Stream: wg-async-foundations

Topic: weekly meeting 2019.05.28


centril (May 28 2019 at 17:00, on Zulip):

:wave: -- what are the plans for today?

nikomatsakis (May 28 2019 at 17:01, on Zulip):

:wave:

Yoshua Wuyts (May 28 2019 at 17:02, on Zulip):

o/

nikomatsakis (May 28 2019 at 17:02, on Zulip):

Well, @WG-async-await, but also @boats, @Florian Gilcher, @Yoshua Wuyts, @Nemo157 -- meeting time.

nikomatsakis (May 28 2019 at 17:02, on Zulip):

Today we had the idea of doing some more general discussion of async-await stabilization

centril (May 28 2019 at 17:03, on Zulip):

Are we doing a zoom meeting for that?

nikomatsakis (May 28 2019 at 17:03, on Zulip):

I was expecting to do it over zulip

nikomatsakis (May 28 2019 at 17:04, on Zulip):

I took some notes on a rough agenda here, but I didn't super refine it yet

Florian Gilcher (May 28 2019 at 17:04, on Zulip):

hi!

centril (May 28 2019 at 17:05, on Zulip):

In my experience these sorta meetings (stabilization and lang plans) are suboptimal when text-based but let's give it a try I suppose

nikomatsakis (May 28 2019 at 17:05, on Zulip):

@Taylor Cramer you around?

centril (May 28 2019 at 17:06, on Zulip):

@nikomatsakis are you driving?

Taylor Cramer (May 28 2019 at 17:06, on Zulip):

@nikomatsakis yeah, I don't think we have anything to talk about still here, though

Taylor Cramer (May 28 2019 at 17:06, on Zulip):

@tmandry 's stuff has been making steady progress

nikomatsakis (May 28 2019 at 17:07, on Zulip):

You mean from the impl side in specific?

Taylor Cramer (May 28 2019 at 17:07, on Zulip):

and the lifetimes stuff hasn't made progress--

Taylor Cramer (May 28 2019 at 17:07, on Zulip):

^^right, there's not much going on in implementation land-- I don't think anything has really changed since last week

nikomatsakis (May 28 2019 at 17:07, on Zulip):

OK, so I was figuring we'd hijiack the meeting to look more at the overall stabilization picture

Taylor Cramer (May 28 2019 at 17:07, on Zulip):

sure

centril (May 28 2019 at 17:07, on Zulip):

@varkor landed a PR to reorganize tests

centril (May 28 2019 at 17:08, on Zulip):

I haven't been able to check what tests need to be written and write any yet

nikomatsakis (May 28 2019 at 17:08, on Zulip):

So let's back up a sec

centril (May 28 2019 at 17:08, on Zulip):

should try to do that today or tomorrow

Taylor Cramer (May 28 2019 at 17:08, on Zulip):

although I did want to ping you again about the lifetimes stuff to ask if you had had a chance to look at it anymore. I spent a while poking around but I feel like i'm not familiar enough with the code to be of much use yet

nikomatsakis (May 28 2019 at 17:08, on Zulip):

Yeah, it's been on my mind

centril (May 28 2019 at 17:08, on Zulip):

@nikomatsakis see how text based communication is disorganized? :P

nikomatsakis (May 28 2019 at 17:09, on Zulip):

but the thing I wanted to start with here is:

nikomatsakis (May 28 2019 at 17:09, on Zulip):

basically talking more about our "communication plan"

nikomatsakis (May 28 2019 at 17:09, on Zulip):

it seems clear that stabilizing the async fn syntax is going to be a big deal

centril (May 28 2019 at 17:09, on Zulip):

What parts are we stabilizing to begin with?

nikomatsakis (May 28 2019 at 17:10, on Zulip):

at the same time, we know that all the pieces we eventually mgiht want are not in place, or are not as nice as they could/will be

nikomatsakis (May 28 2019 at 17:10, on Zulip):

a good question, we could start there to make sure we're all on same page :)

Taylor Cramer (May 28 2019 at 17:10, on Zulip):

bare and inherent async fn, async blocks, not async closures

nikomatsakis (May 28 2019 at 17:10, on Zulip):

my understanding is that we are stabilizing:

nikomatsakis (May 28 2019 at 17:11, on Zulip):

but not:

centril (May 28 2019 at 17:11, on Zulip):

and async move { ... } blocks?

