Node.js Core Technical Committee, 29 Mar 2017

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: 2017-03-29 Node.js Foundation Core Technical Committee (CTC) Meeting

Description

https://github.com/nodejs/CTC/issues/85

A

Live now welcome to the node.js foundation, core technical committee meeting for March 29 2017, and we always start with a stand up and today will be no different, so is on here. Hi.

B

um So nothing out of the ordinary, mostly issues in peer review and in particular the reviewing VV. A key is stable, a bi-stable native module layer and Nexus Bradley.

C

Just started the C++ point of yes, modular I am informally starting on boarding for guy Bedford from cas to help out with loader hooks I I think the next is Nikita.

D

Yeah, yes, nothing worth mentioning for me and hence this colon.

E

Nothing notable issues in core class 9 x-7, another.

F

Them that, for me, Nexus Francisco, I.

G

Cleaned up a bunch of old pull requests, Nexus Jeremiah.

H

um I got the pull request, updated or rather a new pull request in for promised defaults errors for unhandled rejections on GC and stuff I will post a link to that. It is about as controversial as you would expect.

H

So if you have feelings about whether we should do this or not, please let me know, because there's actually not quite as many comments as my avi would expect. I don't know. Next is james. I'm.

I

Finding PR is getting ready for 800 lease and wrongdoing tests, any much tests and.

G

Basically, for me, I don't have to listen from me so much your reasoning.

J

That would be me a little bit uh involvement in the discussions around the x disk speed, so we can get cinch and running participating in the napi review work on the other end and then just the general miscellaneous review coming lands, and I will put out the call again if you are going to get involved in jump into the review for any p. I would be great if you did that in later.

J

Otherwise, brian white is next.

K

Okay, so I finished up the next set of file, FS module optimizations just actually last night, so I got that PR out there I updated some open pr's that I had out there and see just other than that. I think is just mostly the reviewing a PRS and commenting on issues as normal.

K

L

Yeah was that a company office I mean done much and I'm on the phone, so I dunno, who sings I.

M

Think that's me I, nothing too much on the beach side. Acceptor were you know, working on turning on turbofan ignition on in v8 5.9 next step is michael.

N

Is a nothing of note? Next is trevor.

O

Using hooks, be our bridge.

A

I think of know lots of little things. Let's talk about buffers, but before we talk about buffers, did we miss anybody, doesn't look like it cool, ok and then, let's review last meeting, there were two items. One was the buffer thing which were going to talk about it again today and the other one was the adding support for a bi-stable module API. We put. We sent that back. Take kid hub, I think we made a decision. Let's send buddy, remember.

J

I think last time it was just up for awareness and saying you know the tiara. Was there please jump in and review so expect. That was. The conclusion is just go back. Please comic-con the pr they're cool.

A

And I see it has comments as recently as 18 hours ago, so that is, this is working locally or one point, but excellence. ah Ok, anything anything! Anybody wants to add to the agenda before we proceed with talking about the future direction of the buffer constructor.

A

A

It's many many many long threads about buffer constructor, so um I'm going to go ahead and frame me.

A

H

Pardon what was that we can.

J

I think we lost you there for a sec. Okay,.

A

I just said that I'm gonna go ahead and frame the issue unless somebody else wants to take it. I am looking at uh okay, Gia yeah.

H

A

Sorry I'm confused um so.

A

Many times here we go great, uh so there does seem to be consensus. Although not you know it's not unanimous but consensus that we should do something that we can't just do nothing um and it looked like uh one of the things that had sufficient support and not significant opposition.

A

Was the opt-in deprecated flag that James put together a PR for, but now they're just seem to be some mild opposition or questioning um so I, don't know if that's the place to start another possible place to start is having the discussion about whether or not we are going to announce an intention to run time. Deprecated the buffer constructor at a specific version in the future, because a lot of people's support or opposition to other actions depends on whether or not that is included um and I.

A

Think that that's probably one of the bigger sticking points ah and I guess that's I mean there's a lot. There's a lot to talk about, but I guess those are kind of that. The two things I was hoping to hammer out today.

P

I

I would say that there's like there's three distinctive oats that we need to need to have. First is: do we commit to a run-time deprecation of Bernama or any variation of the buffer constructor? That's not required by UN, say to Ray.

I

And we commit to that at a particular time frame so either. Yes, we are going to do it by nine or ten or no. We are never going to do that.

