Rust Programming Language Language Design Team, 18 Apr 2019

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: Lang Team Meeting 2019.04.18

Description

Lang Team Meeting from 2019.04.18

Paper link:

https://paper.dropbox.com/doc/T-Lang-Meeting-2019.04.18--AbebQCdj8aVOFCU7vqueWouKAg-TBw9YD58Ciy431nNQ1Y0d

A

Now, welcome to.

B

All right, let's see, doodle pull oh yeah. Let me share this thing too. I.

C

Apologize in advance, my computer is sounding Jackson I need to clear it. I.

B

Don't hear it nice.

B

That work yeah.

D

It looks like some of the Wednesday times are pretty close, like 112.

E

Based on looks like possibly either Nikko or Josh is meeting them. Maybe one of them should we suppose one of those and then no it's.

B

Plausible yeah I can take a look.

F

That particular context is I'm, really tired of having a meeting during the one time of the day. I can have lunch. So I'll reschedule my lunch to dinner time.

B

Whoa, what is I shared the wrong thing, whatever no.

E

Way no I'm, like my screen, is still I. Don't have access to my computer anymore. I only see all your faces.

E

Here we go, I fix, it is six.

E

Josh, can you have your lunch from 12:00 to 1:00 I, wonder I've in terms of like Wednesday times.

F

Can we like this would be a discussion better had asynchronously.

E

That's probably true, but I just read it. The most promising time on the schedule looks like somehow there's these two Wednesday claims that your job you can't make. Why.

B

Don't you send out an email with with those two times and I can comfort them.

E

Exactly all right, so Thursday I have are off and have evening obligations that everything yeah.

B

I would find it hard to make that time. Personally, we're.

F

Pointing out if you're talking about Wednesday one two, two, that is on top of the car, go meeting.

E

B

Looking at the we had, we did a little pre triage and we went through the pie issue, and so this issue, I think we we were probably you might want to push it to. The main topic depends on what we want to cover in the main topic. I have a few thoughts, but it's basically a stability question that came up in the compiler with the backwards, a kind of cleanup change that has some effect on existing code.

B

D

Is the stability more.

B

D

Indexing with you size, suffixes and.

B

So we won't do it now because it seems unlikely we'll cover it in my rule, of thumb should visit with these triage things. If we can cover in five minutes great I, don't think we'll cover that in five minutes, um so one of them here is five one, four, nine seven. This is kind of an interesting one. It has to do with some quirks of method resolution and whether, like it, basically it comes down to it.

B

If you, if you import a trait by name- and you also have a glob that pulls in a trait it gets shadowed. If you just reference X by name that you always get the one, the explicit import but method resolution doesn't respect those precedences and and Claude was proposing that. Maybe we would want to change that. There are some reasons we might, however, I think in particular.

B

This is an interesting one, so it like the shadowing rules enable us to add things to craters and so on, but but the method resolution makes that sometimes a great thing to change. However, I think it would probably at least technically be a breaking change, because it might be that the met that there is no I don't know we don't base achill II my fall on.

B

This is if someone wants to see this happen, which I'm not opposed per se, but we would need to do some investigation sort of looking into the rules trying to think of corner cases and really lay out all the stuff and that's what we should have, because anyone wants to do this.

D

Proposal isn't making new things breaking changes right. This will only make more code. Work right, I think that's not fair to either.

F

I get how it's saying that it will make future things possible without breaking that would today be breaking, but it's not these things.

D

Will still be breaking in edge cases right. This is just making more things, not breaking.

F

Not obvious that this change itself doesn't break things, but.

C

It makes adding a trade so.

D

Any any this is adding a resolution for ambiguities that exist today, where we error rate I, don't think.

B

That's I, I, don't think that's entirely clear. I think that's the idea, but we would have to.

E

We would want to yeah right.

B

We would want to like dig into that a little bit and also like sketch out what that would mean if I would I think that you might be able to make that happen. But then it might lead to there's that's kind of where I was not too investigating. So.

C

B

C

Think the report and some patrons could be nice, so many informations who I mean.

