments as you don't think, this specific change should go into v7, as is.

E

I'm not sure if that questions useful like should we try to get what this is aiming to do in vcenter right.

F

That's I think the question we should answer is like is: is there a rush and value in trying to get this in or should it just wait and take the natural course.

K

I

K

You I feel, like the behavior, is edge kc enough that the risk of messing with people, the ecosystem, isn't that great, so yeah. If we can get the if we can get the bugs fixed cool, but we should also know it's like okay. Well now we know what screwed up today, and so we can document that. But if we try to fix it a last second, we very well might screw something up and either we have to document that later or we have to slightly break edge case stuff after sevens been released right.

K

So we have to choose which path: we're okay, with um I'm, okay with either, but is.

F

There somebody saying we really need to get it in like with the process. May we want to put it in so.

A

In the issue, tracker on ax and James indicate that they are +1 on landing it in be seven, but I don't know that they really lay out. You know what that means is like. Yes, I really want this v7 or yeah I'm, okay, with it landing and b7, and unfortunately, neither than we're here.

F

And what I think we go ahead? I think.

E

I'd prefer, if it landed like Sudan later than in general, because it's a it's a situation where you have you can do this, but you'll have settled problems that are not visible to you necessarily.

E

We're not immediately apparent, but.

K

The point is in at in previous attempts to fix this other subtle bugs have been introduced right, I, understand that all right.

F

Yeah, like it I mean yes, we'd want to fix it, but yeah I do that risk. Would it be better to say okay, we'll put it in. You know in seven I guess this is going into seven right, which has six months to sort of shake things out and even longer before LTS.

C

Well, seven will never hit LTS Yeah right, I, don't think she just let it you know, take its course and fix it when it's actually fixed.

D

That's how I feel too.

E

So does that mean you feel we should just wait until the 8th I.

C

Feel, if that's, if that's what it takes, then yes I mean this is a pretty edge case thing. I, think all right.

E

Yeah, it's a little bit awkward that we don't have either the people that are really supporting it here. So we probably if you, if you feel like that we shouldn't land it. You should probably take your comments to that thread.

A

Yeah and and and the discussion we're having is about landing it in v7, not necessarily when it lands and master. In my understanding that correctly, yes,.

E

It will land in master when it lands normally.

E

A

And okay, so I would I would I propose that we encourage people who have opinions, Trevor, Colin Evan, anybody else to to post into that issue, and I would like to just that. We stick a pin in this on a might show up during this meeting where she might not.

A

If she does, we can reanimate the conversation and and and and maybe get a little further with it, but otherwise the issue tracker seems like the way to go, but as usual, I'm going to leave a few seconds in case, somebody has something they want to say and doesn't want to move on. Quite yet,.

A

A

Okay, then, if I'm not mistaken, the next item on the agenda is kind of a discussion item that rod, I think, put on the teeth. The tsc has already discussed it.

A

I have not listened to that tsc meeting, so I don't know what was said, but uh it's the in the tsc repo issue, 146 consider folding the tsc into the CTC I've expressed my opinion in the issue tracker, but I guess I, don't know if anybody else wants to try to frame this a little better to try to get at what rod wants to know or if we just want to throw the discussion open for comments from whoever wants to talk about it. I can.

F

Add a little bit I mean, I think, from the tsc meeting. A lot of the discussion was around the pros and cons of you know kind of two things, one having two separate meetings for technical versus, not strictly technical issues.

F

You know I, think there's some we've seen some benefits and sort of separating that discussion out, but on the other side, is the sort of complicating factors like how do you have who's the member of those two different groups? If you make it the say, no, so one contone one way of addressing it that was discussed was like well, let's have one group which is the tsc, but then it's sort of a subgroup that has the non-technical meeting.

F

Then you get into discussions around well, how do you get quorum to make decisions and those kinds of things so it's kind of it was it was around that you know: does the separation between the two groups still make sense, both in terms of membership and then and and then separately? In terms of you know, the people when you get together in meetings themselves.

E

Yeah I think it would only really work. If then, the CDC, like delegated pretty much like made a team that had responsibility and like authority to like sign off on that stuff without having to have CTC boats pretty much ever.

F

Certainly, from the quorum side of things yeah, we wouldn't have quorum in the tsc side generally from you know, current attendance.

A

So looking at who's in this meeting right here right now, if I'm not mistaken- and I might be- but if I'm not mistaken, the only people here who are not on the tsc arse a capri on Nikita and me I'm.

H

Not on the deviated Oh an.

A

F

Yeah and I'm actually only an observer, but ok.

A

Ok, so it's a I just wanted to list out those people in case any of us have any particularly strong opinions that we wish to express, which might not have been expressed at a TSE meeting.

B

And don't have any specific any specific opinion on this, but the arguments that are mentioned- Leticia are understandable. I just don't have anything else to add.

D

I'm honestly and I prettier pond.

A

Second pair energy or something you want to say.

J

I never looked into the responsibilities of DSC merge, but when I try to explain what the node.js foundation governance structure is the question which people generally ask me is: why do you guys have two different deciding committees, I mean yeah, I mean I have to explore more what TAC does but yeah. This is the question which I, which I get commonly some people.

A

A

Okay- and it looks like oh good, someone was taking notes, excellent Michael. Thank you great fear.

F

A

Cool okay, awesome, uh okay, anything else. Anybody wants to add to this.

A

Okay, I think we have two more issues before the QA time.

A

Someone added the general v700 v6 LTS planning discussion was that Jeremiah or somebody yeah.

E

That was me and.