I

That's question. One question two would be you know based on the result of that. Do we put in the pending deprecation for it, which is the conditional deprecation for, for my one PR that I have questions, three is: do we random fill or zero fill by default and a sub kind of a sub bullet point to that is about back back boarding, so multiple decision points and each one of those we go ahead. Sorry, each one depends on the other, so I think we definitely just start with the question of.

I

Are we going to commit to deprecating buffer constructor or no.

A

And I know in the we in the definitely yes, we should do that or at least I think in the definitely s we should do that camp. We have.

A

We certainly have Nikita and I, think Brian 2, I'm not sure, though, um but now, let's put everybody yeah way and yeah and then and then yeah actually then and then and then and then I definitely hear from people who are like no, we can never do this, which I'm sure we will so I'll stop talking now, but somebody take it away, deprecating committing to a deprecation in a future version. I.

H

Mean we can do it. It's come back to bite us.

H

You I, don't really want to deal with them. The.

I

Ecosystem is very eco system in general, based on all the feedback I received from module developers pretty unanimous boards, don't do it don't duplicate!

I

K

Well, there I mean if they have that kind of a response than what is like their suggestions as to how they would deal with the security issue. What not and where they's early.

A

I

K

I

Default that they would be willing to take the performance hit on zero field, but if he's considering, we have the buffer malecon safe is an option that allows them to get to performance spaces. They need it.

C

I

And majority of the the folks that I talked to also didn't even think that we should even backcourt it. Let's just do it made on Ayden forward and they would deal with it. You.

A

A

Wouldn't help just take a straw poll right now about I mean I. Guess we have that mmm buffer options, private spreadsheet, that we all filled out but um take a straw poll on.

A

Because because there's there's deprecated in there, but that I think meant deprecated like this version, but I, don't think there's anything in there about. Oh, there is scheduled yeah so.

A

Yeah, um it's wow. That's that's tough! Yes, that's like a near 50 50. It's not a 52, just split, but it's awfully close.

A

With a lot of people expressing no opinion.

L

Sru tallying the spreadsheet.

J

Referring to the spreadsheet.

L

It was from before the option of of a additional flag was put forth. I. Don't people have the same opinion if there's an extra flag for further on time for the replication origins put together.

A

So there's the opt-in deprecated, which is like the second thing that on James's list of three things to talk about, but there's also a column for just scheduling, a deprecation. At this time, just saying we're going to deprecated in version 9 or version 10 original 11.

C

Or whatever and.

A

A

Yeah, there's, you know, there's there's a significant number of people who feel we should never deprecated and there's a significant number of people feel that we absolutely must deprecated at some point.

I

At this point, I think it's pretty obvious that we're not going to reach a natural consensus and we're not going to find a solution that makes everyone happy. We need to find a compromise solution and basically call for the vote. Our rules, a simple majority which.

A

Awful plum, I would second a call for a vote at this time. I don't know, there's really much more to talk about.

H

When what are we actually voting on them, like is just like multiple choice, so go ahead. No.

A

F

To really deprecation in some.

A

Future of our own- okay, yes or no like- is that something we want to do or something we do not want to do.

I

If it's helpful, I can type a specific proposal into the chat here in uber car.

D

For example, hundred achieve when the ecosystem music pours down, you should hopefully would I.

I

Mentor edit quite a cattery tested negative, I sent.

D

That you can do that when the existing usage will go down. We should hopefully would if we use the option flag.

H

I think that's doubtful. It was fine.

H

I

I

Just type in the proposal, one second.

Q

A

And I guess yeah the proposals about. Do we want to schedule one at this time? Well anyway, I'm going to stop splitting hair, so I'm just going to let this let this happen.

D

You mean the rough schedule on just saying that really implicated in some future version. Oh.

A

I say so so so so we could.

A

So so we could also, instead of saying that we intend to deprecate in a specific version. We could simply just leave it it we intend to deprecate. Is that what you mean right? Let's, let's.

I

Let's put those comments with those questions: let's say: if we are, are we going to commit to the deprecation in the future? Yes or no? If, yes, okay, how do we set that date? That's a separate question: let's handle it separately,.

D

I mean my question: is: when do you set the date north? How about when we say it, but.

A

And I think this this this proposal assumes we're going to set set a date now and then and then, if this, if this proposal passes, then that's what we'll do if this proposal fails, that we can have another vote on a proposal about just saying we're going to do it at some point, but we're not setting a date right now.

