Node.js Core Technical Committee, 3 Aug 2016

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: 2016-08-03 Node.js Foundation Core Technical Committee (CTC) Meeting

Description

https://github.com/nodejs/node/issues/7948

A

Part is recorded, but it is not broadcast. Okay,.

B

C

B

Rod will edit that out of the record right.

D

But if we like, if you are setting up the broadcast now well,.

A

Whatever sorry, hey basic yeah, the private parts at the end, okay, we're into the public part first, and so basically all that means is, if you want to insult rod, don't do it.

E

Right I'd say this was live. Okay,.

A

Cool we're broadcasting excellent, ah let's see here, welcome to know JSC TC foundation, meeting for august 3rd 2016 and we're gonna get right to it with the stand up. So, let's start with honor.

F

Hi, just the usual issues, NPR review, that's it and next is Bradley.

G

Sure so I went to tc39 last week. Most of what we got back was we need to take our idea of modules in a different direction. A lot of things we thought could work are not going to so we're trying to figure out what that means. Now. Yep.

D

With with we hear, is that the tc39, or is that.

G

Ya know it has to figure out how to make it work.

G

So when I say we are more talking about, node has to figure out how to make it work.

A

Okay, that sounds like something that'll be on an agenda at some point soon, I have to imagine. Okay, um next up is uh been.

H

Nothing special to reports, I, don't know who's next, but it'd.

A

Be nikita nikita is next it's. He says the chat is having mic issues or are they resolved.

I

Can you see me yes, oh okay, good ah so I'm working on the internet search the new me to the one that boosts their data set in that a use for grapes and some comments on initial sample quickest as usual or nothing else.

I

Next, tues, crease or goin.

J

I was on vacation last week, Chris.

B

Was just reviewing issues in four-plus next.

J

Is Evan yeah I've done a little work, cherry-picking septa v6 and then I've been working on giving commit validator run for PRS with some of the build team. Next is Jeremiah hello.

E

I was mostly late last week than not I've really just been doing. Some experimentation with creation, X's nucleus, J us and that's pretty much it fodor's next, two piece here: no thank you that would make James next.

C

um Note summit last week in your exploration of the possibility of an HTTP to implementation in core has an interesting expressions are on that sanity. Violation of the weather, bureau implementation, some foundation, ET SES miscellaneous book and reviewing PR, is catching up still on vacation named josh jana.

K

Nothing exciting year and mostly internal stuff and bit of vacation next is Michael.

L

Okay, so the way it notes from it last week so catching up on issues after getting back I'm, starting to add the Linux one release machines and jobs for Knightley's I just got access to some. The first of three new ax machines from 0 cos else is starting to configure thats, but we can add it to the CIA and landed a few PRS four minutes from past meetings.

L

Next is Michael Rogers.

C

Leave Michael it here, no I.

A

Don't think he is so I think Brian is next.

C

Thank you talking here, I'm.

H

Right step out so.

A

H

A

Right Brian just wrote, commenting on issues and PRS next tues alley so.

M

I was busy with the note summit and in internal issues most last week. Besides, that I've been working on shepherding some backboards and I also started working on writing up a proposal on how to manage v8 LTS branches, the saturation of a discussion, eid that she is working group I think next is Burt.

D

Hello yeah, you never see me anymore. I, coincidentally, commented on eight on an issue with that. So I did I'm mostly here to announce that I'm going to step down from the CTC and I was drawn here by a discussion about real path issues that I am interested in.

A

The next I don't know next is miles who is not here so that makes it me. I've been updating, CTC governance documentation uh in trying to schedule some onboarding. One person had to postpone trying to onboard franziska next arm and welcoming additional nominees for onboarding fixed, a flaky tests investigating some other flaky tests, and that's all I've got. Did we miss anybody who's on the conference? Oh.

N

Yeah I'm here it was dragged by the real path issue, but they don't really have anything to add at the moment. I think everything is on the issues I'm sort of around to clarify whatever is needed.

