Node.js Core Technical Committee, 4 Jan 2017

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: 2017-01-04 Node.js Foundation Core Technical Committee (CTC) Meeting

Description

https://github.com/nodejs/CTC/issues/54

Anna Henningsen - https://github.com/addaleax (CTC)
Colin Ihrig - https://github.com/cjihrig (CTC)
Evan Lucas - https://github.com/evanlucas (CTC)
Jeremiah Senkpiel - https://github.com/Fishrock123 (CTC)
James M Snell - https://github.com/jasnell (CTC)
Josh Gavant - https://github.com/joshgav (observer/Microsoft)
Michael Dawson - https://github.com/mhdawson (CTC)
Brian White - https://github.com/mscdex (CTC)
Myles Borins - https://github.com/MylesBorins (CTC)
Michaël Zasso - https://github.com/targos (CTC)
Sakthipriyan Vairamani - https://github.com/thefourtheye (CTC)
Trevor Norris - https://github.com/trevnorris (CTC)
Rich Trott - https://github.com/Trott (CTC)
Bradley Meck - https://github.com/bmeck (observer/TC39)

A

Just ttc, meeting or january fourth, the first one of 2017 start the stand up so on a Europe first.

B

B

Almost at this, she had a pretty bad cough, so not going to be much audio from her, but the ok stand on the notes as an anything penfield are yep.

A

Definitely feel better, um so next is do we have Bradley.

A

Have pita crisps calling yeah calling eunuch.

C

Just reviewing issues in for quests uh some testing, pr's and enjoying the holidays um next is Evan.

B

Even we're getting more loud, a buffer and crypto random bytes and then been working. All I'm cutting setting up for auto yeah next is Jeremiah.

D

Holidays other than that just miscellaneous minor stuff um next would be Tracy if she were here so I, don't think the doors here James.

A

Taking some vacation time, I've been reviewing and landing the PRS and then looking on some URL and do stuff and next would be Josh. Sir.

E

So yeah come back from vacation like everyone trying to catch up on some of the diagnostic stuff this past couple days. Next is Michael Dawson same.

F

For me was on vacation last couple weeks and this week catching up on issues and reviewing some TRS I.

A

Think Brian's next.

G

Ok, so I'm also worked on uh submitting PRS for a bunch of various performance improvements in core other than that. Just typical commenting on issues and reviewing PRS next is miles.

H

Hey so released a new version of version four and six yesterday reviewed PRS commented on issues changed, my handle will notice I'm no longer the Alpha nerd and just general work on canary in the gold mine. Next up is ally, I, don't think, is here so I'm, just gonna go down and I think this target us here Michael. Thank you. The next one is on the list.

I

Yes way so I tried the holidays and I watched a few issues and pull requests not much this week. Next is sexy. Prien.

J

I was working in the debugger test and I was trying to ah it's the child process to choose the kill signal thing up on that for the code, review and I was trying to fix this. The streaming team.

A

Okay, I think that is everybody. Did we missing you missing? You know Trevor's coming in, let's get to the second and everybody sees highlands. You.

A

Occasional joining you should pop in a second hey, Trevor.

D

F

Getting bigger right.

D

Now, if you have an update to get Trevor.

K

Yeah there we go, he kept moving, my icon around so I couldn't eat myself, um vacation, mainly acing cooks, I think, ah but having some great help, writing up the tests and docs and as far as the API I'm, just about done fixing up all my remaining two dues, so I should I get those tests. Uploaded I think they'll be ready to be fully review.

A

Hi, when I hope, everyone had a good holiday and nice and relaxing seems like most of us, took a took some time off before getting to the agenda. We have the previous medium meeting review so ever call us. We didn't really have a lot of folks on the call.

A

What we did, this video talk to a few things talk about the NPM upgrade to 405 in master and b7 and as its landed and masturb I'm, not sure the side of indy, seven yet talked briefly about console log format, supporting the I integer conversion, some cleanup on timers the copyright attribution stuff I, pier that I was working on a PR I have for adding codes to warnings, and then we talked about the CTC meeting times the note on the CTC meeting times in case. Not everyone has seen.

A

The notice is that we we've added a third time to the schedule to make it easier for our international participants.

A

So we have the two times: we've had already I, don't know the conversions to you know what the PCC conversions are within 8.

C

A

Noon, pacific and then the other one is at nine p.m. pacific.