I

A

N

I

Just say: are we going to commit to doing this to to deprecating and leave the question of men and house to a separate question? You.

D

So what are we exactly the world in for now.

A

Oh James this proposal, which one is this is this: is this we're committing to deprecating? We just don't know when right now we may or may not know whether, yes.

I

At a specific point in the future could mean, when you know at a particular release version, or it could mean, when usage drops below a certain point. This proposal is not clarifying how we determine what that specific point in future is what.

D

I

Committing to deprecating it at some point in seizure, definitely or not, you.

D

D

H

I already wants to vote on that then.

A

So my girls record it in the dock, so I don't know. Maybe it makes sense to just type your name. Let me type your vote in the in the sidebar chat or actually.

P

A

In the in the Google Doc, actually the most sense.

Q

I don't see Oh.

I

The minutes talk minutes talk with what you men right.

A

D

So how we have to do it over there.

A

I just created a yes kala, yes line and no call an O line in the stain line and just type your name on the line. Okay,.

D

D

H

And so far we've got 44 propagating at this time and.

Q

Seven for not coming to dedicating this time and one stain so.

A

Wait a minute I mean now, I'm now confused about the proposal because we're we James we're talking your proposal here is that we commit to deprecating at some point in the future, and we don't we may not decide what that point is for for for many months. Is that correct, or am I misunderstanding, that.

I

That's it I mean how we set that specific point in the future is a separate question, and it's no proposal on table for that. Yet it's just are we going to commit to deprecating it at a specific point in the future? Okay,.

A

So should we so basically this this could be read as should, should we deprecate at all very much.

H

P

H

If we're not committing to deprecating like at some point, commit to deprecating right.

J

It's it's it's more precisely. Are we willing to commit at this time and no sense, I'm.

H

Not gonna say that they might not deprecated this like two year, two three, however many years from now like proposal.

I

H

D

We going to announce now that this is going to be duplicated from time to placate it at some future point. Yes,.

A

Renounce it at some point and we do not we play what that sometimes right now later is a site is a separate question.

P

L

J

Many people do, we need to get pass.

A

Now you don't need a leaven votes, ah but abstentions read. The number goes live stats. Now it's intense.

O

J

So we're so much there there's.

H

A couple people that haven't voted yet.

D

James I alter the dog beach. Is this correct.

F

B

Seems frightened me yeah.

P

A

Would say possibly without defining that tuck that time right now.

I

A

Looks like those.

I

What this is telling me right now, since we don't have everybody, we need to take the vote to github, give the other CTC members who are not on the call, a chance to weigh in my proposals that we give them 48 hours to do so. James.

O

H

O

Does this is this saying we made, will defecate at some point in the future and we may deprecated the entire thing or we may deprecated. Only the non un aid array.

I

This this says that, right now, as of as of today, we are not committing to deprecating with upper constructor home at all or or we are saying that we will deprecated it in the future. The height.

D

Of the question was all.

I

About aging accepts those reports, all combinations except those that work on required by you and data rates. So it's just it's a statement at a point in time we may have come back.

O

As one at this.

I

Hour a year from now well.

O

Like I, don't want to say yes, I'm, okay with deprecating, if that means that, in the future we're going to be deprecating, the entire thing period, I'd be okay, saying we're not going to deprecate it. If we're going to deprecate all non you at eight array, it's.

I

Specifically, all non-us eight array: okay,.

J

It's called out and what's written there right, so this is deprecating all combination of art for constructor that are not specifically required by subclassing. You went a terrain.

I

I

So- and it be absolutely clear- a yes folk here says, as of today, we're going to announce, we will deprecated.

I

No vote says we are not going to announce that today we might decide to do it again later at some other ones, future / not made to it today.

O

Either way we're going to be talking about buffers for a long time, yep.

A

That's forgive everything and, and and we're not we're not done with it today so I'm so looks like on this is gonna have to go back to github or private email or something but pride yeah. That's right better, um but I will try to get that squared away in the next. You know as soon as possible. I hope it doesn't know they're there.

I

Is a second vote right.

O

The second to code is.

I

Very specifically, do we do a conditional or pending deprecation warning for my PR for buffer? This is a and all in all respects a normal runtime, deprecation warning. That is only printed if the pending deprecation flag is used.

I

We can, we can choose to add this independently of whether we ever is to do a full, regular, unconditional, runtime deprecation, which is the first question right.

