Stream: wg-traits

Topic: weekly meeting 2019-01-07


nikomatsakis (Jan 07 2019 at 18:46, on Zulip):

Hi all! Meeting in 15 minutes.

cc @WG-compiler-traits — not sure how correct this alias is though. If you don't get a ping and you expected on (or you do and you'd prefer not to have one), send me a message and I'll update it.

Alexander Regueiro (Jan 07 2019 at 18:46, on Zulip):

speaking of aliases... any luck with that PR? ;-)

nikomatsakis (Jan 07 2019 at 18:48, on Zulip):

I thought today we'd talk a bit about "roadmap plus process" -- basically, brainstorm a bit how best to attempt and organize this effort =)

I created a google spreadsheet that contains some notes on various goals and status items, to start. This was constructed from my scraping past threads and the code. I'm probably not going to try and manage things in a Google Spreadsheet long-term but seemed like a decent place to start.

nikomatsakis (Jan 07 2019 at 18:48, on Zulip):

speaking of aliases... any luck with that PR? ;-)

I started rebasing, but decided working on the roadmap thoughts were higher priority

nikomatsakis (Jan 07 2019 at 18:48, on Zulip):

will try to finish up later today or tomorrow

scalexm (Jan 07 2019 at 18:49, on Zulip):

I will be around-ish, but I’ll probably be following the meeting more or less asynchronously

nikomatsakis (Jan 07 2019 at 18:55, on Zulip):

Traits Paper Document

nikomatsakis (Jan 07 2019 at 19:05, on Zulip):

So, I've been thinking about a few things.

nikomatsakis (Jan 07 2019 at 19:05, on Zulip):

But wait, I get ahead of myself

nikomatsakis (Jan 07 2019 at 19:05, on Zulip):

@WG-compiler-traits :wave: starting :)

tmandry (Jan 07 2019 at 19:07, on Zulip):

do you still feel that rustc-guide is a good way of communicating that design?

nikomatsakis (Jan 07 2019 at 19:07, on Zulip):

I was just writing something about that

nikomatsakis (Jan 07 2019 at 19:07, on Zulip):

yeah, that's the question

nikomatsakis (Jan 07 2019 at 19:07, on Zulip):

it feels like maybe?

tmandry (Jan 07 2019 at 19:07, on Zulip):

or should we have something in-between RFCs and the guide, say, "design docs"

nikomatsakis (Jan 07 2019 at 19:07, on Zulip):

I mean the only "downside" of the rustc-guide to me

nikomatsakis (Jan 07 2019 at 19:08, on Zulip):

and I think if we put too high of a barrier

nikomatsakis (Jan 07 2019 at 19:08, on Zulip):

is that it feels like it's a bit "finished"

nikomatsakis (Jan 07 2019 at 19:08, on Zulip):

but otoh I think it's kind of a logical place

nikomatsakis (Jan 07 2019 at 19:08, on Zulip):

if will not grow enough

nikomatsakis (Jan 07 2019 at 19:08, on Zulip):

but maybe it makes sense to have a place where the design/docs are "evolving"

nikomatsakis (Jan 07 2019 at 19:08, on Zulip):

and a pointer to that from rustc-guide

tmandry (Jan 07 2019 at 19:08, on Zulip):

yeah I tend to agree

nikomatsakis (Jan 07 2019 at 19:08, on Zulip):

sort of a "meta repository", I guess

nikomatsakis (Jan 07 2019 at 19:09, on Zulip):

hmm, does anybody know, if you have two repos in the same github org, can they reference one another in their "github projects" and things?

nikomatsakis (Jan 07 2019 at 19:09, on Zulip):

I've been wondering if those would be a good fit to help us track things

nikomatsakis (Jan 07 2019 at 19:09, on Zulip):

they gotta be good for something

tmandry (Jan 07 2019 at 19:09, on Zulip):

not sure on that one

tmandry (Jan 07 2019 at 19:10, on Zulip):

personally, my go-to for writing out a design is something like google docs or dropbox paper

tmandry (Jan 07 2019 at 19:10, on Zulip):