A

Okay, cool that looks like everybody, so, let's, let's get to it, so lets us I yeah. Let's start with the nomination of for adding on ax to the see this is issue number 7, 6, 07, I, don't think, there's well I nodes. Is there any discussion we need to have, or should we just go right to the eyes and nays and what not.

H

A

The addiction Philip go okay, all in favor, please say: aye aye, aye,.

B

A

Hey all opposed, say, nay, please.

A

All abstaining, please indicate.

A

Okay, onna welcome and congratulations by unanimous ascent vote, whatever it is you're on the core technical committee, hi.

F

D

A

Think, oh I'm, sorry you're trying to say something: I'm.

F

Sorry that was just my microphone. Okay,.

A

Yeah I'm, pretty sure rod will add you to the appropriate mailing list or lists and probably send you other information, but I, don't think anything that will be a surprise, so I'm very, very happy about this and but I think we should move on to the next item. If that's okay with everybody, which is reverting, FS changes, this is issue 7, 8, 46,.

A

A

Where this was left at the end of the last meeting was there was a bunch of conversation. I was going to be moved back to github, and it was this.

H

The change that makes omitting the call back and ever what is it, I'm trying to say an immediate error right.

A

And I think there are two PRS in to revert it, and I believe that I don't know that. There's anything. Is anybody able to frame this I, don't know, there's anything to talk about, except that, like what do we need to do to get those pr's to land.

F

So yeah we talked last week about 7 8 46, the one that is on the agenda: mild split off one of the commits into the enter 7950, because that was requested here on the call last week. I, don't really think its controversial or anything yes that! Well it's what Ben said like it's, the one that made FS thought something without a callback for an immediate error. Those two rivers are necessarily necessary because one of them builds upon the other and yeah that they are holding up the real path we word because they otherwise would conflict.

F

That's everything I think.

A

It appears to me that both PRS have been open long enough and have LBT ms and I don't think, there's anything stopping any of this at this point. They might need a CI run or something, but that's I, don't even think it needs that I I.

C

Would like to see a CI run, a fit to run and just just some additional testing to make sure that we have the right exactly the right set the river, because right now, man, we've had we've had a number of different conversations that I think kind of are bleeding into one another that all relate to FS. Michael I just want to make sure that we are very clear on what for voters.

A

So normally I think miles would probably be the person who would kick off those canary in the gold mine runs but he's you know, I hope at cascadia Fest this week and then flying across the continent any chance you or someone else might have the bandwidth to tackle that yeah.

C

I think I could get off the line, yeah picking.

E

Up stitches such an runs is pretty easy yeah. It's like running the normalcy I, almost I.

A

Did it once and I thought it was easy and then it blew up, but it turned it yeah so anyway, yeah the errors from it sometimes can be hard to read, yeah so cool all right ah great. So so so James you'll uh you'll pick up the ball and push the CI end of it forward and it all looks good landed and get to solve. The agenda is that the plan at least yep.

H

What about making it Prince deprecation warning, because I believed it was a plan, or at least so it mentioned somewhere, I.

F

Think that's pretty much agreed on that. That will be the plan after the rivers have landed um yeah. Oh.

H

Goodness, someone take no responsibility for that. I.

F

Think that is what the fourth I sector Briana's planning to do: okay, okay,.

H

C

It is that going to be a process warning or a deprecation learning, Rousseff.

H

C

Going to be a plus this morning or a deprecation warning, I.

H

Think you mean a warning: it prints at run time when you were you call it what, without a call back.

C

Right right, so yeah we have I just want to. You know just double check our we're going to do an actual deprecation warning or not like with you till print duplication. One.

E

E

Are we talking about reverting this and then issuing a deprecation warning that it's going to be coming again? It.

H

Was plentiful and it makes sense because.

O

H

Has been well not exactly deprecated, but at least discouraged for four years.

A

So James do you have a strong opinion about one or the other.

C

No strong opinion about one of the other, but deprecation warnings have always been cember major, a process that warning would be a sunburn, minor or even a patch. So.

A

Why don't we let why don't we let the one will that set the prion submit something and we can hash it out on github, so.