I

The intent here is to give something to somebody like running in CI or a test environment to use to catch these easier.

I

But not make it noisy for ordinary users. The argument has been no one's.

L

I

Use it, which is you know of essentially a valid argument. We also have other solutions like limping and those other social engineering approaches. We can take.

J

And if we do add that that would be a general flag, which is you know, people could use for other pending deprecation type things right. Yes,.

I

Yes, the pr introduces the general mechanism and specifically the thing for buffer. Now we could choose to land pending deprecation anyway and just not do the ones that buffer does a separate question.

J

Was there any discussion of making this like a warning, as opposed to the deprecation specific, like I, think it's we'd like to tell people don't use this right, yeah.

I

We could we could make it just a regular morning. This this particular boat is specifically about whether we make it a deprecation morning. I can go back and change that to be just a regular warning. Yeah.

J

I I mean it's like yeah, it would do, but those warnings there's already a way to opt into warnings.

I

Yes, but its particular one, we may want to eat a little special because they're the performance impact right.

H

So I'm not really sure that it matters. If this is a deprecation warning or not, because the crux of the thing is: does your nose print.

P

H

Mean feedback but some sort of notice. Essentially our warning when you use the flag and buffer to Kruk is used so whether that is a location or if it just prints, don't use this, and does that really matter, I mean I, don't think so? Maybe it really does I was.

J

Just exploring whether making it more general made it any more palatable to people.

J

P

But let's not get site. If that's not doesn't resonate with anybody. It.

D

Makes sense to enabled edge for something gauges already Dobbs duplicated March, not for everything the ox's be granted, for example, degree usage in cases where we are going to run time duplicate something soon or in cases where the ecosystem music is just to work for an interpretation by for the buffer theme or in some other cases when do just to want to multiplication more visible, but don't want to turn it on by default.

D

I

I'm, actually I mean it's mighty or but I'm actually torn on it. I think I need to know people, but at the same time, I do agree that very few people will actually knew it. Actually, the flag boy.

A

What would recognize us we're going to use it, but we don't know until we try it and my support for this is around the idea that we have to you know, wait. There are no good options and we have to do something um and I'm I'm. Definitely not not in favor of we have to do something in there go get up there. No great options, let's at least try this. You know um that's kind of where I'm coming from yeah.

J

I think I I agree and what's the downside, if you know okay, nobody uses it, but have we really harmed anything in that case, so I don't see much of a down side, so it is worth trying.

I

There is a definite performance impact if there is exhausted. It looks like adding that in there and the reason has to do with this workaround for fast buffer and the best fast over constructor.

I

Basically, it does it as an additional case that, as we looked at so that if the constructor is called directly, then the morning is printed, but if it happens to be through bulbs for like map or filter or women or mesh or the other ones from.

D

You and date array it.

I

Doesn't show it and it requires this kind of a half work around the performance impact of that is definitely nonzero, but.

N

You may need an impact, even if we don't enabled the flag, yeah yeah, but could.

L

N

Could you make a different constructor based on the fact that the flag is set or not keep the current one if the flag is off ah I.

B

Think that will work and I think you should do that. He'll.

H

Q

I think I, a PDF icon.

H

Q

H

Is that if you really, if you really want to get max performance out, I mean you should probably be ahead of the game anyways and you should just use like you need to be, then careful and just use a lock on safe, like if you really want to squeeze every inch performance out actually.

L

H

Have right cut internet.

L

Oh cool, we don't have to cut the palms any further, because I think that suggestion from Michael is pretty good. They shouldn't be a performance impact.

D

Yes, then, there should be workaround for the home stretch here. If it's just the new coach is enabled on you in the Fergus best. Alright,.

I

I will see what I can do to minimize performance impact. The I have to make sure there's. No, there are no other side effects in terms of I, wouldn't want to end up having a December major style change between whether it be flag it on or off. You know, I don't want to see any kind of behavioral differences with buffer. There shouldn't be, but I haven't actually played with that variant. Yet so let me prove it. Let me demo and make sure it works and then also paid to PR. Accordingly,.

I

How we doing on necessary month, seven, there are two abstains that brings us where we need nine, yes, votes and motor Ford's pass for at 127 of them. Right now,.

I

Colin evident in Brian, do you guys want to say anything here? It's been as the no votes.

F

I feel a lot of because that has been asking us to not.

F

If there's going to be like a performance impact like that's kind of what, from what I've gathered all mean, you should tracker anyways.