D

Certainly, if it's implemented correctly, a career run should be clean right, cuz it can't it shouldn't be resolving. It shouldn't be doing anything to code that already compiles. It should only be changing the behavior of code that does not yet compile right, sure, let's check yeah, yeah, yeah, yeah, I'm, I'm, sorry, I'm, just making sure we're on the same page and that I haven't misunderstood what the features I think I.

B

Think you're right that would be the goal. Does anyone want to try to do.

C

That which I looked like thousands it, but back to my saying like do you want to try.

B

D

To create a run, a PR, adding this feature right, that's what we do run and then we find out if I'm.

G

Also I'm already working with Matt cloud on some next name resolution, so I'd be a semi natural choice. I try to at least assist okay.

B

I feel, like the only thing caveat I want to add to that, is.

B

Even if we get a clean run, I wouldn't miss, like that's a good break data point, but I'd like to consider it further after that. Yes,.

D

It's necessary, but not sufficient for.

A

Us to limb.

D

B

We would want to consider the overall effect.

B

Sounds good, okay, I'm trying to remember what I had a I had a corner case. I was trying to write about when I close my tab accident I can remember it later all right. So next one so this issue, four seven one six eight has to do with hint expect the hints we used to give. So if you get a type error like this program, you have a you size here and you expect to you 32 I guess when they filed the issue.

B

We would give all this help like a bunch of random methods that almost.

F

B

Did not want to call, but now we give no help at all, and there used to be some code that suggested that you do as you 32. However, that is about, because that will truncate and not to you what you actually.

C

Information here is that we have to.

C

Try from the conversion rate right.

E

So I really I. Don't let me go with you if I go and rapid be a better choice than I think you, you could have.

C

The context is is like: was the method mixed or no result? I mean that is just trying.

C

B

Could certainly just suggest that they look into the trying to treat or something but they're, not actually. No one actually asked us give our input on the situation, but you're welcome what.

H

So we've already recommended the try into trade in the past back in January there was a discussion about that.

B

Right, so all the extra banks- it is, what is the languages long term deal and I think the answer is it's trying to just correct. We won't have any other plans that, during this I.

E

C

E

I think, as is for hire against as anyone elses I think that, rather than a meal, so okay.

F

I mean the third obvious answer would be. The right answer is the same for link to take sides right.

B

It's not clear sort. Presuming the other types are correct. Trying to is the preferred way to do a conversion. Yeah.

E

So yes unwrapped now.

F

Try into and manhandle the error I don't.

B

Know that we should suggest a specific change, we might just suggest look into the trying to treat and let them, if they're, trying to try to give more context, but I think I think we.

E

E

C

Try a famine, those things might be relevant, but only.

C

D

Is a stupid question, but does anyone know why the error type for try from you 64 for you size on 64-bit platforms? Isn't bang is that is that just because they one of the types to a little maximum portability.

H

Lint didn't happen, so you don't want the same. Always. Okay,.

B

So for the nominated RFC's, so it's someone's gonna.

H

B

Like this, alright.

E

Everybody I can write that okay.

B

Scott berated.

C

So we'll write about the other one solution fit the right, Scott Felix and this you are involved. Maybe you can write something. Yes,.

G

Yeah sure yeah, yes,.

B

Yeah so nominated RFC's put two five, eight two. So this this RFC I would.

C

Like the other one, first.

B

Two five one: five yeah.

C

So have you thought anything about this good.

E

Yes, well, I am actually in general,.

E

Optimistic that I'm, just everyone whose we're not speaking, because someone has nothing.

E

Like I can't think well when I'm my own sentence is few because I got back at me. That's all so. Yeah I actually have been my actual mission, and this is that I've been coming around to it and I want to mainly like think about the objection that I had before and write out. Why like be able to justify? Why I?

E

Don't think that anymore and the main thing is an issue of focus and being able to focus on it and I want to some time in a few weeks, focus on it, because I feel much closer to the consensus, but everyone else has on it.

E

It feels like a good way to work on that experience prior to working like dealing with the awake, syntax question and so I think that possibly they're concerned I registered I will remove without any further discussion with anyone else, once I focus on it, which I was possibly earlier this week, but then didn't and instead did like for await, and so sometime the next, like maybe next week and the week after I hope to essentially go through my top process on it and possibly I'll. Just like write a post which removes my concern.

E

So that's my status on this RC. That sounds really good. So.