F

Yeah, just so that's clear, like he already opened and PR to print a warning instead, but like that would probably work instead of just reverting but like we agreed on reverting last week and it's holding up the real path. Pr. So I still think that's a good idea. But if anyone wants to have a look, that's PR, seven, eight, nine seven, which is just forewarning and I, think it's a duplication, warning yeah. It is I.

E

H

Think look. Oh.

E

H

E

I, don't think um unbreaking something that was already going to be broken in a major and then moving it to a deprecation is really somber major worthy.

F

I know that someone on some of the issues around this a while ago commented that they actually like to do that sometimes calling FS functions without callbacks just like, because they don't care if an error is thrown I, don't know if that makes any difference, but I prefer to be careful. I didn't see any reason not to do this in assembler major and as.

A

For whether or not a deprecation wearing a semper, Major James just submitted a PR outlining explicitly our deprecation policy and according to it that will cember measure right James, and so so we should so so if you feel that deprecating it should not be so. Oh, but you're you're in your time.

E

Talking about I'm breaking it, it was already broken in a major we could have deprecated in the last major, but we didn't cuz. We well ya,.

F

D

E

Effect of that point screwed up, so why not just change it to a deprecation it like. We already had the major for that effectively. That's what I'm wondering okay.

P

E

Mean that might be a little bit weird logic, but I'm looking for my.

O

C

O

C

Going ahead and making a deprecation since we changed it with a major or fixing it, but it's gonna change it to a major deprecation. Those.

A

C

A

Would like to ask that, since we do have a full agenda, that you know, we.

L

All agree that we should.

A

Revert and revert and then the deprecation warning whether or not thats ember major or not, that can be discussed on the issue and github. If that's okay, and if there's no consensus on that, then we can talk about. Then someone can slap. A ctc meeting tag on that issue is that, okay, with everybody yep.

A

Okay, if that's the case, then let's talk about real path, which so last week we talked about real path and the this some the end of it was that on a Trevor or Alexis would move it forward, and we also saw will hear today. So my understanding, based on the conversation we just had, is that basically don't think all the up at this point is the two reverts that need to happen that we just talked about. Is there anything else that needs to be talked about with real path? I.

F

Don't think so I have.

N

One tiny thing I mentioned in the current PR that reverts it, which is the fact that it adds one test for a condition that it might not be correct depending how you look at it. It's not dig is it's about the substitute drives on windows. So, basically, if you have long path and you substitute zet the desert unit, the new real part would actually result in to the right one, one that points to the real deal, but the one in the jas version gives you like jet colon, whatever file.

N

Some people have apparently relied on this for shortening paths and I, guess I'm. Okay, with that, the question is, if we add a test for it, does it mean that if we ever find a way to make the sea version fast and good, but it works the other way it's incorrect or not.

D

So I mean the question: here would be what the rights of Magic's I know exactly clean slate. What would be the right semantics? In my opinion, it should resolve subsidized to their original place and I.

D

Don't think people are relying on the fact that real bath won't like they use they use subs as a tool to make sure path in general that won't be broken by this, but the whole reason we do real path in the first place is to avoid a module that is accessible from like through different paths, thanks being loaded multiple times, and so you know, if you imagine, that you know, d might be substitue, see who and like you're loading the module from both Sifu and from deed, and you will still get some coffee, so the goal would be really to to resolve to the really the.

N

Final path, I agree with that, but the other thing was brought up by someone, as in you go beyond the short path length on windows. Some tool gave some problems, I can't remember what and where so we're like substantive proof to a unit. So you will get like a very short, but that's.

D

True, but like what you will do in no Jas I, real, fast you're, probably going to use it the path in nodejs and then the path length limitation doesn't really apply. Although I agree, the path length limitation is 64 K, so yeah yeah. It's long enough for pretty much all everything yeah.

N

D

N

Point is that, right now there is a test that checks for this and I. Don't think the test is correct.