A

And there and they yeah they're, going to rotate each week, so I'm definitely make sure that you're taking a look at the schedule. That's been updated in the ctc repo. For that.

A

Any questions on the previous questions or comments on the previous meeting review before we get into the main agenda.

A

So main agenda items first, one we're looking at the node.js node repo number 1 05 05 landing in p.m. for into node. Seven Jeremiah and II know where this is a sad.

D

I I mean so NPM for is in node. Master now contains relatively few and generally very obscure, breaking changes. So the discussion is if we should pull it back in the node 7. As we discussed a couple weeks ago,.

D

And there's support reasons for wanting to do that. He we get new features then still in node 7 4pm, for we also get to test mpm for a note 7 already before we go into our next LTS.

D

Also, it might be a bit of a stopgap if an NPM 5 comes out before or even after the next LTS and additionally, it could be used as a sort of gauging mechanism, given that everything in it that is breaking is generally hard to hit. We could potentially pull that in and use that to to look for how much reaction is, depending on that sort of stuff, for the people that actually use the npm that ships with node.

D

D

C

Essentially, we need sign.

D

Off if we are comfortable with this or not yeah,.

A

Is anyone recalled have any objections or reservations about learning if you have 47 invisible.

A

A

Right so not hearing any objections, I mean I. You know I kind.

D

Of have like reservations to but I think they're like acceptable.

A

Yeah, I think my mikey reservation would just be just wanting it to sit and master for a couple of weeks to make sure that people actually had some time to test it. I'm, not anticipating that any any actual issues given helps to the breaking changes are.

D

A

B

A

B

Feel like it's ending.

D

In master, doesn't really I, don't really yeah, so I don't have data to back this up, but I would highly suspect that mpm pre-released branches get far more use than our master Brant. Well, not okay, not necessarily than our master branch, but then the MPN. That's in our master branch right.

B

Yeah I feel like they put in v7, would kind of give us that kind of data yeah.

A

I'm kind of leading the same way we.

D

Could do like an RC or pre-release build with it in and see if we can get a couple more people to test it, but I think it'd be easier, just to ask people to install NPM for and see if it works for them. If they're using on p.m. 3.

H

Yeah true one thing that we can consider as well as, like you know they, we know the life cycle of version 7. We could especially it becomes much easier for waiting a couple weeks. Anyways is we could wait until we're like two months out from the next major before landing. This know that way. If there are complications, it minimizes the amount of time since we're going to be dropping support for seven anyways when it's done I.

D

Would figure that if there were complications, we would probably revert back to mpm three.

B

You know that's what I would think I feel like. We would want to know those complications sooner. Yes, that's why I'm I think I'm. How would like for it to land in it b7 like sooner rather than later, that way we can start preparing people for when VA does or version 8 does come out.

B

F

I was already landed in master yet or yeah.

A

It said master yeah.

F

How long has it been there.

A

A week sitting.

H

Yeah I landed a last week after at the ctc meeting.

B

You would anybody object to be cleaning it and 7 dot for dot of today.

A

As long as we are okay with doing a quick reversion, if any any significant issues, you come up diagonal direction.

B

Yes, I will take responsibility for that like if we do use our Roberta yeah.

F

I think that's I'm on the same wavelength as long as we're going to revert it quickly. It seems like it. The sooner the better gets us more advanced notice that there are going to be issues.

A

Is this any other drug good.

B

You're saying is this something that we need to vote on or.

D

B

D

I, don't think when we need like at least CTC review it, because it's a weight well, it otherwise would be a cember major right, I guess we should so so long as like we've looked at it. We should be okay, I.

E

Think that the process is, if there's any objections, then we have to go to a vote. But if there's not then- and we have consensus I- don't don't think we have to go two of them agreed, but.

B

Do we have enough people for it to be consensus, considering not everybody's here, it's kind of the point I was gay, not.

A

I think that is this point. We have enough consensus to get a PR open benefit. If anyone who's not here objects to it, then they can weigh in on the github thread.

A

A

C

A

Tag it to make sure we mentioned CQC taggat for CDC review.

B

Well, there is an issue, I mean there's an issue that have been that was on the ctc agenda and nobody has well. It doesn't. Look like anybody's objected to the bear.

B

I'll plan on including it and if anybody does object, then they can do that in the ER, but I'll get that updated after this call.

I

A

Any other comments on that for removal.