C

I think that's top.

B

Five, I muted okay, yes,.

C

F

B

We'll check back in in two weeks otherwise sounds great. That.

F

Sounds wonderful one quick note on the other item, the async/await, without going into depth on it in this meeting at all, the only comment I was going to make was I saw the for await. Post I was really excited about the idea of that syntax.

F

It sounds really interesting to me and if you would be interested in having a sounding board or discussing the idea of ways that the pattern like syntax could make sense with question mark in a way that doesn't have the right to left left to right problem, I'd be interested in having that discussion in order to make sure that they're, a reasonable alternative on the table sounds.

E

C

Some of my thoughts on being transferred yeah.

E

I need to reply to the internal threat. I'm gonna try to do that tomorrow, because.

F

Also, there was like.

E

Ross Johnson's restaurant, whose post about.

B

You might comment here: does it sound fine, live, live update all right.

B

Okay cool so to come to this back to this one so to five a I don't want to discuss at this moment, because I think it will take us more than five minutes guaranteed, but I wanted to propose that we should schedule the time to talk about this in more depth and try to have Ralph here, for example.

B

This is the operator around for all references, and at least I was surprised because it completely changed from the time when I looked at it to the next time that I looked at it because I guess there was a lot of conversation in the meantime. I'm not sure that I disagree with changes, I haven't really thought about it, but seemed like it might be useful for us to all get on the same page or at least talk about it all together. So.

C

This I've seen might change slightly again. I would like to do crater on and test whether we can have at sorry, Amazon, draw and X and expression without having monster or mute that that needs to be checked. The Creator different and I need to be, because apparently this is not possible for some reason. Okay,.

C

F

Parser subtlety of effectively a contextual keyword row next to ampersand and seeing if that would break yeah pretty.

E

Much I didn't really serve the syntax and if I was starters, the change that I was referring to what happened here. It's.

D

Certain, it's certainly will break if you do it. That's W right here all right, but anyway, like it seems like I'm, not sure.

F

D

Can tell me, why does it am wrong and not star raw.

F

Well, for one thing, because the dereference operator not to take the interests of all operator so.

B

As it I think, that's not a stupid question that was actually proposed by gob or initially I. I think this is exactly the kind of thing that would be good for us. I would be fine.

D

Because I, don't think of em and presented as the address of operator and rest well.

F

I think like star is definitely the dereference operator not recreate a paper from the operator. I haven't.

G

Met I agree: yeah go ahead. Does anyone else feel like it was sort of mistake for this RFC to be rewritten this much yeah.

C

G

D

Because commentary is not correct.

F

Would have been a good idea to simply this RFC, or at least postpone. This are at the in favor of a new are out circles in I.

C

Mean the commentary that exists: it makes sense to explain the change. I don't care. Well, you can refer to other RFC yeah. So.

B

Actually, this is interesting. I want spend years I, but it is interesting that a point that is pretty important, which is that, like I, think we should be moving more to summarizing and not requiring people to read all the comments to know sort of accumulate the history of the issue, and maybe this would have been this would like pace for. Let's summarize, what led up to it hidden.

H

Comments, it would be very helpful, yeah.

D

Note that summaries get hidden as well, also like the by summary I meant urs opened.

H

C

Having it in the text, so I think it's really useful to just be able to link the text from top to bottom and understand.

C

E

The original RFC, which is not about surface in text like it just in terms like an MVP like, is that like something, is like a subset of adding the syntax and possibly because just like movement or with that and then make a new super sentai receiving syntax, and that could be like something we discussed separately from the original like core proposals. I think.

H

The original one semantics.

E

B

Think we should talk about this in this scheduled meeting yeah we're only at four minutes from when we started. I have a clock, but this seems like a good yeah I, don't feel, there's total consensus that we want syntax for one thing: yeah.

A

So I think it'll be good.

B

For us to have a meeting to talk about the issue, what are the pros and cons of those things and go back over it unless everyone feels they all remember all the context.

F

Review of this with what were we trying to solve, and why did it change to surface intact? Okay,.

B

C

I can answer that question now. If you.

B

Or any other questions, I'll.

D

Talk to Ralph, we'll.

B

Put it on okay for a future event, in fact, I think.