I find the whole edit/preview/commit/push/pr cycle to be a bit much in the early stages

nikomatsakis (Jan 07 2019 at 19:11, on Zulip):

seems like you sort of can: https://github.com/rust-lang/rustc-guide/projects/1

nikomatsakis (Jan 07 2019 at 19:11, on Zulip):

yeah, I mean, a paper or google doc could work too

nikomatsakis (Jan 07 2019 at 19:12, on Zulip):

I think PRs don't have to be onerous per se though

nikomatsakis (Jan 07 2019 at 19:12, on Zulip):

that is, we could just merge quickly

nikomatsakis (Jan 07 2019 at 19:12, on Zulip):

:)

nikomatsakis (Jan 07 2019 at 19:13, on Zulip):

(it's also easier to have e.g. a discussion about a design and so forth)

nikomatsakis (Jan 07 2019 at 19:14, on Zulip):

I guess i'd have to move some of the docs from rustc-guide to start

tmandry (Jan 07 2019 at 19:15, on Zulip):

I mean I see projects etc as a great way for tracking progress, not sure about design

tmandry (Jan 07 2019 at 19:16, on Zulip):

PRs can work as a discussion point for design, yes

tmandry (Jan 07 2019 at 19:16, on Zulip):

but it always feels _to me_ like "we have to get everything right before we merge"

nikomatsakis (Jan 07 2019 at 19:16, on Zulip):

my plan would be to have design docs living as files in the repository, and issues tracking things that remain to be figured out

nikomatsakis (Jan 07 2019 at 19:16, on Zulip):

the projects were meant more for tracking impl progress and "things that need to get figured out"

nikomatsakis (Jan 07 2019 at 19:16, on Zulip):

but it always feels _to me_ like "we have to get everything right before we merge"

yeah, that makes sense

nikomatsakis (Jan 07 2019 at 19:16, on Zulip):

I guess I like editing markdown files in emacs

nikomatsakis (Jan 07 2019 at 19:17, on Zulip):

so there's that too :)

nikomatsakis (Jan 07 2019 at 19:17, on Zulip):

anyway, beyond this question ....

nikomatsakis (Jan 07 2019 at 19:17, on Zulip):

.. we can always try and experiment with a few things

tmandry (Jan 07 2019 at 19:17, on Zulip):

yeah I'm open to experimenting and can certainly work with that

nikomatsakis (Jan 07 2019 at 19:17, on Zulip):

I guess I like editing markdown files in emacs

there are always wikis, btw

tmandry (Jan 07 2019 at 19:17, on Zulip):

we just need a place where it feels safe to put something that isn't quite finished

nikomatsakis (Jan 07 2019 at 19:17, on Zulip):

(github wiki I mean)

nikomatsakis (Jan 07 2019 at 19:18, on Zulip):

but in general I think I'd rather shoot for PRs and just try to encourage half-finished stuff to land :)

tmandry (Jan 07 2019 at 19:18, on Zulip):

true. main advantage of PR or paper/docs over wiki is the commenting ability

nikomatsakis (Jan 07 2019 at 19:18, on Zulip):

basically refer to open an issue with stuff to be filled in than to "get it all done" in the PR

Sunjay Varma (Jan 07 2019 at 19:18, on Zulip):

what functionality do github wikis provide? Can you see the history and stuff and write comments?

Sunjay Varma (Jan 07 2019 at 19:19, on Zulip):

what if we just opened a PR with the design and no changed files. Does GitHub allow for that?