E

So I'm against having a flag to have deprecation zon in like two different styles of deprecation Xin, the code base. It just doesn't make sense. We should either print all deprecation, zor, don't print any of them.

D

There are some deficiency, for example, the Jermaine application that doesn't make sense to actually stuff on that.

D

I

Alright, it looks like we're not going to get enough votes on this fall. So it's again, let's take it to github, get additional CTC members to vote, give it 48 hours.

A

So so I don't want. You said 48 hours, on the other words who I want to point out that there's no such thing as a time limit on votes. We have to get you.

I

A

Yeah I'm proposing a time limit. We can hang on it the violets our flights are governance, but we'll talk about that after the meeting.

D

He my time image do me to count everyone who is not to who has not voted by the time. I was up stained, wood or something or other.

A

You, but there was no process for this to Nora Nora. You know, um but let's, let's, let's move on yeah.

J

I think at this point we have to get enough votes to go one way or the other. We can't just just sign them.

I

Yeah: okay, okay, third, third vote.

I

And this is there's a part, a and a port betta fish part a is: do we switch to automatically fills, but for now I'm a new buffer numb.

I

Part B is, is that a0 fill or random fill.

A

And then, of course, to backport or not to backport yeah.

I

And that's: let's make that vote for.

I

Right on Election.

D

Day, the google doc is having its role right now.

Q

Wait: um okay, Mars Indies.

I

So this would be zero feel from 80 on we'll do a little 44 back for dinner before.

H

We do a boat for back porting is that? Is there anyone? You don't need to explain right now we can get into that when we get into it, but is there anyone? They would actually object to back porting.

D

I

D

I

Feedback and speed from one of the bell first, there are quite a few who would prefer to not to back for you not.

H

To okay, I'm, very surprised, ah I I.

J

Could see the the argument that you know if you do it an eight and there's a performance hit, you know you can upgrade its it's a major. You should upgrade to the new api's once you're back poor it. It now means that you're telling people they have to do work back and earlier releases as well.

D

Once they passed a or drug world, I think that.

Q

D

Something out then watch a CD to a google doc right now.

I

A restart for this in the google doc just a slight bit, instead of doing the a B and yes not abstain, I'm going to put it as.

I

Zero random nitre Alex.

P

N

J

You James how much of the differences are between the zero and random fill and four terms of performance. It's.

I

Not it's not significant, depending on how we do it and my current Pilar does the allocate and then a separate fill which is slower. Does it all in JavaScript, because it's a computer number of code changes we.

P

I

The grand and fill with the memset right after allocation and the performances and difference is negligible. Okay,.

J

Because the I just seems there's some comment about fifty percent, so.

I

But just switching to filling by default, whether it's zero or random, for the significant performative.

I

N

D

I

Between the two choices.

D

Just for, like a huge a note, my current role depends on the code, one which I see is not forget to pass. This is why I motion neither here.

A

For the more possible for someone in this in this particular vote to vote for both zero fill and random Phil, if someone is there resent, is that right.

I

Voting for both would count as a vote.

D

I

Of automatically spilling and then we would separate those 20 field, random chills in Sweden right on the first part, you.

I

I

A

A

This one, it's really hard for me: I, can't even decide if I want to abstain.

A

I

Looking at the boat, so I think we have to say yes for filling by default in me.

A

You know I, don't think we quite have enough. Neither two more not do. We need. Okay, oh.

I

That's right, Epstein's bullets.

A

Boy down yet okay, but it's awfully close to passing.

J

Hey did we need wheat, we need dirt like there's 13, if you add zero and random together right.

I

Are voting them.

J

Oh yeah double counted. Sorry I forgot that part.

E

I

Those of you that have voted in both will need some indication of which, which is your first choice, which is your second I I.

A

Think I don't well I, don't I, don't say random. Fellas.

D

What that would be needed, Joanie for both places.

H

Would not only be applicable if those two actually got the same number of total votes for each like. If there's no contest, then.

Q

I

Yeah I think so I think it's this point. It's pretty clear. We just have to make sure we have enough votes to be able to call it.

A

All right, it looks like everybody with me, I think his finally weighed in and I'm gonna weigh in later on, get up or something cuz, my boat's not going to tip it won't might have at this point. Okay,.

I

And then there is the fourth of.

M

O

I

Fill do we back port, it move it backward dis to seven six and four, maybe five, but we're not doing any releases on five. So that doesn't really matter if we did backward 27 able to just be until June.