H

B

For me, I think.

H

To the can we make progress on this now, the the compiler to you may change to put this in mirror and stuff like that is probably going to have to happen, but also doesn't need RFC's. Definitely.

C

B

Here regardless regardless well, it needs to fight beside mean you can put it. You can put it in here. You.

H

Can put any mirror and not expose it.

C

That's just all right.

B

Back to this so good we're moving on five minutes. 38 seconds done.

B

There's a lot of unassigned RFC's.

B

We didn't really get to those in our pre triage I, don't know if any of these feel like things that people want to jump on.

G

I'll take the script bits. Oh.

B

C

Quite hectic all about the discriminate sites and how's the news, but not I,.

B

Don't know what it is.

B

F

I talked about this one. A little bit hang on. One second I need to actually get it open. It will open so.

C

I signed it to Felix right.

B

Yeah, why don't we? You know why don't we hold off on talking about it in too much well, I, don't know, do we want to so what do we want to talk about? We have we're coming in on thirty minutes, which is why I'm trying to decide do we want to go into I? Think we could.

B

Can you briefly discuss what we will discuss with the main meeting? I want to throw three things out.

C

But before we discuss anything else, I would like to quickly go through the lead, Rock, semantics or unused arguments and Mason Blake.

F

F

Today, I would object to the use of the term quickly, because that will be the discussion for today. If we start that discussion, that's.

B

My sense, the I have three things that are on my mind. One of them is discussing stability, I, agree that that I think that would and buy stability or what I mean is I've seen a couple of topics basically are seem to be arising. First off, we had the NLL challenge like coming to consensus around what to do with you know well, but in durban, and then there's this other question mark this question about the tuples and in general. My sense is that stability seems to be a rising source of conflict in it.

B

It's been a while context has changed since one dot o, and it might be good for us to talk about whether we want to revise our stability policy. If we can all come to agreement on what exactly the policy is, future compatibility, warnings as part of the story and so on. I, don't expect us to come to agreement in this thirty minutes, but I would hope that we could better understand where we are and think about.

B

If, if and what next steps we might take on one thing, the other would be I want to talk about. The FCP process and how we could integrate it with meetings in particular to talk about how we should handle blocking reservations, which also I think have been a source of a lot of conflict lately. And finally, there was the possibility of discussing days in Kuwait status, update and issue of draw parameters which I think we should discuss, or we could also have.

C

It so I would suggest two things we discuss, Basin weights drop and then we discuss the absentee process in that order.

B

What is your motivation for having that honor, because.

C

I want to get through the unused stuff, because it's.

C

F

I feel like that one is not going to become a critical issue until it's the last thing blocking that and considering we still have to sort out the syntax I. Don't think it will be that big a deal if that one waits a bit I would personally prefer.

E

Asking the process.

F

Proposal and then the async/await bits.

E

The drop thing: if we decide to change it, I think that it is not an easy implementation change, and so it's something that sooner than later, it might be beneficial. That's.

F

B

That you're one of the complications was implementation. There actually was a lot of discussion, I.

C

B

I generally agree with judges. Take that I don't feel as much priority on this as other people seem to in part, because if it's not the last blocking issue and in part because I don't know why the other part is I, guess it seems. Okay, if we talked about it next week,.

D

It's not the last blocking issue. What you considered to be the blast blocking issue in tax I. Consider.

B

Syntax tree blocking I consider the region entrance stuff to be blocking.

D

Which? Which region.

B

D

Stuff are you talking about the whole lifetime's thing? I am okay. Well,.

C

If everything that is Diagnostics related seems looking at least maybe not generators and.

D

Newcombe I still I'm still waiting around on you to do some research for the.

D

The lifetime resolution thing great.

B

Maybe let's talk about that, have fun I, think I posted some notes. I could do more. I wasn't.

D

Sure I wasn't sure. Last time we talked it sounded like you were gonna, do more investigation but I all.

B

Right it wasn't so the one thing: okay, the one thing I, would say, I'm happy to stop talking about we're. Gonna talk about I just talked about, but the thing I was gonna say: is that I think the ICP as this might be short, but I guess that's probably.

H

We can we talk about.

B

One thing: yeah problems.