Sunjay Varma (Jan 07 2019 at 19:19, on Zulip):

Or actually maybe a meta repo would be better because then you could comment on individual lines

nikomatsakis (Jan 07 2019 at 19:19, on Zulip):

I don't think you can make a PR against a wiki

nikomatsakis (Jan 07 2019 at 19:19, on Zulip):

but you can check it out and then push commits

nikomatsakis (Jan 07 2019 at 19:19, on Zulip):

which is what I usually do to read/edit it

nikomatsakis (Jan 07 2019 at 19:19, on Zulip):

it is just a git repo

nikomatsakis (Jan 07 2019 at 19:19, on Zulip):

with markdown files in it

Sunjay Varma (Jan 07 2019 at 19:19, on Zulip):

Probably a rustc-design-docs repo or something would be better then

Sunjay Varma (Jan 07 2019 at 19:19, on Zulip):

WIth the intention of putting that stuff into the rustc-guide once it is implemented/done

nikomatsakis (Jan 07 2019 at 19:19, on Zulip):

I was imagining a rust-lang/wg-traits repository

Sunjay Varma (Jan 07 2019 at 19:20, on Zulip):

That would be fine too

nikomatsakis (Jan 07 2019 at 19:20, on Zulip):

I think what is missing from my giant list is a way to prioritize, to some extent

tmandry (Jan 07 2019 at 19:20, on Zulip):

agreed, the repo idea sounds good

nikomatsakis (Jan 07 2019 at 19:20, on Zulip):

I feel like a good idea would be to start by focusing on some of the chalk work that remains

nikomatsakis (Jan 07 2019 at 19:21, on Zulip):

that is, we've made some progress — and shouldn't ignore — on integrating chalk into rustc but

Josh Huber (Jan 07 2019 at 19:21, on Zulip):

what I wouldn't give for an emacs-based collaborative markdown editor... perhaps layered on top of a gh wiki or repo. sounds (to me :) like a perfect design UX.

Sunjay Varma (Jan 07 2019 at 19:21, on Zulip):

You can prioritize with GitHub projects

nikomatsakis (Jan 07 2019 at 19:21, on Zulip):

there remain some important questions and also some obvious next steps that were raised by that

nikomatsakis (Jan 07 2019 at 19:21, on Zulip):

and I think that working on chalk would be good for folks

Sunjay Varma (Jan 07 2019 at 19:21, on Zulip):

e.g. https://github.com/sunjay/caves/projects/1

nikomatsakis (Jan 07 2019 at 19:21, on Zulip):

to help understand the way the whole design fits together

nikomatsakis (Jan 07 2019 at 19:21, on Zulip):

You can prioritize with GitHub projects

sorry what I meant was more like: "I have this big list, but it's not obvious where to start within it"

Sunjay Varma (Jan 07 2019 at 19:22, on Zulip):

You could have lists in the project where you categorize things based on their status

Sunjay Varma (Jan 07 2019 at 19:22, on Zulip):

That gives you an idea of where to start

Sunjay Varma (Jan 07 2019 at 19:22, on Zulip):

"Ready for Review" "Not Started" etc.

nikomatsakis (Jan 07 2019 at 19:22, on Zulip):

right what I mean is: we should talk about what we think we should focus on first

Sunjay Varma (Jan 07 2019 at 19:22, on Zulip):

"Author Disappeared Into Life"

nikomatsakis (Jan 07 2019 at 19:22, on Zulip):

realistically, creating mentoring instructions and things can be slow-ish

nikomatsakis (Jan 07 2019 at 19:23, on Zulip):

so I think it makes sense to set some first goals to focus on

Sunjay Varma (Jan 07 2019 at 19:23, on Zulip):

Question about the spreadsheet: does coherence need to be added at all?

nikomatsakis (Jan 07 2019 at 19:24, on Zulip):

like, if we zoom out, there are kind of a few categories of tasks

nikomatsakis (Jan 07 2019 at 19:24, on Zulip):

Question about the spreadsheet: does coherence need to be added at all?

good question, I suppose so, yes

nikomatsakis (Jan 07 2019 at 19:24, on Zulip):

it would certainly be nice to be used the formulations you came up with (or some variant of them) more directly

Sunjay Varma (Jan 07 2019 at 19:24, on Zulip):

Which section should I add it to?

nikomatsakis (Jan 07 2019 at 19:25, on Zulip):

I think I made the spreadsheet read-only because .. well, not sure why

nikomatsakis (Jan 07 2019 at 19:25, on Zulip):

oh, I feared spam for some weird reason

nikomatsakis (Jan 07 2019 at 19:25, on Zulip):