A

I'm, just looking for already halfway through the time to call we don't have a lot to you have several knives on. So it's keep moving console.log detail format supporting the opposite is a jerk conversion 1029 to.

A

Some are quite sure where this is at. Anyone have any connections to this.

D

C

D

Tagged us, but like two weeks ago and they're like this, this issue is not really just like this one thing, so I'm not sure what we're actually discussing at this point in time.

E

D

I was thinking the same thing when.

E

I, look this over. It looked like there are a number of non-standard console, behaviors or quasi standard, and this is one of them, but maybe it's a bigger picture. How well kind of standards.

D

On console, do we adhere to their there? Never was a standard right like this is the problem. Again we had this sort of like with the you. This is it's like the same thing: we've experienced the URL parsing stuff. There never was a standard. So now, there's like one it's by the what we G working group, or rather there's like a work in progress, spec.

D

And I mean it's: a good amount of our behavior is not the same as the browser's and has never been.

A

To be on it to be fair, that's actually a lot of larger reason. Why they're looking at creating aspect, because there have been difference to the cross implementations and it's always got back to hey, there's no specs, so we're going to do what we need to get done.

A

So the fact that putting a speck in place is consider that positive step we just have to decide how close he was going to want to track it once it's there or if we even want to get, we may even want to get involved in helping shape it right. So.

D

A couple of us have gotten involved in talking I guess like the spec process, for that, what we g working group is mostly based off of like specking. What's there that's most common, but like that's like really problematic for some things like console dot us or for the last four or five years, we've had like a completely different behavior and changing. That would probably be painful.

H

You for a cert, there was the discussion about perhaps making a new API surface that had converged behavior between the browser and node I. Think, like you know, the meta issue here is whether or not we want to make an effort working with the what working group to try to find conversion convergence and the biggest struggle.

H

There is going to be that a lot of the browser vendors have similar implementations and in a lot of these cases, nodes behavior will just lose out due to you know, just pure numbers as to how many implementations exist that are different.

D

Yeah I know I, don't really expect us to be able to follow perfectly whatever spec is there anyways.

D

A

Not now without either taking the Signet the significant break and change or taking a route like what we do, URL basically in parallel, which is you know, has problems of its own I mean.

D

That yeah well, we can't really do that right because, like the entire idea so like we don't need the console API in note right. The entire reason why it exists and like the entire reason, why, like set timeout and set immediate exists- and you can look back through like the commit history for this- is to make it like- have the same API roughly as a browser. But then not quite because things weren't like as uniform, four or five years ago,.

F

So that would be an argument towards making it consists as consistent as we can I.

B

Think Sam Roberts made a very good point of the issue about how like how like what it does, what it's meant to do in a browser versus wouldn't admit to do and notice pretty different yeah.

D

I mean so I mean the spec allows for this, but, like those so um console dot clear the spec says it can do nothing but like I mean that's something I that maybe people expect it to do something most the time. But like there's no way we can implement console dot clear because we don't have control of the terminal or really whatever is consuming the standardout streams.

L

So we should be a little careful because the inspector actually hooks directly into our console API I- have people thought about this.

D

mmm What do you mean.

L

If you console dot log and you have the node inspector open, the node inspector also displays it in addition to that, it actually patches console if you fire up node with the inspect flag on.

D

Maybe it shouldn't be doing that.

L

It needs to do that if it wants to show you the console yeah.

A

I think it's that the fact that we we've never really produced any way any clean way of hooking into the console output other than that is basically hacking into it. I.

D

Mean you can listen on the stream, but I guess like that, doesn't work for like some of these other methods that.

L

Doesn't entirely work either because they need to actually implement things like consult group. Yes,.

C

L

Added and other things also hooking into the stream is after it's been formatted, and their formatting is slightly different, which is part of what this issue is.

A

Ok, but do we have enough information right now to be able to make any kind of decision on this call? It does not sound like it. I think we probably need to return a conversation to github and revisit it.

D

I mean the only thing we can well stay is that at least whether or not I guess most of us are interested in following some sort of console, spec or not.

A

um I would say eventually: yes, anything you know, depending on how much pain is actually involved in doing so.

A

At the same time, I would I don't want to see those smaller changes like you know the the percent I, and you know the this particular pull request. I, don't want to see these kind of things get held up by that, because the changes that I see in this percent I in 10 to 90 to seem useful.

A

Whether we're following that standard spec, precisely or not,.

A

So I mean: is this designing this PR make it? Will it add to our pain later on this.