E

If you mean not necessarily correct, if we resolve its 1 resolve it somewhere else in the future, that kind of comes to the deprecation policy and breaking changes, policies that we have already. So if we message that we're changing real path- and that comes with it, that may or may not, may not be a problem, it could still be a problem. It's hard to know the.

N

Thing is: it has been added as a sort of regression test in the sense that well, if we add real path again coming from leave, you be, for example, let's just make sure that these passes, because otherwise wind up being enchanted territory and, unfortunately, all the conditions that were broken with it are weird edge cases that are very hard to test for, but this one, which is easy to test for it's actually not the nature.

N

In my opinion, which bird seems to share Oh testing, for example, I am bisque Ram discs, which apparently more use than I thought is Anna hard to test. You rely on the machine having some software installed in some unit mapped, but we can't test for that. Let's say, and that is broken we know, but this one can be tested for her and it's not broken depending how you look at.

F

Just one thing for clarification: reverting to the effort, jazz, realpath implementation would mean restoring the behavior from free version. Six right, yes, even worth respecting, so I.

F

Don't think it's worth debating that right now, possibly because, like we pretty much agreed that the revert should happen, and it should happen quickly because it has been hanging around for a long time and yeah yeah.

N

F

N

The test doesn't need to be part of the revert, for instance, it wasn't there. In the first place it tries to test what the new code did. That is now deemed incorrect, and what I argue is that if we land that and then tomorrow, I come with an implementation that works, but that's not pass this test I'm going to get shouted at because oh that's impossible and I will say that that's this incorrect and then so. He ok.

M

N

F

Just just discuss that later, when, when that new implementation is there I guess.

A

F

A

Yeah, another option is just to submit a pull request to remove that test, or just separate of everything else just make the case in a pull request that you know the test is invalid. Here's why and I'd like to remove it. Our.

F

Move it to know Nicias test, if you prefer so.

D

Just for my understanding, so the old implementation would not river resolve substituted rice, but the new one did right. Yes, that's the case, but I think I like it is desirable that, like an ideal future implementation does resolve them. Those like you. We will have to revert the test now because it will not past pass whip like when we go back to the old implementations. Now.

N

But ideally it will pass because the old implementation doesn't resolve it. It will say that the file is in z, not that this in see users, bird programs right.

D

Ok that case, you should probably just remove the test. Thank you because it's like it tests for undesirable behavior, yeah.

N

So the test is part of the revert PR, so I'm just saying we could draw that commit and it doesn't delay the process of landing everything I would.

D

I would argue like if the in decades leave the test in because at least it helps catch changes in behavior, and that means that if we, if you change the behavior further in the future future, you will know you know you will be deliberate about it and change the test as well. I agree.

E

With birther, um can we take that back to issue yeah.

N

Sure um yes up right up, but that's all I guess so I, don't.

E

Think the intention is definitely not to make to revert to like not even necessarily to revert to a better implementation, but the revert to an implementation. The ecosystem has been functioning better on, so we should make move towards making more incremental changes if possible, then to make it actually better degree.

A

Okay, so it sounds like there's agreement on the path forward here. Is there anything else to discuss about real path.

D

What I would recommend in the future is to try to restore the like this staff this this thing that June I was getting referred it but start with UNIX only because it's very simple and you know, will probably work there.

E

um Just a question: are we reverting to where the CAF cash argument still exists?.

F

No and that's being removed in to revert PR and reply study, encoding option I have was a concern about that. Be there because, like I think it will require that all intermediate paths that may be encountered when resolving siblings must have the same encoding or all utf-8 or something but like I, think that's just something one has to take now.

E

mmm That doesn't sound like a full riverton, okay, it'll say it's.

C

A revert of the internal implementation, not the API, came not the signature change. Last.

E

I heard looking.

C

Around just the.

E

The new perimeter that was added would still not break the counter argument and it was better to take to work with the cash argument due to performance reasons that were removed because of the internal change that may be in yeah.

F

If I understood the discussion from last week right, we did not agree on that because, like I think what Trevor said it was that I'm, not including the cash right now, would give more freedom in the future to do to do other kinds of performance improvements like a hidden cache in lip UV or something.