I should change that

nikomatsakis (Jan 07 2019 at 19:25, on Zulip):

at least temporarily

nikomatsakis (Jan 07 2019 at 19:25, on Zulip):

(done)

nikomatsakis (Jan 07 2019 at 19:25, on Zulip):

I would @Sunjay Varma put it under "chalk integration"

Sunjay Varma (Jan 07 2019 at 19:25, on Zulip):

Yes you're right I can't edit it

nikomatsakis (Jan 07 2019 at 19:26, on Zulip):

I think I fixed that now

nikomatsakis (Jan 07 2019 at 19:26, on Zulip):

anyway I can add it

Sunjay Varma (Jan 07 2019 at 19:26, on Zulip):

Added

Sunjay Varma (Jan 07 2019 at 19:26, on Zulip):

Oh nevermind

Sunjay Varma (Jan 07 2019 at 19:26, on Zulip):

You beat me :clap:

scalexm (Jan 07 2019 at 19:26, on Zulip):

With respect to the chalk integration bugs and the holes in chalk, I have at least a clear list of things that need to be done; and which I think are fairly atomic so that we may not need to prioritize them

scalexm (Jan 07 2019 at 19:27, on Zulip):

Well my guess it that the only thing we may want to prioritize right now is the goal selection algorithm in chalk-engine

nikomatsakis (Jan 07 2019 at 19:28, on Zulip):

maybe we can just do some of both

scalexm (Jan 07 2019 at 19:28, on Zulip):

Because we may want to use it for answering ?X: Trait goals, as discussed with @nikomatsakis

nikomatsakis (Jan 07 2019 at 19:28, on Zulip):

I think one big thing I was concerned about with chalk itself was trait object integration

nikomatsakis (Jan 07 2019 at 19:28, on Zulip):

yeah, that's on my list

nikomatsakis (Jan 07 2019 at 19:28, on Zulip):

I also want to see refactoring for specialization

nikomatsakis (Jan 07 2019 at 19:29, on Zulip):

but that's a bit lower in priority

scalexm (Jan 07 2019 at 19:29, on Zulip):

Actually I have some ideas on trait objects, which I think may solve the soundness issues we know about

nikomatsakis (Jan 07 2019 at 19:29, on Zulip):

oh yeah? interesting

tmandry (Jan 07 2019 at 19:29, on Zulip):

even if it's somewhat arbitrary, my feeling is it would be good to have some guidance on which issues to pick up next

tmandry (Jan 07 2019 at 19:29, on Zulip):

github projects do seem pretty great for that, actually

nikomatsakis (Jan 07 2019 at 19:30, on Zulip):

OK, I'm feeling pretty good about the idea of a wg-traits repository to house some of this stuff

nikomatsakis (Jan 07 2019 at 19:30, on Zulip):

and trying to use projects to track stuff

scalexm (Jan 07 2019 at 19:30, on Zulip):

(Re: trait objects; I’ll spawn another topic to explain myself later because I’m on my phone now)

nikomatsakis (Jan 07 2019 at 19:30, on Zulip):

even if it's somewhat arbitrary, my feeling is it would be good to have some guidance on which issues to pick up next

I agree with this, we need to focus on some goals :)

nikomatsakis (Jan 07 2019 at 19:31, on Zulip):

but maybe the divide between "work on chalk" and "work on rustc" isn't really needed

nikomatsakis (Jan 07 2019 at 19:31, on Zulip):

the main thing is to pick a priority or two

nikomatsakis (Jan 07 2019 at 19:32, on Zulip):

I think that we can add a simple API to control predicate clause ordering in chalk for sure

nikomatsakis (Jan 07 2019 at 19:33, on Zulip):

I guess some initial weekly goals can be:

and in terms of code

nikomatsakis (Jan 07 2019 at 19:33, on Zulip):

In terms of stuff outside chalk:

nikomatsakis (Jan 07 2019 at 19:34, on Zulip):

@Alexander Regueiro you were doing some stuff on the impl Trait in let binding work?

nikomatsakis (Jan 07 2019 at 19:34, on Zulip):