D

Isn't the place.

A

I'm sorry, this.

D

There, as far as I'm aware, there is no PR.

A

F

Thing is an issue.

A

F

It does suggest the code change, which is like a one-liner.

A

F

Says they can create.

A

F

If it makes sense.

A

B

Know if we should I mean this can be done in you to land right, it's already done in usual and like in there a four-and-a-half package.

D

Yeah I mean the idea is that it's like standard and always in that you can always rely on it being in there by default. I.

L

Think something that should be considered is we actually ship two different consoles with the inspector having its own implementation entirely.

L

D

I mean I, think I'm, okay, with like moving towards more spec parity, but I'm, not on any real way interested in losing existing functionality, which is part of why I'm not very happy about the % d thing because, like our thing shows it if like, if it's a if it's a floating point, it just prints it out as a number.

D

So you still see the decimals if it has those and if it doesn't it just prints it as an integer and we'd, be losing something like that, and maybe I mean, like % j for jason, probably would not be like spec. So would we I guess we wouldn't necessarily be required to remove that, but, like I mean, why would we want to.

L

We could try to add it to the spec. I.

D

Was told that was not particularly an option at this point.

L

A

Visitor any fight.

L

To keep it the truth, I can move some stuff to that. One.

B

Maybe we could get somebody from that. It works on the console standard to join, leading and kind of see like where the differences.

D

Are I mean the differences are mostly like documented between this thread and there's a couple links in this thread to the what we G repo, which is what we g / console all copy that and paste in the char.

A

Okay, I think im just looking at the time and we have a few other issues. um What are we? Why don't we return the conversation to to github? It seems like we have more discussion that we need to need to have around this.

A

We need more investigation into what into what the what WG is doing legion. If we're going to get we're going to try to get more involved in what WG effort to actually define this, we need a decision on how much we actually care about following the spec.

A

So and I just don't think we're going to be able to get through that all right now.

A

So unless anyone has any really strong feelings or pins market may so express right now, I suggest we move on.

D

A

Okay, um timer is cleanup extraneous property on immediate 10, 205 Jeremiah civil fears, yeah.

D

This is mine. Can we please get rid of this property? It was added kind of accidentally in a commit that I, reviewed and I missed this, but there is no an extra property on immediate the duplicates. Existing functionality does not work as expected. It was added recently, but.

E

You realize you're.

G

D

D

Please don't meet me there I'm not going to notice and the current setup um so right.

D

What was I saying to myself um right, so there's there's now this extra property on immediate. That is partially my fault, because I merge the commit that added it. um So it's been in there for a couple months. It duplicates an existing functionality, that's already there. It doesn't work as you would expect you can't always assign to it, and it won't always work the way. The previous property that also still exists does and there's no there's, not really any reason to keep this like again.

D

It just duplicates behavior and doesn't even work the way you'd expect it was added about two to three, maybe two and a half three months ago and in the meantime, we've also removed properties that have been a way around way. Longer and I've been wasting more visible, so I don't really see any reason why we would not just remove this property to.

A

um Trying to look enough, no Nikita did some looking around on one of the southeast k-member fit. This is one that he did to see if it was actually used anywhere and whether been it is there any evidence whatsoever that is actually used.

D

You um I not sure he has looked at this thread, but um this property name is not.

D

I think you would get a lot of false positives. It's just underscore call bak, in fact they might have already conflicted, is something that someone else in ecosystem was doing for all we know, um I mean we could lock.

A

But if it is an underscore property, it's never been documented to live in there for a few months, yeah.

D

If anyone wants to be using this, they are going to be using underscore on immediate, because a that has been around for the last I. Don't know how many years and it actually works.

M

Hi, this is rich I, gotta drop off soon. I just wanted to really quickly register on a micro level, I'm totally okay, with this PR landing on a macro level, I'm not comfortable with the way. We do these things, just in general, that we don't seem to have a concrete policy about what we consider cember major and we kind of like yeah this one's you go I just want.

J

To register that.

M

Before I drop off, that's all.

D

Also, I don't want to be set. Various property into the future did.

F

This go out in any of the LTS like. Is it in sex.

A

Think if it's, if it's in six, we should take it through a deprecation cycle. If it's not in six, it's just in seven, then I'm thinking, okay, we're just dropping yeah.