H

F

I heard the echo I also heard what sounded like a scream in the background.

E

F

C

Should we just go through the stuff.

B

C

E

Wants drop even whatever we start talking about no one's gonna like stop meeting till J. Let's just talk about whatever.

C

So so, how so I think that I'm very much inclined feature at moments? That's my personal view, but it allows us to take your time so.

B

All right, everyone knows the context here and.

F

B

C

Did you say I said.

D

Yes, we're from the context we discussed this in the last meeting just.

B

Make sure so I have my opinion is that I am I, find it pretty hard to imagine that the right thing to do is not is to keep the current draw order. I would certainly be more amenable to feature gating than stabilizing eyedropper order. I think is sorry.

F

There were a lot of there, a lot of negatives. In that sense, were you saying you don't believe? The right thing to do is to keep the current drop order. I think.

B

What I'm saying is I think we should change the drop order. However, I might be amenable to feature gating. If you know I would rather basically I might be like the weekend to feature gate.

F

What the motivation is for changing the drop order, a given that we have never had a semantic specific definition for the drop order of async functions. What would be wrong with keeping the current order.

B

My argument would be that I think that if you went through and took a bunch of code and changed it to use a put a sink on all the functions and then what you say is that is.

D

That a real thing that we anticipate people doing like going around an existing code base, that is amazing and like async affine functions, I mean.

F

I, wouldn't care about draw order when they do I'll.

B

Give you example where I have a ones do this and B I would care about draw order, which is their salsa, currently ditches? The compiler incremental compilation, library is currently written in a threaded way, but I was considering porting it to use a significant salsa has locks that it holds, which are not data. They don't actually protect data they're like logical locks and they would get dropped, possibly at the wrong time. If I were to go through and make great transformation.

E

That make sense, I found all the arguments in favor of changing it very compelling and they every time someone puts to make MIT's favorite like yeah. That's a good point and I feel like we should change. I I.

G

Have a very quick question about this business: a future meeting, if you don't mind me interrupting um so, if I understand correctly, if we the drop water as it exists, then we bypass actually make sure you use all your arguments. Is that correct.

G

Unless you'd always get the worst case scenario of whatever the space overhead that the tailor is concerned about, that's motivating trying to stable work trying to commit to this current drop order. Yeah.

E

Currently, able to cannot ignore.

C

Or arguments totally not to have it, and in that case, if it's a problem, performance-wise people won't tell us hopefully well.

H

You can always use an async block instead of an async fun right. You want to not have the space overhead.

D

Explicitly opt into this it sort of opting into a more efficient thing, but yeah, even you could even conceivably imagine having like a cookie lid or something that you enable everywhere. That says, you know, tells.

F

You to drop everything.

D

Early before a yield to so it doesn't like you know when the generator stayed or could tell you to really make something interesting block like we can make trolling around this.

G

Thing doesn't have dropped glue. Would you still be allowed to not.

E

So well, it is a great.

F

B

F

This me, or is anybody else staring at mutex of unit and thinking the compiler is doing the correct thing? Optimizing it out of existence.

D

B

Know I used mute executed all the time and.

D

I didn't know: I use it for meaningful things.

B

In another language, right, yes, exactly it's just a lot of thank yous, you're.

F

Passing it into a function and then counting on the fact that it doesn't get dropped in till the end of that function, yeah, yes! Rather, let's start of that function. No.

E

Until the end of that, I mean yes until the variable goes out of I.

F

Thought the current drop order of a non async function is that an unused argument may be dropped at the beginning of the function. It did only if you use an underscore pattern.

D

G

Even then, I only doesn't matter like it still will still provide another function for a function. Cranham interesting.

H

E

I think judge explained was that if you drop it explicitly in the button, the first statement of the function that it'll drop then.

H

I

H

A meeting about this many months ago, because there was some discussion of what the draw order should be for underscores there.

E

Like a future, you understand, and nothing has happened, nothing happens until the future starts getting awaited or spawned or something, and so it's different from it being before the future is even constructed. The problem here is that it's like this for async functions just like this would be like an earlier phase than even the beginning of the function, because if you drop something supposedly to be in the function that won't run until the async monkey demo honto you serving music, the future.