Taylor Cramer (May 28 2019 at 17:11, on Zulip):

Yes, those are async blocks

nikomatsakis (May 28 2019 at 17:11, on Zulip):

(I presume there are two kinds?)

nikomatsakis (May 28 2019 at 17:11, on Zulip):

(I'm actually not sure, I've only ever seen move blocks in practice)

Taylor Cramer (May 28 2019 at 17:12, on Zulip):

Yes, they behave just like move and non-move closures

centril (May 28 2019 at 17:13, on Zulip):

This corresponds to what I was expecting :+1:

nikomatsakis (May 28 2019 at 17:13, on Zulip):

Right. So it seems like this already starts to point at some of the things I meant about "roadmap" -- notably, if you are working with async I/O in practice, you may need to (for example) have an async fn in a trait. There are workarounds for how to handle this, but they can be a bit difficult.

centril (May 28 2019 at 17:13, on Zulip):

Tho we have some bugs around coercions and whatnot

nikomatsakis (May 28 2019 at 17:13, on Zulip):

I think other examples would be rough edges around interacting with futures combinators and streams

nikomatsakis (May 28 2019 at 17:14, on Zulip):

I think this bears on the docs that we wind up with in the end, and probably also on the communication leading up to the stabilization

nikomatsakis (May 28 2019 at 17:15, on Zulip):

(But also perhaps on some questions of what should be blocking -- e.g., how important are "polish" issues?)

nikomatsakis (May 28 2019 at 17:17, on Zulip):

I think one question I would like to settle is how we will communicate the status etc-- but maybe best to dig into that after we've looked at a few other questions in detail?

boats (May 28 2019 at 17:17, on Zulip):

I think the big thing is messaging to the community that we are aware this isn't 100% feature complete, that we are working to move forward even after stabilizing this MVP, etc

boats (May 28 2019 at 17:17, on Zulip):

I'd like to revamp the areweasyncyet.rs website for that. Right now it has good links but there's no context for understanding what it says unless you're already involved in the process

nikomatsakis (May 28 2019 at 17:17, on Zulip):

There is an async blog

centril (May 28 2019 at 17:18, on Zulip):

I think that can be taken care of relatively easily in the report/relnotes/blog-post

nikomatsakis (May 28 2019 at 17:18, on Zulip):

Yes, I'd like that too

nikomatsakis (May 28 2019 at 17:18, on Zulip):

I think this probably merits a "main blog" blog post

boats (May 28 2019 at 17:18, on Zulip):

Yea, this needs a lot more information than release notes

centril (May 28 2019 at 17:18, on Zulip):

@nikomatsakis you mean beyond the 1.37.0 blog post?

boats (May 28 2019 at 17:18, on Zulip):

Because its going to be drawing attention from people who aren't invested in our processes

nikomatsakis (May 28 2019 at 17:19, on Zulip):

Yeah, I meant a dedicated kind of "Async I/O Update" post.

centril (May 28 2019 at 17:19, on Zulip):

@boats by release notes you mean RELEASES.md, right?

centril (May 28 2019 at 17:19, on Zulip):

@nikomatsakis sure that sounds good

boats (May 28 2019 at 17:19, on Zulip):

@centril I mean more than what we would do for a normal feature. All the things you listed are not sufficient for communicating the status of this feature, because it will draw attention from people who do not read any of that

centril (May 28 2019 at 17:19, on Zulip):

I expect most of the 1.37.0 blog post to go towards async/await anyways

nikomatsakis (May 28 2019 at 17:20, on Zulip):

So let's assume we will have such an update somewhere -- and I agree with @boats that areweasyncyet.com should be updated to be simplified. I'm not sure who will be doing said work--

nikomatsakis (May 28 2019 at 17:20, on Zulip):

we could (probalby?) branch off that discussion for later

nikomatsakis (May 28 2019 at 17:20, on Zulip):

i.e., how such a page should look to be most approachable

nikomatsakis (May 28 2019 at 17:20, on Zulip):

(though I think that this "wasm post mvp future" blog post is good inspiration)

Yoshua Wuyts (May 28 2019 at 17:21, on Zulip):

@nikomatsakis I can reach out to the lead maintainer of areweasyncyet to start communicating that we want to do _something_ to give them a heads up. Agree we should probably leave the details for later

boats (May 28 2019 at 17:21, on Zulip):

Yea what would be great in this meeting is to resource the work.