F

I'm sort of like that'sthat's to what I was thinking like I, think it's a much easier decision as it's never been in an LTS that I think we could. You know give ourselves. The flexibility is make that one of our policies that you know it it hasn't been an LTS that it doesn't necessarily need that deprecation cycle, hey.

I

It is in v6.

A

D

Like it might be in v4 to.

A

Ya, get all right, give it I get my.

F

Clipboard its will raise.

A

A

Yeah um Jeremih I definitely emphasize that just not wanting to keep things around that or just kind are broken, but I think it's just safer, even a even if it's extremely unlikely that anyone who's using this. I think it sends a much better message to the ecosystem. If we are consistent about these things and even if it's an absolutely pointless and work with api, we still take it through a proper deprecation cycles and sista an LTS.

D

uh Why would we not just remove it in v8, then, like it's an underscore property.

D

It's not even like documented is that that's not.

A

A proper replication cycle I mean we've, there's other things that were you know. We would like to see just moved to right.

D

Yep, but does actually have like usage. This is ridiculous.

A

M

All we needed some cup some kind of exception policy like.

D

Okay, if you want it like deprecated this thing, you're, probably going to d, opt or like make a set immediate, not that fast, because it's kind of inconsistent on how it works. I mean I, know may-maybe could but like it'd, be weird I, don't know if we make.

A

All the in, if we go ahead and change all of the internals, you.

F

A

It's using the on immediate and just make the public accessor for let's roll call back I think that deprecated, why would I diop to it in the code inside me?

A

None of the internal usage would use it. It'd just be an alias for on me. It the way it is now.

A

It would only do if somebody's actually using.

B

And we see if there's a performance difference and adding the deprecation warning for without it being used good I'll assume we just do like a gator on it. No.

K

Okay, I feel like further discussion on that requires that somebody actually ready test first, so less on the dress hypotheticals now on a macro level, again it almost everything and node is in some way publicly accessible, even if it's through a ridiculous way. So does this mean right from the discussion? Does this mean that every time we want to remove or change a property on internals that we have to deprecate it first.

A

It means we need a better policy in what we consider an internal, it's okay to just remover, certificated, well unknown. Oh.

K

It's what I'm saying is: how do you define internal like how internal is so internal that we don't care if we break somebody like anything on process bindings or not.

K

Or anything on a constructor that you can only get via a detail, constructor property from an object from a you know, kitten handle so.

M

These are, these are really large discussions. That's probably going to take some time to sort out and I guess. One question is we want to hold this up while we're having that discussion, or do we just want to make a decision on this landed or whatever and and- and you know, move the discussion of what really you know? What really is our semper policy move that to a separate you know issue you.

A

I think we definitely need to have a separate conversation about. What does cember policy is I'd started? It p are kind of refined that it kind of stalled out on discussions. I can now refresh that conversation. We can continue it there and you saw that.

K

Ink the mid knows.

A

K

A

The as far as this PR is concerned, I think this we can. We can get some experimentation. That shows you know, adding a deprecation message whether it does or does not cause a performance impact on this one. It's not used, then we can make for da I keep coming back to I. I would rather this go through the deprecation cycle, since it is an LTS if it.

C

A

Been delivered an LCS, then I would say, go ahead and remove it, but since it's there and we have that stability requirement and you'll see I said we should err on the side of the code of more caution.

D

Well, we can not remove it in the LTS if you really want.

D

Again, keep in mind that, like I mean these might have even been back part I didn't check, but we removed stuff in optimization commits that was around way longer. I.

A

Know things would come back and did know.

D

Those have come back and bit us why.

K

Woman, why don't we just call for a vote in the PR, because this discussion is going in circles.

K

Okay, because I mean it.

F

K

This needs a vote and we definitely know how consensus yeah.

F

I think try prints. Other point was that we really sorry richest. Other point was really: we need to also come up with a way that we have. You know a great set of rules so that we don't look at each of one of these as a one-off and just say, yeah. Okay, this is yeah.

K

Jane said he's going to link a PR in the dark notes about a PR you created about that specific thing. I didn't know what had been created. So hopefully we can get a special yet.

M

Okay, I'm, maybe label let Peter ctc agenda yeah.

A

Ap updated, I mean what I put in there originally was largely straw man, so it's really kind of figure out what we, what we wanted to do is open to start the conversation about what we wanted to policy to actually be looking for the link right, you you.

A

A

um Yeah, so that's we can open up for more discretionary vote in the Indian PR. For this again. They that is 10 205.