C

Or even bringing back the new implementation should be very issues. Would it be resolved.

A

Then had a question about the fs revert since we've gone back to talking about that. Oh.

H

Yeah, so I was wondering I'm a little confused. These are change, is targeted at fee; seven, not fee six right. So the next.

L

H

Maker, okay, yeah yeah, then I don't see an issue with refloating in and putting in a duplication warning because hey it's going to be a major anyway! So.

E

It will be landing in v6, though right.

H

You may be duplicating ordered, recruits or.

E

All of it pervert all.

C

Of it, the idea is to get this revert in for v6 before it goes to LPF.

F

Did the changes from that? We're inverted actually make it into v6 because they were attacked several major too yeah.

H

Yeah exactly yeah. That was why I was confused because I don't think they were able leonard in physics, I think so too.

F

Yeah they didn't make it into g6 I. Think.

D

We changin talking about the real bath change and we're now back talking about Davy previous topic, the fs cool, back change, yeah.

C

And this is actually part of it part part of the challenge. The conversation is too many things going wrong at us, but we need to make sure stay clear. The real path, change that will be reverted is in v6 and that change that revert would land in v6. The other changes regarding the eighth throwing the errors on the callbacks. That is only in master that would not land in v6.

E

I thought you said that the real past changes were dependent on that they can't be dependent on that. Then. Now the real path changes occurred before the white revert, the other one of them.

C

That they're, having other issues.

F

If the river for real path were to take place on the v6 that X branch, there would be no conflict, but like the summer measures that are to be reverted by miles PR, they would conflict with reverting the real path to change. On.

A

F

E

A

The change was okay changes. The change was back, poured it right to well. Okay, yeah I'm, gonna stop! Actually the talks. If.

C

We do eight like a back court room type of revert where it's not just a straight-up reverting to commit right. Then we can do it. You could possibly do it without it. The modifications, but because of the changes that have occurred, we can't back porch. We can't simply revert to commit to a real path without also referring the other changes that sense on.

B

Master a master, oh, the teams are.

G

Great, wouldn't.

D

Be possible then, to revert a fast River, real bath and then reapply fast like and then picture the conflicts. Yeah, that's against a yes! It's.

C

Just more work, you know, and that's right, that's why I'm kind of concerned here that I'm not sure if we have all of the steps kind of lined up yet and they're all on the same page with those steps are like we're all on the same page on what needs to happen. But I'd.

A

Like to hear Anna say what she was going to say and then I think Nikita might have something to say.

F

Yeah I think the path was actually kind of clear. Like you for first two miles River PR, then we do the real path we work and then probably sectarian will have to reapply his semper major dressing, changes that were previously reverted and turned that into the deprecation warning. Yeah.

A

I agree that it seems pretty straightforward I.

I

Wondered that the comments that we needa watch it edit some regulated death. Well, they are related, but they won't be broken by divorcing those comments. So I don't see why we stood Josh, those test that I still work knew and when they actually decided orchard, could we keep them.

H

Not sure quite gotta.

A

Are you referring to cells previous comments.

I

I

The pull request that you watch some comments. Some previous comments, also reverse new tests that will introduce the net comments that are not necessarily related to the actual changes that were mentioned that comments. So could we alert the changes watch the tests that are broken with reversion to the actual changes, but keep the tests that are not broken with reverse the actual changes our I see.

D

Yeah, so it sounds like this is the same as the the test. Sal was alluding to, and consensus was to keep the test. So I think we are in agreement with Nikita.

A

Is there, is there anything else to talk about with real path or FS or FS in real path?.

A

Okay, in that case, great, let's, let's hope carrying the gold mine goes smoothly and we get some quick reverts and life is beautiful. Again, let's move on to vie for 50 proposal. This is miles as he's not here, but he really wants people to test the rcs I, don't know if anybody else wants to add something, maybe James or someone.

L

The only x-ray info I have is, I know, we've been building them internally at IBM and haven't come across any problems yeah. I.

C

Was gonna say that sir, it was on you same thing, I've been doing some checking out, I have not seen anything opens, run awesome.