(ah, that's something else that chalk probably needs to be extended to understand, though it may be related to trait objects)

tmandry (Jan 07 2019 at 19:34, on Zulip):

selection ordering selection..?

nikomatsakis (Jan 07 2019 at 19:34, on Zulip):

no idea what I was tryping

nikomatsakis (Jan 07 2019 at 19:34, on Zulip):

edited :)

tmandry (Jan 07 2019 at 19:34, on Zulip):

:)

nikomatsakis (Jan 07 2019 at 19:34, on Zulip):

the idea is this:

nikomatsakis (Jan 07 2019 at 19:35, on Zulip):

if you have a list of things to prove, which order do you use?

Keith Yeung (Jan 07 2019 at 19:35, on Zulip):

i think niko means that we are implementing prioritization of subgoal selection

nikomatsakis (Jan 07 2019 at 19:35, on Zulip):

chalk right now says "start with the last thing in the list", mostly because it means I can call vec.pop()

nikomatsakis (Jan 07 2019 at 19:35, on Zulip):

prolog traditionally starts from the front

nikomatsakis (Jan 07 2019 at 19:35, on Zulip):

ultimately I think we should use some heuristics, perhaps examining e.g. the impls for traits and things

nikomatsakis (Jan 07 2019 at 19:35, on Zulip):

but to start, we have some cases that we know are a bad idea

nikomatsakis (Jan 07 2019 at 19:36, on Zulip):

e.g., if you have ?T: Sized where ?T is not bound

nikomatsakis (Jan 07 2019 at 19:36, on Zulip):

then trying to prove that is kind of pointless -- right now, it does something semi-bogus, but to work properly, it would basically have to enumerate all types that are sized (which is an awful lot of types)

nikomatsakis (Jan 07 2019 at 19:36, on Zulip):

what we would prefer is that if you have something like ?T: Sized, X: AsRef<?T>, you start with the latter goal, which might yield you a specific type

nikomatsakis (Jan 07 2019 at 19:37, on Zulip):

for context: when the solver is "solving" a goal, that basically means it is enumerating values of the type variables for which the goal is true

nikomatsakis (Jan 07 2019 at 19:37, on Zulip):

so it makes sense to start with the subgoals that have the fewest true values

nikomatsakis (Jan 07 2019 at 19:38, on Zulip):

in some sense, controlling this ordering is the challenge if you are trying to be efficient

tmandry (Jan 07 2019 at 19:38, on Zulip):

makes sense

nikomatsakis (Jan 07 2019 at 19:38, on Zulip):

or at least it can make the difference between getting hopelessly lost and finding an answer

nikomatsakis (Jan 07 2019 at 19:38, on Zulip):

(there would be other options too, I guess -- you could imagine something adaptive maybe, but it'd be a big change to support)

tmandry (Jan 07 2019 at 19:41, on Zulip):

ok so we'll get a repo and project set up
in terms of prioritizing, if there are no obvious places to start, perhaps we should pick things that either a) seem like faster wins or b) have a lot of people requesting them

nikomatsakis (Jan 07 2019 at 19:41, on Zulip):

so meta question:

Of the people here, are any of you thinking "geez I'd like to write some code on traits this week, but I don't know what to do"?

That is, I plan to focus this week and next on mapping out steps in more detail so that I can put out some calls to try and pull more people in. But I'd like to prioritize people here now

nikomatsakis (Jan 07 2019 at 19:41, on Zulip):

(I think I have some ideas what to prioritize)

scalexm (Jan 07 2019 at 19:42, on Zulip):

Note: the current semi-bogus implementation of ?T: Sized is only buggy with built-in impls generated by the compiler, because you cannot even enumerate them all (because you cannot e.g. represent variadic impls on tuples)

nikomatsakis (Jan 07 2019 at 19:42, on Zulip):

Note: the current semi-bogus implementation of ?T: Sized is only buggy with built-in impls generated by the compiler, because you cannot even enumerate them all (because you cannot e.g. represent variadic impls on tuples)

I'm not sure if this is true

nikomatsakis (Jan 07 2019 at 19:42, on Zulip):

That is, doesn't it just generate an ambiguous result?

tmandry (Jan 07 2019 at 19:42, on Zulip):

yeah I would like to dig in to some code this week

Sunjay Varma (Jan 07 2019 at 19:43, on Zulip):

Of the people here, are any of you thinking "geez I'd like to write some code on traits this week, but I don't know what to do"?