A

A

And let's: let's open it up for discussion and voting also the next week and then, let's make sure this is on the agenda for next week's call or final decision. Okay,.

D

Okay, um so I'm not really like very happy about like the politics it's kind of going around this stuff like it seems to depend on like who knows what and like can move the most amount of people so I did the other poll requests that I'm talking about that remove sighing I actually objected to that being removed, and nothing was done about that and it's okay.

K

Like this isn't politics, this isn't politics its policy and our policy is crap, well non-existent, and so it the fact we need policy is just coming up strongly right now, right, yeah.

A

Yeah I agree with this. Many it really comes down to just you know. What is our policy going to be in we're going to stick to since you.

C

A

Having having to go through this process of you know we're going to make an object exception for on a case-by-case basis, it is, it makes it really difficult.

F

A

Okay, yeah, let's just until it lets us open up for vote in the in the github thread, mom and then my next week. Let's have a decision.

A

Okay, we got like ten minutes left in the call. So you know we have Morgan items together.

A

It should be relatively straightforward, but a lot of time left moving on issue, 101, 16, actually PR, 101, 16, add code to learning design codes. Duplications when I put on the agenda largely just to get books looking at it need some either sign off or objections to be right for this. Before can land, it's been open for a couple of weeks now, like s folks to please go, take a look at it and weigh in I appreciate it is a fairly large PR.

A

What it does is add a static ID to all deprecation and prints that ID out when the warning is printed. It also adds a new command line flag or redirecting warning output to a file.

A

This was done from a result of a conversation at the collaborator summit in Austin to make dealing with deprecation easier with regards to NPM installs and this parrot ecosystem thanks, basically, NPM would be able to set this command line, flag or environment variable I wanted to install so that the installs aren't as noisy, and they can do more intelligent things with the deprecation warnings that are printed.

A

But just again, just large million people now folks to go out and take a look at it.

A

Does anyone have any comments, objections on 101, 15 formula.

K

Looks like it follows a similar idea to the code to error PR that you create in the past.

A

Yep same basic idea: the error code, ID, is still 1, I'm going to be pushing on. I purposely haven't been because we've had other things going on in it and that one is such a disruptive PR touches. So many different things. I want to make sure we get it right. It's actually. It takes a smaller chunk. So if we can get this one right did kinda gives us a stepping stone to do any errors, cool.

A

And if this PR also gives us a single page in the docs that will list all of the deprecation, even doc will need applications and runtime their fabrications. So we have a central point to point central place. Appoint people to for information opt out to the deprecation, so the decision that we made a pie and what else to do that kind of thing give us a document all that.

D

Alright guess I should have a comment in the thread, but I think we should at least for like really old deprecation. Eventually, we should at least have policy to just remove everything, except for like this was the defecation. This is what was deprecated and use this API like that that information couldn't be stored and get history similar, but once it's like five years old, like we really know, don't need like tons and tons of wordiness on it, it was I.

K

What about deprecation entire modules like domain domain is completely deprecated, but it itself is not listed. Every.

A

Every deprecation, whether it's modular, believed a little whatever.

C

A

Would be given in a static eight ID, it would be listed in that document. Net I'd, never mind ever. Yes, sorry, my search food.

K

Just failed I see it sorry. So yet.

J

Requiring domain.

D

Would just be like a separate ID? Essentially, yes,.

A

Yeah and those ideas would never change, so, even if that code eventually is completely removed, the ID for that deprecation would still remain in that document and would guide IDs would never change. So, whatever.

C

That, whatever that.

A

Deprecation is if we complete, will move the domains module at some point in the seizure, the documentation, for that would still be there in that one and it would wait to be reduced around this thing is pissing no longer exists, don't use it.

D

A

D

A

Deprecated old debug protocol, one of 320 I, think.

E

My girl zasu added that just now that right, Mako.

E

I

I added it in the google docs, but it was added by rich to the agenda.

E

Actually me I just copied it from the minute section.

D

Like okay, so what I'll? What other choice do we have for this other than like shim it in v8 and deprecated in v8, which is like massive overhead? What.

E

What exactly is the question here? Whether it should be deprecated whether question.

M

Is this the deprecation the court docs only deprecation should probably happen now, if, if the debug protocol is going away and be in inversion, aid of nodejs but Doc's only deprecation is semper major, at least that's James's.

A

M

And probably the opinion of a lot of other people so.