A

Hopefully, you've communicated that to miles, but if not I bet he'd love to hear it. He probably knows, though cuz he works with you guys.

A

All right, in that case, I'm going to move along to drop support for viet for visual studio 2013 in v7, based on the PR. That seems uncontroversial now that who's who's, who can frame lyst.

H

It's too bad joe, is not the call, but so I think the conclusion is that we can move away from PS 2013 without too much trouble yeah. I think everyone was in agreement that there is a good thing, so I don't know don't know if.

O

There's anything else to discuss. Thank you. It's. Why? I'm no longer a windows person.

D

But it seems that the alexis is also in favor. That's a big plus one. For me, the only question I have is what this means for a compiled add-ons and whether no chip already supports vs. What have we nowadays 16.

K

Yeah I talked with joe this josh. I talked about a little bit last weekend. He didn't find any problems, although we don't completely guarantee, if you compile with a later and that modules are compiled with 2013 the ABI compatibility, but everything that he tested was ok. He ran into such him and he added note jit touch to the CI system, so I don't think its problems. I.

C

K

C

A few people look at this and tested it and haven't had any issues with eight ABI issues in practice.

C

Doing a switch so yeah.

L

I and I think 47, it seems pretty safe 46. It might be a bit late to do that. Change.

D

Yeah and I way, since we also guarantee avi civility, it seems that a lot to change a compiler for an for an existing version, I think.

K

One of the reasons we were thinking about it was cause of the LTS that we wanted to be on 2015 4 LTS that we're not stuck with 2013 all the way through.

C

Yeah, it does the main argument for for aligning it in six.

D

Right, I would suggest to make this on a separate issue like the discussion here is clearly about v7 or like clearly I don't know that's what the issue says and I think we are an agreement that we can do it for v7 and for physics, yeah I, don't know I, it's not that I care very much, but yeah.

C

D

C

The complexity they're really just comes down to like a CI environment, but it doesn't seem controversial in any way with refers to be seven I think we get two slanted and master. We can do some additional tests between now and october and if it just looks like things would be, ok then I said you know maybe number time primary visit and decide if we're going to go ahead and do it in basis. Yeah.

L

It really comes down to how much community testing will will get your not last few months right.

P

P

C

There right now we're only good month away from that, so yeah.

L

Exactly it's just if there's not a lot of time, and once he gets landed, I want to be switching it like just a month before.

A

Okay, anything any other comments on this topic.

A

Okay, in that case, let's move on to don't inherit from object. Prototype I believe where we left this last was that Brian was going to seek tried to try to get Chris. Try to ask chris dickinson about his concerns and I pained Chris last night. I think, but he's I think he's traveling or something anyway. um Brian. Is there any news on this? Oh.

P

Not really, uh you know I last week, right after the meeting here, I had been Chris. Also and I guess we're just waiting on him. Then, whenever.

L

P

Gets time to I guess respond or whatever in that throat.

E

Do we remove that from the agenda them.

E

It doesn't sound like we necessarily needed on for next week.

P

I suppose so I don't yeah I guess until we hear from Chris we could probably just remove it for.

A

Now I mean so so: I just removed the the label from the from the issue. Just now, um I mean eyes: they're.

C

Living it from the agenda is fine and it's like yeah I. I I'd whether just not be an open question indefinitely, though yeah.

A

We need to be able, as I says they're like can we can? We, like you know, is there? Is it? Is it problematic to you know, I mean well yeah go ahead games. You should finish. I'll stop I.

C

Just going to say it's been, it's been sitting there for four months. There's been plenty of discussion about it. We've heard from folks like Doug Wilson, where the Reds express who sees it as a positive change. He understands that. Okay, they'll need to be some stuff that the changes and Mike it could break some folks, but he's waiting, objectivity on it, I'd, rather not hold it up waiting for a response without any kind of indication that there may be coming any mention, so I mean I. We.

B

Have two months.

C

We've just over two months before b7 comes up and it's a major change. If we don't land it with the next two months, we have to wait six more months before it be released. Yeah.