D

My back porch in do mean by pushing to all the rations I'm in 20 branches or just your.

I

Orderly clearly range Oh all released four inches, so we'd go Landon, 4, 4, X, 6 X to the next.

D

The clothes on my interest.

I

And it would be December minor in each.

H

Shocker I just.

D

Want to say that I might have different opinion for the coaching to just 26 or to both of them.

H

Right yeah I was going to ask if anyone had opinions about. If there was specific branches we should or shouldn't push it to.

I

Fact or tall order, yeah I would say um no I say that we should be. We should definitely break it up, you know, might make sense. It's back toward it to 76 illness might make sense to spec for it to seven. So we should break up the end of the questions. Backport, yes, or no to be back for it to all release branches.

I

2764 six step or whatever combination.

D

D

I, don't think the edge made it better for me because I'm against back porch in each 42 version, just to version 7 our direction. Seven and six by checking that I will abstain on the poster image for version. Seven. Six and four.

I

So should we separate it out in two separate votes for each really fun.

D

J

So we're more like a seven six and a four like that.

D

B

J

Now your tree, are you want the only if it's all of them right.

D

I think that I will explain on your fridge was over then and 10 pitch much against, but fortunate to a subset of the devices.

D

D

D

To be clear, my.

L

D

Some sort of haha to clear my opinion is.

P

D

Partial back porch is worth then not back portion at all or then the torsion to all the branches.

D

H

I'm not actually sure I'd agree. I prefer to just see it backward to at least some, even if we really aren't going to do all.

H

I think we should just back part the hall.

H

The new api's are already in the things if anyone really wants to squeeze every last inch of performance, so they should just use a lock on, say, finger.

H

H

Or if they're, not I, guess initializing an empty buffer, just buffered up from.

H

I

Okay, so let's look like again just: we need to go back to get home to collect the rest of the votes.

J

H

You cut off their.

J

I guess I was just saying: I was just saying we're already over time, but do we want to move on to the next issue on the agenda which was related but different.

J

Okay, we know who added this one to the agenda.

H

Wait isn't there only one agenda line looks like it was rich, but.

J

Though the sucked the next one was 95 31, which was discussed future a buffer constructor API, it seems related, but was listed separately.

H

Think it's all tied up in the same issue and say: isn't this what we just finished discussing pretty much parent.

D

Who chose discussing days.

H

Anna is the only thing that we must.

B

um I, don't think so. No yeah, okay,.

I

At least for now, I mean it was going to be other conversations based on the outcome of this later. On he's, like es6 class.

I

Things so, but those are I, think balcom of those depend on me out from these those yeah.

J

I, you know I think closing on this. First one makes sense and then moving on.

J

I guess next thing is: if there's any question and answer on the.

J

H

Can someone go over the meeting times.

J

Sure so the upcoming meetings I can always check the foundation calendar, which is in the link. But the collaborator summit is upcoming, May. Fourth, fifth, so if you can try and see, if you can make that there's a CTC meeting which is marked as March 30th tomorrow, um I guess the clutter.

J

Oh sorry, that's sorry I've misread, that's the call, there's a call with the v8 team, which is tomorrow it's about 12 hours from now and then finally, next week, there's the the CTC meeting and I believe there's yeah next next week, there's a CTC meeting april fifth at the regular time.

H

Okay, so someone asked our questions only related to current ctc I'm. So I is the questions that can be asked. I.

H

Saw a general literally questions can be asked, but if they wanted help request, so they should put that in the help, repo I'm, not sure. If anyone else wants to clarify or if there is much to clarify.

H

You someone else asks which database or I'm would you suggest for SQL database I would say that this question is probably better for somewhere else, but I don't know if that is the help, repo or stack overflow or what anyone else would suggest.

H

I

Yeah I probably say you know, and at least ask a nun help if somebody specifically looking for mature and rescues with no note and the help, repo would probably be better.

O

Lots of all facilities.

I

Javascript developers are definitely enjoy having their different depends.

Q

D

Q

Questions I got there.

H

Other in the comment that says the person with cherry MX, blue key switches should meet or moves to MX Browns. Someone has a noisy keyboard but imma dance, I. Think multiple of us do probably.

J

All admit, mines noisy.

J

H

All right, I think.

J

That's everything.

H

I think we can call it okay, great thanks. Everyone for tuning in okay, thanks.