F

We had a discussion about this quite some time ago for non async functions and I think the conclusion we came to there was there is some code that counts on drop order, and we can't change that. But I think that still leaves us with the question of how much do we need a sink.

F

F n drop order to match a pin drop order, and how much can we say this is a new feature and we can define the drop order as we see fit as long as it isn't too surprising, and what is not too surprising me.

E

Yeah, that's the that's. The point is that I find the arguments compelling. That would be surprising. That would be inconsistent to do it like the thing that the samples people understand is for them in the work the same way and like I also think, like words like the argument in favor of changing.

E

It is sort of this like argument from like there being a lot of small thing like these death by a thousand cuts and most of the other cuts we can deal with, and this is like the one cut that we like automatic locations and so.

I

V by thousand cuts, then you.

E

Can handle one of the thousand cuts that we can like take? The overhead is like by default things stay around until and.

D

Someone and someone actually didn't bring up their meal and someone actually did bring up the related thing on the issue, which is that the other thing we don't do, but we could, if we wanted to fix one of these other sort of death by a thousand cuts issues, is dropping arguments prior to yields right.

E

E

D

The difference to me between those and I feel like this is still a little bit silly of me to be making making this distinction, since, like I, I, think I'm on your side that we should just go ahead and change the behavior. But I think that the difference to me for those is that one is a lot more common and a lot more observable.

F

D

Whereas the other impacts lot mark open breaks, a lot more users expectations, but it's.

I

Also, the other other comment.

D

I had to make about it is that one of those you can easily sort of or you should be able to easily assign step by by calling drop early, whereas the other one of them. You cannot, because you have to accurately refactor it up and they were a nascent block. It.

F

Just so happens, the drop doesn't actually right now because of.

D

Other generator optimizations work and where we tax storage, Ted, but that's a mess so I don't know it's a well.

E

I mean I think that specifically what you're saying you can't do these is not have it ever become a future state, but you can like drop it as the first thing you do in the function right, and so that will you can be part of the later always do like only the initial state will have to have it.

E

Yes, that's cruel, it would be so like you can't keep it from growing the interstates for the Metamora state and they will be larger than that's. What that's my point. Insisted I mean.

D

I, don't I mean.

E

That doesn't actually work.

I

Especially for copy.

D

Types but like the weird details are probably not worth coming in to.

D

E

Work at some point in time, yeah right, Bobby actually never used. We like could optimize them because we know that their drop is pure, because the other doctor.

D

Yeah yeah yeah, we are if there are unused copy arguments we can theoretically, like.

G

D

G

Away yeah, that's the one detail actually like ties in the world's bring it before I. Think I worry about with the future comparability for those who are the future gating. They would lead to a pattern, I think of people hard for sleep, their arguments in order to avoid the future gate, including things that are copy, and thus they would artificially assume.

D

We would not end on onion copy arguments, but I mean that's a detail. I mean.

B

D

B

When we say feature gate like I,.

B

Think it should be one of them, I, don't think we could consider this mvps futures done well. This feature could mean so.

D

So I would like to make a proposal for what we do here and, and people can say they disagree with that. I think that we should go ahead and do whatever we can to try and change the drop order to be the thing that would match the me havior of other functions and as part of that, if it comes time to stabilize async/await and we have all the other blockers covered, we should introduce a feature gate error on non non own. Only drop implementing argument patterns that are unused does that sound right to people.

D

Well, things are dropped, blue and.

E

Like make it possibly don't want to drop it.

D

Yes think things with drop clip you're, you're right. What.

E

Or if you were contacted.

D

Rights that aren't coffee things alone. It won't pop. If you yeah, we can we can. We can make it copy if you want I, think that that's slightly less expensive, but the distinction is not very significant, but it's just about like yes, I, agree, I, don't care if.

C

You are confident about that implementation of the feature gate sure so.

B

It's the idea, Taylor is basically we won't block on this, but we'll make a best effort to alter it. Correct.

E

I think also we make the decision design decision and we're not blocking honesty, implementation.

D

I'm posing that we decide this now that we want to match other functions and that if we don't ship with that behavior initially, it's only because we ran out of time to implement I'm.

G

That's exactly what I wanted to do. I will point out that I think the NLL project made a similar choice.