A

I wouldn't object to landing it with the understanding that you know someone might you know if that makes someone panic and we need to revert it in the net. You know a week later: that's okay, but I, don't know if everybody else feels that way, though,.

M

I'm slightly a negative on this one I, don't think we should make assembler major change once and then another one in the future. I think we, let's figure out. If you are going to change this, it would be much better if you figure, if and how we can get the maps.

E

Or what, if we just did the map implementation, now explain this last time it needs a new API. If you do the map implementation now, there's way too much stuff that people rely on this object. This is being a regular object for regular use. Also do not do that.

M

But I IV with Jeremiah I I, just think that breaking the ecosystem twice doesn't make sense so so, or four Timeflies negative on this I think we should target something around maps for for v8 or longer diode timeframe and I. Think that would have a much better performance behavior, because that's really the motivation here right. We want to improve performance for and we are willing to break the ecosystem to do that.

E

No, we primarily want to make sure you can't set properties from outside sources that could break your application. Yeah.

M

That's that's fair, but yeah I think if we were to like it. If the goal is to eventually move to maps, let's just do one day.

I

I'll proposing to postpone the changes until then you I pass for for that.

E

The new which sorry Adam and gotta.

I

Then you okay, rich you're, proposing so is what early proposing here's to postpone all the semi major changes. Until we have a a final view of what the ufa will give I.

J

Agree without north until we move to a man.

E

J

E

I agree with that, because I think it's impractical, I think an API to make that will not necessarily get made. It will get stalled out and take very long, probably not by VA by version eight. This is mostly driven by past experiences. With looking into how you might have to change those API is very involved. I.

P

Think if we, if we're going to switch to maps, I think we'd probably want more than enough time to get at least some kind of you know at least all of their documentation updates. We would have to do at the very least, but you know probably some active messaging or something to let people know that sometime down the road, you know that it'll change so that enough people can be prepared for it.

E

Right, yes, it probably needs it probably is like a series of VPS or something.

E

So if we, if we do, that, change will still if it's not very communicated in a long time and or symbol for some significant amount of time will run into an issue where there's only so much the ecosystem using the old part, it will be extremely hard to switch off of so this, like you, really need to think about like moving off of things longer term, and because, if it's not that easy to maintain a map limitation and then also guarantee that this old and from API works, even though this old API is not documented, it is very wildly news for anything that manipulates headers as an example.

E

Express uses it pretty extensively because it's the only way to get at this set information.

M

One idea might be to convert the old API to a proxy and it could use a new sort of map based API and then we can slot in replication warning inside the box. Sita yeah.

C

I can take a first stab it. It is a photo because that would actually take too much effort to have a channel to do go to proxy rap. But you know it's worth. You know what, when taking time and make sure that we have the proper tests and place and make sure it's stable, so I'll take it to do to do a first positive. Oh.

P

I was just gonna say that I think well, I mean I, haven't checked recently with like v8 5.1 or two, maybe, but the last I had looked that like proxy performance. Wasn't really that good. So that might be a serious issue there as far as accessing and setting here, whatever.

M

Yesel idea would be that I mean for for the version 8 time frame. We want to defecate it so so we know that if we were not breaking the ecosystem, but we really would like people to move to the new API, so support deprecated did pass I think it might be acceptable to take a performance set. I.

E

Do not think it's well moving to an entirely new API is reasonable for the ecosystem until v8, but if we have people punch that or something that can help you use as move over to it, that might work. If you, if people feel strongly about this, we should put together a proposal and I might be willing to help write that, since I did some amount of Investigation working a lot in the past, but I don't think it should hold up this because it's not really, I mean they're, pretty different changes. Yeah.

A

This is not a very the change we're talking about now is not. He is not a the type of ecosystem breakage that will see with you know the api change to maps. If, if things go, that way, right, I, don't.

J

Necessarily agree with that, if that lands I'll have like personal good I'll, have to change if.

A

The current removing over the prototype properties from HTTP plans, that's what you're talking about yeah.