:person_tipping_hand:

scalexm (Jan 07 2019 at 19:44, on Zulip):

@nikomatsakis I hard coded the ambiguity for the Sized trait

scalexm (Jan 07 2019 at 19:44, on Zulip):

But for other traits, it will just enumerate all known impls, which could only have been user-written

scalexm (Jan 07 2019 at 19:45, on Zulip):

So it may be slow (if there are a lot of impls), but at least I think it’ll be correct in all cases

nikomatsakis (Jan 07 2019 at 19:45, on Zulip):

Hmm, so re: actionable things... actually looking at my list, a lot of them are in rustc.

One that leaps to mind is universe mapping. Contra to what I wrote about rustc hacking, this would be extending the rustc code to help make queries smaller.

Josh Huber (Jan 07 2019 at 19:45, on Zulip):

for me, i'm more like "geez I'd like to write some code on traits later this month, after I've made in through more of the existing documentation" :)

nikomatsakis (Jan 07 2019 at 19:46, on Zulip):

In terms of chalk:

nikomatsakis (Jan 07 2019 at 19:47, on Zulip):

maybe there are some smaller steps though

nikomatsakis (Jan 07 2019 at 19:47, on Zulip):

I suspect there's just some chalk cleanup or other things to be done

tmandry (Jan 07 2019 at 19:47, on Zulip):

I assume salsa-ification can be done independently of the rest of rustc

Keith Yeung (Jan 07 2019 at 19:47, on Zulip):

i'm more of the learn-as-you-work type: i need to see actual code when i read docs

Keith Yeung (Jan 07 2019 at 19:47, on Zulip):

which is why i'm working on the clause selection order

nikomatsakis (Jan 07 2019 at 19:47, on Zulip):

I assume salsa-ification can be done independently of the rest of rustc

yeah this is specifically for the chalk test runner code (not the engine itself)

nikomatsakis (Jan 07 2019 at 19:48, on Zulip):

I can give some more context, but basically in order to handle specialization, we're going to need to be able to do the lowering a trait at a time

tmandry (Jan 07 2019 at 19:48, on Zulip):

interesting

nikomatsakis (Jan 07 2019 at 19:48, on Zulip):

right now, the chalk test runner begins by

but specialization needs to be able to construct relationships between impls, and that sometimes requires doing trait solving

nikomatsakis (Jan 07 2019 at 19:48, on Zulip):

so it needs to be able to lower some traits to rules, do some solving, then lower more traits to rules

nikomatsakis (Jan 07 2019 at 19:49, on Zulip):

probably too much to do in "one step"

nikomatsakis (Jan 07 2019 at 19:49, on Zulip):

the last time I was looking into this, I started by breaking up chalk into more crates :P

nikomatsakis (Jan 07 2019 at 19:49, on Zulip):

never actually got to the real task

nikomatsakis (Jan 07 2019 at 19:49, on Zulip):

there's also stuff like this https://github.com/rust-lang-nursery/chalk/issues/168

Keith Yeung (Jan 07 2019 at 19:49, on Zulip):

heh, as if chalk doesn't itself have enough subcrates

Sunjay Varma (Jan 07 2019 at 19:50, on Zulip):

Are you sure you want to use failure?