A

Yet good we'll do that one first guy.

E

So we're now I think two weeks out from v7 um made release James is doing that who is not here and one week out or actually pretty much like six days told the scheduled v6, LTS, release and rod is handling that I believe not sure how much there's the comments on my end. But if anyone else has comments and then this week's current v6 release should be I'm out today,.

D

Related to v7 going out um I opened a PR for headers a better invalid. It doesn't tell you which header it is, and so it fixes the error message. So I know it's ember major, but would anybody I probably should I should add notes, but does anybody like opposed to getting that into v7.

E

For that somewhere, yeah.

D

Let me pull it up on some changes.

F

The message from something less specific to more specific, yes,.

D

It tells you what header it is. Okay are.

E

We talking about like uninstall.

D

Now this is in HTTP, oh.

D

Wait paste it in here I'm, sorry, I, just paste it into the uber conference chat.

A

For ball, listening in who might want to follow it's in the nodejs / node repo, its pull request, 9010.

A

A

A

Yeah III don't know jack.

F

Yeah, what's the likelihood that anybody's any scripting or tools are using that.

A

There's always a chance that somebody is is parsing error messages, but I mean, as last minute, semper majors go I'm; okay with it.

E

Yeah so like parsing errors is definitely a thing, because you can't tell what it is otherwise. So, like people do do that.

F

Worrying, if the only mitigation might be, is this instead of the change? That's there. Are you left the existing message exactly the same and added in like an addition at the end, I don't know if that makes it less likely of racing people or not.

E

Nikita I'm gonna I'm going to ask you in the issue that would be. Do you think you could search for the the error message.

B

mmm For the exaggerated, a message, I think yes much, let's move that Felicia could yell just mention me. Yeah.

A

Okay, so I don't know, maybe if we had a ctc review label, I'd, ask you to add it to that hit to the floor, request everything I.

E

Like my stance on, it would be yeah. Okay, if not everything in the world is depending on it.

D

I got another alternative would be to add a property to the air that has the header name. I, don't feel like that that could help solve it. I would.

E

Prefer to just wait for James's errors, rewrite to yeah get in otherwise we're just going to add properties here and there and people start depending on them at some point, and it will be a mesophilic gas.

D

A

Yeah definitely prefer change if we have to choose between changing error message and adding a property. Definitely prefer changing the error message for the reasons Jeremih stated.

A

Okay, anybody else have anything they want to sound that before we move to the issue, miles added a mild suggested that it might be a good idea to talk about note here, I'm going to paste it into the issue track into the uber conference.

A

G

A

Discuss this here, I think we just want to make a note of it that this is might be kind of important. Some of us so be aware of it that, basically now, since we're going to have two versions in LTS, v4 and b6 for a while, um it's going to change the layout of the no just a torque page and you might care what it looks like and how easy or hard it is to get to the LTS version that you know this or that so be aware of this issue.

A

Does anybody have anything they want to exhibit a actually want to discuss this at this time? Sorry.

F

Which issue is that look? It's.

A

It's in the issue: it's in the the chat sidebar, it's okay.

F

A

Know just /, no guess: that'll work, / issues, / 9, 53,.

A

Okay cool! If and if miles wants to put on the agenda for next week, we can do that too. I guess but I'm. Otherwise, oh yeah got Gemma. It.

E

Will need to be live before our meeting next week right.

A

But true, but it also, but it doesn't mean it doesn't mean that revisions can't be made after its yeah birds live yeah.

F

The other thing yeah I, want to make sure we get the links updated for the pic and Linux one downloads as well. Oh.

A

Yeah yeah, I think I saw that poker good good.

F

My gums are but I guess it's related to this as well.

A

Okay, so encourage everybody to look at that issue and get your opinions in and make if you think it doesn't, if you think it's irrelevant to you, read it and at least make sure it's actually irrelevant to you.

A

If no one has anything to add it's it's time for QA from from public channels,.

E

Yeah so we'll open up QA on the YouTube shot and in no dev IRC on freenode. um So just as a note, I'm going to read off what our rod put into the meeting issue. So he'd like the CTC to be aware that we're migrating to a new noches, org, I, O'jays, org, Libby, Vorg, etc, back-end server, um some of them are fronted by cloudflare and we've already moved everything, but no just like org to a new host, um so we're planning to do that tomorrow or that might be today for some people.

E

So if you could, please watch for any noise about things breaking and let Johann know about any anything that you hear. That would be great there's a build issue for that to build issue. 5 16.

E

E

E

We have six people watching apparently.

E

Again, we're just doing Q&A, anyone has questions, put them into the YouTube shock and to.

A

Fill the dead air while we're waiting I'll also just mention that there is a tsc meeting tomorrow at I, guess it would be 8 p.m. UTC, 1 p.m. pacific time, if you're in if you're North America the next ctc meeting will be on one week from today about four hours later than this meeting started there. Let's see that's in the past, so that can't read that meeting off there will be a Diagnostics workgroup meeting the first week of November but I, don't yet know the date or the time.

A

There's an LTS working group meeting on the seventeenth of this month and that.

E

Is next week, yes,.

A

And that's all that is currently in the minutes doc, but that's not necessarily all that's happening. Alright,.

E

So I don't see any questions. Someone says hi, nothing else. All.

A

Right well, at least, we know someone someone's actually listening and not just like streaming with the sound off okay, if that's the case, I think we're going to close out the public part of the meeting last last chance for anybody to raise anything. Otherwise, thanks and see you next week, four hours later, I have some private business and others might too. So if you want to hang on the call, that'd be great.