A

We've we've always treated it as finger major right.

M

I would argue that kindness to the community means getting it in the blog mutation now, rather than waiting for version 8, because by version 8 will already be gone. I think.

D

We should do a run time to application of it, because.

B

It just won't exist in the next.

D

Major we we can like ducks deprecated it like back further. If we need to, but like it's just not gonna like even exist, you might as well tell people like right up front that, like this thing is not going to exist, and there is a PRF.

E

That, although I wrote it in there, the other thread with the PR for the runtime deprecation encode I have to update it. To use I think that to use James's new codes and thanks, um but the pull request, is there already I? Think that's what you're saying for my event, one I don't.

M

Think we want to run time deprecated, so we have no to inspect right.

F

In terms of the.

M

Dogs are offered in.

F

Terms of the doc deprecation, so if we don't want to actually make it a deprecation, can we just put the same text in saying we know this is going away in in v8, therefore be warned, it will be going in v8. Why.

D

Not just up again it then that's when the deprecation is yeah.

M

I, don't really see the point deprecating something without deprecating. Well,.

F

M

F

We're looking could.

M

F

Be I can't see why we wouldn't give people a heads up that it's going to go away. If there's any concern of those the fact that deprecation have a certain process they have to follow, then you know that would be a way to get around that.

F

If we, if we think making an exception, doesn't affect the process and there's not a concern, then yeah we could just go ahead and if, but if we say well, we don't want to be seen to you know violating our own processes, then just basically giving people a heads up, there's another way to go.

D

F

We're we're violating our.

D

Own process anyways by removing well.

F

I think well: I, I guess, for V. We've always basically taken a position that we can't fix all the VA changes, and you know that they just happen right.

A

Yeah, unfortunately,.

A

And it goes back to you know not having a great policy on this, and this was this was kind of being forced on us without being able to get through a proper cycle.

A

Thank so much as it sucks doing a runtime deportation in v7. It's probably the best thing to do.

A

The most responsible thing to do at least so.

M

This pull request is just a doc's deprecation, so I mean, if that's, if that's, if there's consensus, that that's the way to go, let's make sure we document that back yeah.

F

Let's agree on that and get that that part done for sure yeah.

E

So we're selling the doc deprecation go ahead and do now runtime deprecation, we're not sure that we want to land now.

D

I mean I think they should probably just happen like one commits like why? Wouldn't they well.

M

My only thinking is that um until we land note inspect yawns thing until you were talking about bundling, it went when we have that will have a command-line debugger again, and so so that might be the time to do the runtime deprecation rather than doing the runtime deprecation when there's no alternative. Yet what.

N

M

K

N

The difference, though, I feel.

K

Like that's the same thing, we're doing with like the domain module, we all hate that I'm a module we all want to go on, but it's only its documentation only because there's no alternative to it.

D

Yeah but I that that saying we can actually decide not to like remove an ax version. I, don't really see why making a not why having note inspect as an alternative really um makes that much of a difference I mean we can update the message that this might print then, but whether or not we have an alternative. It's going this this thing is going away. We can say this is going away and we're discussing it. An alternative at URL yeah.

A

E

Be clear: actually, um there are actually two deprecation zai guess when you think about it, there's the one which is dash dash debug, which would activate the listener and then there's no debug Newton space debug, which spins up the CLI debugger, ah there's there's some other related api's yeah, but I'm getting at is that you know, even if we don't have knowed inspect that we do fully have the other. You know the new listener. Is there how it's just the CLI debuggers?

E

That is the note inspect, so maybe that shouldn't hold up, including that run through, since the main thing is not the CLI debugger, at least in my opinion, it's the listener. You know whether we're listening for the inspector protocol or the old protocol ah or disagree, you think the CLI debuggers. Yes,.

D

E

We could include a warning on dash dash, debug and wait till note. Inspect lands to get the warning on node space debug.

E

We could break up the two deprecation 'he's right now, they're all in one PR 102 76, but we could split them into two and then we could hold on the node space. Debug deprecation till node space inspect lands. Maybe that's systems are.

K

Your good starting point and put that over vote in the PR um it's way easier to deprecate, passing a parameter.

D

I'm still not sure why we wouldn't like deputy the other one too, unless unless we really still think that this discussion of maybe making node space debug an alias to node space inspect in the future like, unless that happens, it's just going away.

D

K

Like it's just gonna disappear next word is, is we're out of time? Is this discussion going somewhere or should it be pushed over to the PR.