B

Enema but we didn't feature gate: okay,.

E

Lesson learned that seems good one.

H

Random thought this means that we can't say: async fun is Addie sugar to an async block and imple to write, because.

H

B

The do sugar, if we take this decision, the D sugaring would be delayed. Well, maybe not depends yeah.

D

It's um it's a it's! A more complicated series of rillette bindings inside of the a syncopal. Yes,.

H

We weary shadow, all of the function, arguments into the async block and it does the.

E

Same thing like that, yeah, okay,.

H

Something that we meant to get Ansonia.

E

Explaining to users we're not going to talk.

C

There is one thing: I need to double check with Nico about inference and compatibility. But yes, this sounds good. Okay,.

B

To do we feel we need to do any. This is sufficient just copy this onto the thread. Do we need to do it FCP or anything.

C

Aaron is not here and Eddie will bully just check your books and will.

H

Not have an opinion, yes, should we have an FCP for process reasons, but enough CP doesn't block feature getting it or implementing it. I.

B

D

It like it, you don't make fancy peas for every decision.

B

D

Since we have an FCP, async/await, syntax stabilization itself, I think it would fall under that. Okay sounds.

E

D

E

A sense that we have enough to be four: isn't it.

B

Great I can do that RFC.

H

Had been that, it's supposed to be this way anyway, so.

C

It's just that's the key and everyone checks their votes here. Doesn't.

E

C

Really don't think enough.

E

Cp is needed, I think.

B

E

Generally, a good idea to move away from the idea that we do have enough to be for every decision. Virtuous, alright,.

B

It's 348 that was pretty productive. Now the question is some: do wheat we have 12 minutes. Do we talk about any FCP process proposal? I? Think stability is too big, or do we end early and go home to our well-being at home, for those are not noscopes to go back to our jobs so.

D

For the my office caught on fire and apparently we're all being sent home because the network's about to go out.

F

Priority was oh, we have to go home because the network went out.

F

Of consensus, I look looking over to this proposal. Does anybody see a problem with this proposal? I'm not suggesting we discuss it in depth, I'm more asking. Can we get a show of hands, for this sounds good one.

C

E

Already told Niko that I thought it was great, so yeah.

F

I liked it yep.

E

E

B

I read so yeah basically will try to bring FCPS and automatic.

F

Nominate, yes,.

C

That sounds good. The one thing I will say is that people in the team are very good at just looking at at CVS on mice, about Doris and they're. Usually there are some like non-controversial things that just hang forever, because no one reviewed them and if so, we shouldn't like counselor CPD's, because of that, because people didn't have time to check them. People just like main alarm every ones every once a week or something having to get that are see the box and check their boxes. If there's part.

D

Of the one of the issues I have is: is that like I I, don't know, I feel like I'm, pretty good about this I'm? Looking at my list again now but like there are quite a few things on here that are just like things that I have intentionally not checked and I. Wonder if there's maybe some way that we could mark those as being exceptional, that it's clear when there are new things that we should pay attention to.

D

You should that's what it doesn't, notate that on RFC bot, RS right, oh I, want to see a list of like all the things that I have neither listed a concern on nor have checked a box on and in.

F

D

There's places were then, and even in that list it's a little ambiguous because there are places where I haven't left your concern. But my concern is just yes what they said right and someone else has left a concern. Yes, yes,.

C

D

Seems like a feature request.

B

So basically allow a way to exclude I think if your concern is their concern, you should check your box but yeah, but I think when I was kind of hoping to get away from a little bit was some of the asynchrony in our current process, because I think it doesn't always work that well. So, like part of my goal here was that when we have an SCP, we would nominate the issue and when we talk about it, ideally someone would discuss because I think a lot of times. That's also be helpful.

B

What is the rst? What's it about, and hopefully, at the end of the meeting, we can all go check our boxes or register or if we have concerns like we'll know, so we won't have his problem that it's.

C

In general, it always works.

C

C

B

A

Run for your life.

B

Well anyway, I agree with you that central I'm sure it won't work quite how I projected it, but at minimum I think we should try to bring it to the meetings and yeah.

E

Setting norms and knowing there will be exceptions, yes, that.

C

Makes no sense we'll.

B

Work it out all right, very good thanks.