nikomatsakis (May 28 2019 at 17:22, on Zulip):

let's circle back to that maybe, because i'd like to get an overview of what has to be done and by when

nikomatsakis (May 28 2019 at 17:22, on Zulip):

but i'm jotting it down inmy list of things...

centril (May 28 2019 at 17:22, on Zulip):

@boats Can you help out with the 1.37.0 release blog post?

centril (May 28 2019 at 17:23, on Zulip):

I usually write it but having domain expertise is nice

nikomatsakis (May 28 2019 at 17:23, on Zulip):

One other question that seems clearly important is what sort of implementation bugs ought to be considered "blockers". I think we plan to try and trackle some of the limitations around using multiple regions (and @Taylor Cramer I will talk to you about that afterwards). Do we consider the "making futures smaller" bug to be a blocker?

nikomatsakis (May 28 2019 at 17:24, on Zulip):

(Is it a moot point because @tmandry will have it fixed?)

nikomatsakis (May 28 2019 at 17:24, on Zulip):

this is basically a question of runtime efficiency

boats (May 28 2019 at 17:24, on Zulip):

@nikomatsakis we discussed that in the previous lang team meeting

nikomatsakis (May 28 2019 at 17:24, on Zulip):

The danger from not fixing it boils down to "we may look bad in benchmarks", I think?

centril (May 28 2019 at 17:24, on Zulip):

@nikomatsakis what is particularly important to me is to know that it is fixable at some point

nikomatsakis (May 28 2019 at 17:24, on Zulip):

OK, great, what was the answer :)

boats (May 28 2019 at 17:24, on Zulip):

We were hoping not to make it a blocker but that it seemed like it would resolve before the cut with tmandry's work

nikomatsakis (May 28 2019 at 17:24, on Zulip):

OK

boats (May 28 2019 at 17:25, on Zulip):

its a pretty serious problem because it can cause stack overflows pretty easily

nikomatsakis (May 28 2019 at 17:25, on Zulip):

The other outstanding bug that was not super obvious to me is the RLS integration, but hopefully we'll get it fixed too. -- I guess in short the question is, given that this is MVP, how bad is anything short of "future incompatibility"?

tmandry (May 28 2019 at 17:25, on Zulip):

I think we're close to landing the work, but there may be cases that aren't fixed yet

centril (May 28 2019 at 17:25, on Zulip):

(I don't think I ever agreed to it not being a blocker on the meeting)

nikomatsakis (May 28 2019 at 17:25, on Zulip):

My take is probably "not particularly bad"

centril (May 28 2019 at 17:26, on Zulip):

@tmandry where are we with the discussions you, eddyb, Ralf, et. al are having?

davidtwco (May 28 2019 at 17:26, on Zulip):

A minor thing that might be worth doing (because it probably wouldn't take much effort and certainly not a blocker), could be updating the various syntax highlighting definitions to include async and await, things like rust-lang/rust.vim?

nikomatsakis (May 28 2019 at 17:27, on Zulip):

good point, we should file an issue on that

tmandry (May 28 2019 at 17:27, on Zulip):

tmandry where are we with the discussions you, eddyb, Ralf, et. al are having?

basically I'm trying to sidestep the whole issue by doing it a different way

centril (May 28 2019 at 17:27, on Zulip):

@tmandry ah; and the new different way doesn't have any hairy "can we do this" / soundness issues?

tmandry (May 28 2019 at 17:28, on Zulip):

right

centril (May 28 2019 at 17:28, on Zulip):

@davidtwco I think that's definitely not a blocker but worth filing as an issue :slight_smile:

nikomatsakis (May 28 2019 at 17:28, on Zulip):

OK, so, is the attitude of "the real blocker is future-incompat" what we are working towards?

centril (May 28 2019 at 17:29, on Zulip):

@nikomatsakis except for tests

centril (May 28 2019 at 17:29, on Zulip):

insufficient test coverage would be an absolute blocker for me... tho that's fixable by writing tests =)

nikomatsakis (May 28 2019 at 17:29, on Zulip):

I see. Yes, seems ok. I consider test coverage important for ensuring future compat :)

centril (May 28 2019 at 17:30, on Zulip):

Also a good point :P

centril (May 28 2019 at 17:30, on Zulip):

I'll try to see what sorts of tests are missing tonight and see if we can divide that work up somehow

boats (May 28 2019 at 17:30, on Zulip):