A

Push it over to the PR I think we have a consensus that a doc's deprecation is something that definitely should happen. We just need to get clarification on whether and when to do the run tight, deprecation, yeah.

A

The next two are related. We've just got a couple of minutes: I don't know if we need to do much discussion. I just want to draw people's attention to it.

A

The restoration of the copyright attribution discussion I have updated to PR that restores the original copyright and license header in only the files that are still remaining from the PR that changed the copyright headers in the first place, and then I opened a second PR that reverse that and then changes the copyright header to using new copyright in a link to our license file, it's kind of a torturous process, but what needs to happen for both of these and they have to happen in order- is that the node foundation legal committee asked to review them and make sure that it that it's a bit adequate and that we have to sign off on it as well.

A

The one that restores the original attribution is the legal committee is basically telling us we have to do that, so there's not much that we can do there. There are some other conversations that are happening for to get clarification. I know. Ben is asked to talk to the legal committee to get some questions answered. If anyone has any specific questions for the legal committee, I I encourage you all to reach out directly connect to them through Michael, but we need to get review on these very least.

A

We need to land the one that restores the original contribution original attribution fairly soon and then the other one that changes to that attribution is going to be a larger discussion. We have to get the legal committee sign up before we can line them yeah, so just encourage folks to go. Take a look at those and review.

C

K

Okay, so you're saying I can ask the legal community a question via Michael, yes, they're.

A

Apparently is no direct way of contacting them, so you can connect through Michael TG.

K

So do you know why I'll top your head, we couldn't I mean it seriously sounds like all they want is to readd it for whatever unknown reason, um but then we can change it over.

K

Like ooh I, don't see why we can't discuss what we're going to change it to and then essentially just put the whole thing at one gr where one stupid commit we introduce some it very next commit we swap it over to the new one.

A

If you look at my second commit, that's exactly what I do so I have the one PR that restores and then I branched off that reverted that commit and then King, you know added a new commit. It changes to the noob. What.

K

um Which the was the fewer number, because the one I'm looking at only has one kinetic. They.

A

G

Their stuff up- yours, oh.

D

K

Super pure okay, I.

D

Think what you're asking is that if they can land at the same time, yeah.

K

Promotions in PR, so they're like they're right next to each other I thought.

D

It's very unlike it happened, because um the the legal side of things will want to review what is done once and.

K

That is gonna, say it's because they can review it in in like diff format. That is absolute crap. If you tell me it has to land in code for them to review it. No.

D

It was sorry it's not really about the review it's about like well, it's like it's about joint. Actually, then, seeing that we've done something about it and like like actually done slang.

K

It's not great.

N

It's in the committee tree, how can you not.

K

See that it's been done well,.

N

K

N

The commit history until you merge it. Yes at some point so.

A

The leads the point, the point of having the separate PR that restores first SEO, is making a good-faith effort to make sure that we're following our own license requirements and.

K

A

That good faith effort ah the second p or that that changes a license also shows that hey, okay, we're going to take this through our our actual defined policy process of having a legal committee review. These China changes before we actually lay on them. Okay,.

K

So are you have a run both PRS okay, then bulls and then land both pr's right next to each other? If.

A

If, if we can get the legal committee to say, okay, that's it goes, the new changes. Look good right. You can.

F

A

Having those those you know, the old copyright holder agreed to it, then we can just land my second PR and we get exactly what you want, because it has all three commits the original.

C

A

Its original restoration, the revert of that and then be a change. Okay.

K

A

How do I put pressure on their to do that.

K

A

K

Process goes anything like what this has been like. I, don't want to be stuck with that massive header at the top of our files. Next, six months, we can. We.

A

Can try, but we have had an issue with a lack of responsiveness from the legal committee and attack right.

A

So you know the best we can do is knock on the door, a bunch and just say hey. We really need you to review and you know give us a timely review. We.

M

Move this to the private side that this means going way over.

A

Okay, so yeah we're definitely done with the public agenda. Oh, do we have any anybody queued up for Q&A.

D

Not really any questions someone says: they've been using mpm for for a while no issues. There was someone that made some comments about the console, spec conformity and I directed them to the issue where they have left a comment, and then someone else says: oh, the levels of frustration is comedy, so I think we're good to em pay. Okay,.

A

Cool and thank you all for driving on the public portion and the other part of the me comedy.

M

Tragedy, whatever.