J

J

Mean I, it's probably more prevalent in application code, the module code, but I, don't know, though,.

E

You you're saying you object to to landing this, but you'd be okay with an API changed about no.

J

I'm not objecting, though I mean it would still be a simpler maker right. Yes,.

A

J

No I'm, okay with that, as always, it's ember major, because it actually improves the HTTP module I.

J

Mean wasn't that the main point behind it to access the headers that or saying this like a built-in name.

P

Pretty much yeah yeah.

A

All right, so we we still have one more item to talk about. We are now two more items. It's like somebody, I'm guessing bird added, a second item there, um so I'm wondering if we I don't think we're going to resolve this in five minutes, but is it still everyone's feeling that this this does not need to be talked about next week or should we should we talk about this next week?.

A

Okay, I'm not hearing anything so I'm going to just leave the label. Often, if someone wants it to be back on the agenda, they can just put the label back on.

C

So we're talking about the which one I'm sorry take the label off, of which one the.

A

Label is off of the HTTP object prototype one, but but anyone can add the label back. uh Okay, yeah that that that one didn't go the way, I thought it would. Oh, okay, let's, uh let's uh quickly, we were going to touch in only briefly about the what WG, URL standard implementation and then slot a big chunk of time next week to talk about it at least that's what we agreed about agreed on at the last meeting. So is there anything to talk about here, James and everyone else, just.

C

Need everyone to keep reviewing it, the idea would to be would be to land it as an experimental, undocumented, okay in master between w no experimental is it doesn't matter if it does, if it doesn't go into v6 or not, so you can just keep iterating on it, but please take your time to review it and we'll discuss it. D jealously.

E

Just as a note on this week's meeting agenda, sorry on the meeting thread, Trevor nars, says I believe my opinion. Url beavers it represented by others, but to make sure minus well 2-1 is a global URL right for what it's worth. I also object, adding any new Global's that are not language features.

A

Is that even in the EPS anymore, James.

C

Yeah they suck in the EPS. It was added based on the conversation in the EPS me that, because because it's a global in browsers most you see, users of it would expect it to be a global in the implementation I step. That is a separate commit, but I could easily you know, switch on and off, basically about whether it's a global or not so.

M

Really just you.

C

Know depends on where we land next to be cousins. Coming conversations great I can live with it either way. I.

E

Wouldn't be surprised that it in the future euro becomes some sort of es whatever global, but ok, you have a conflict. I can for say that as a potential possibility, the way things are going to get.

A

A great so anything else, hopefully it can be in the github repo on that issue, or we can talk about that next week we have two minutes and um I assume Bert. You wrote kick bird out of the ctc.

D

Knows no, yes, I did yeah, no I mean like it since you asked um whether I would finally officially step down yeah, that's yeah, I joined here by and so long thanks for all the fish and I feel free to invite me. If there's anything that actually needs my attention, don't that's all. Does there is there any formal process for it is like do? Does the ctc you have to vote for it or no? No voluntary.

A

Resignation is voluntary resignation, but thank you for your service and I predict I predict you will be invited as an observer, but that's just my personal prediction go ahead, Jeremiah so well.

E

I mean you can do it if you want, otherwise, one of us will will make a pull request to remove you from the ttc list on the nerdiest repo I assume Lee. We just put you into the collaborator section them unless you, for some reason also want to be removed. Not currently, there's no removal process from collaborators so sure I'm, hideaway,.

D

C

Think is the the tar and feathering should come within a few weeks.

D

Yeah but like it now, I can feel free, free free again to like say on Twitter about no JSO.

D

Okay, well, anyway, done all.

A

Right we're right at the top of the hour, which means we're all we're at we're out of time and I. Think we got through everything. Is there anything anybody feels we did not cover that? We need to talk about.

C

Nope, we didn't have a private section at all, so well.

A

Maybe we'll do is we'll wait for the streaming to turn off and we'll see if anybody has any private matters to talk about yeah.

E

um In terms of QA know what I put up a solicitation for questions already on the stream? No one has asked any questions, even though there's five viewers so I think we're get to end here.