Sunjay Varma (Jan 07 2019 at 19:50, on Zulip):

Isn't it going to be superceded soon?

nikomatsakis (Jan 07 2019 at 19:50, on Zulip):

I don't know :)

nikomatsakis (Jan 07 2019 at 19:50, on Zulip):

i'm not sure we want to use any such crate really

tmandry (Jan 07 2019 at 19:50, on Zulip):

salsa stuff sounds interesting to me

Keith Yeung (Jan 07 2019 at 19:50, on Zulip):

this is news to me, i thought failure is the way to go

nikomatsakis (Jan 07 2019 at 19:50, on Zulip):

salsa stuff sounds interesting to me

what we should probably do, @tmandry, is schedule a slot to talk it over

nikomatsakis (Jan 07 2019 at 19:50, on Zulip):

if that works for you

tmandry (Jan 07 2019 at 19:51, on Zulip):

improving queries in rustc too if that's more important

tmandry (Jan 07 2019 at 19:51, on Zulip):

yep that works for me

nikomatsakis (Jan 07 2019 at 19:51, on Zulip):

what I did in the NLL group that I'd like to continue doing:

nikomatsakis (Jan 07 2019 at 19:51, on Zulip):

is to do sessions to talk over how stuff works and record them

nikomatsakis (Jan 07 2019 at 19:51, on Zulip):

then post those videos on youtube

Sunjay Varma (Jan 07 2019 at 19:51, on Zulip):

I would love to join in on one of those

nikomatsakis (Jan 07 2019 at 19:51, on Zulip):

actually, no reason not to have >1 person

tmandry (Jan 07 2019 at 19:52, on Zulip):

this is news to me, i thought failure is the way to go

https://github.com/rust-lang-nursery/failure#evolution

nikomatsakis (Jan 07 2019 at 19:53, on Zulip):

I gotta run soon but I'd be game tomorrow to do a session to talk over salsa-ification and/or the universe-map stuff I was talking about

nikomatsakis (Jan 07 2019 at 19:53, on Zulip):

not sure how flexible y'all are, I could do say 11am UTC-5 (Boston time)

Sunjay Varma (Jan 07 2019 at 19:53, on Zulip):

Please ping me. That time works :)

tmandry (Jan 07 2019 at 19:53, on Zulip):

Works for me

nikomatsakis (Jan 07 2019 at 19:53, on Zulip):

ok

Sunjay Varma (Jan 07 2019 at 19:53, on Zulip):

(I may have to run at 11:30 though)

Josh Huber (Jan 07 2019 at 19:53, on Zulip):

I'd love to join if i can swing that time!

nikomatsakis (Jan 07 2019 at 19:53, on Zulip):

I am also avail at 13:00 and 15:00

nikomatsakis (Jan 07 2019 at 19:54, on Zulip):

yeah I'll just do a general ping

nikomatsakis (Jan 07 2019 at 19:54, on Zulip):

but we should pick a time :)

tmandry (Jan 07 2019 at 19:55, on Zulip):

15:00 can also work for me

nikomatsakis (Jan 07 2019 at 19:56, on Zulip):

voting time. put :+1: on times that work for you :)

nikomatsakis (Jan 07 2019 at 19:56, on Zulip):

11:00 UTC-5

nikomatsakis (Jan 07 2019 at 19:56, on Zulip):

13

nikomatsakis (Jan 07 2019 at 19:56, on Zulip):

15

Sunjay Varma (Jan 07 2019 at 19:56, on Zulip):

:thumbs_down: if you do not prefer a time/can't make a time

Sunjay Varma (Jan 07 2019 at 19:57, on Zulip):

(in case we tie)

nikomatsakis (Jan 07 2019 at 19:57, on Zulip):

Sounds like 15:00 is the winner...

nikomatsakis (Jan 07 2019 at 19:57, on Zulip):

OK, let's do that.

nikomatsakis (Jan 07 2019 at 19:58, on Zulip):

Gotta run, thanks all! :wave:

Alexander Regueiro (Jan 07 2019 at 20:21, on Zulip):

@Alexander Regueiro you were doing some stuff on the impl Trait in let binding work?

Sorry Niko, I got called away. And yes, getting back to that. Working on the associated_type_bindings feature too (the issue I mentioned to you earlier was about that).

Alexander Regueiro (Jan 07 2019 at 20:22, on Zulip):

Is Chalk integration going well? Lifetime bounds coming soon?

scalexm (Jan 07 2019 at 20:55, on Zulip):

(PS: 15:00 works for me)

Last update: Nov 12 2019 at 15:35UTC