@nikomatsakis the featurefulness bugs that seem like possible blockers are not supporting multiple lifetimes and stack overflowing because the generator size is exponential

boats (May 28 2019 at 17:30, on Zulip):

otherwise i think we are in a backcompat and sound implementation attitude

boats (May 28 2019 at 17:30, on Zulip):

but i think with those two features, we should work on them as if theyre blocking and then can re-examine the situation closer to also, if they run into pretty serious issues

nikomatsakis (May 28 2019 at 17:31, on Zulip):

ok, I guess we will revisit those two major bugs somewhat closer. I plan to invest some time this week into the multiple lifetimes question specifically, I'd like us to make progress as deadline is getting overly close :)

nikomatsakis (May 28 2019 at 17:31, on Zulip):

yeah, seems reasonable.

nikomatsakis (May 28 2019 at 17:31, on Zulip):

Should we turn to discussing documentation?

centril (May 28 2019 at 17:31, on Zulip):

16 days until the 13th from tomorrow

nikomatsakis (May 28 2019 at 17:32, on Zulip):

16 days until the 13th from tomorrow

a good reminder :)

nikomatsakis (May 28 2019 at 17:32, on Zulip):

Documentation goals

centril (May 28 2019 at 17:32, on Zulip):

@nikomatsakis I'd like to discuss the test coverage drive

nikomatsakis (May 28 2019 at 17:32, on Zulip):

so a lot of people have talked about the lack of docs -- and I think it's safe to assume that stabilization will lead to a lot of new eyeballs...

centril (May 28 2019 at 17:33, on Zulip):

are people interested in helping out with writing tests?

centril (May 28 2019 at 17:33, on Zulip):

(cause writing all the tests alone is something I can do, but it gets boring and repetitive after some time...)

nikomatsakis (May 28 2019 at 17:33, on Zulip):

I'm trying to decide if I can responsibly volunteer any time :) I'm definitely interested at least in so far as I have some ideas for edge cases i'd like to make sure we cover and I can try to write those up

boats (May 28 2019 at 17:34, on Zulip):

On docs, talking to @Yoshua Wuyts and @stjepang earlier, we wanted to focus on the "low level book" that explains the futures model and our async/await

centril (May 28 2019 at 17:34, on Zulip):

@nikomatsakis if you give me directions I can try to cover the edge cases if you don't have the time

nikomatsakis (May 28 2019 at 17:35, on Zulip):

@centril (why don't we schedule 30 min either later today or tomorrow to talk over the test effort? we can discuss in a separate topic, I'll open it now)

Yoshua Wuyts (May 28 2019 at 17:35, on Zulip):

Referring to seeing through the work @Taylor Cramer has started in https://github.com/rust-lang/async-book

boats (May 28 2019 at 17:35, on Zulip):

I think thats probably the thing to focus on now while our ecosystem still catches up

nikomatsakis (May 28 2019 at 17:36, on Zulip):

The good thing is that this has a deadline of august

boats (May 28 2019 at 17:37, on Zulip):

what about documentation we ship with the release; that can go in beta backports or it needs to be done july 4?

nikomatsakis (May 28 2019 at 17:37, on Zulip):

How much do we think we need to have "described" and by when? e.g., the existing TOC seems pretty decent..

centril (May 28 2019 at 17:37, on Zulip):

@boats docs backports are usually safe so I think so

nikomatsakis (May 28 2019 at 17:37, on Zulip):

..but a lot of the sections are blank

Florian Gilcher (May 28 2019 at 17:37, on Zulip):

Hm, I'm not sure if I agree with the direction of focusing on a low level book. Docs for how futures work in Rust are ~2 years delayed already and people want to become productive. There's _tons_ of tribal knowledge in that space. Not starting with that kind of documentation delays us even further.

centril (May 28 2019 at 17:37, on Zulip):

cc @Pietro Albini on backport

nikomatsakis (May 28 2019 at 17:38, on Zulip):

(Which docs are you thinking of, @boats?)

centril (May 28 2019 at 17:39, on Zulip):

the async-book is a good start

boats (May 28 2019 at 17:39, on Zulip):

Don't we ship the rust book with the release, and shouldn't there be docs in there?

centril (May 28 2019 at 17:40, on Zulip):

@boats we do; but we don't have @Steve Klabnik or carosl10cents here atm

stjepang (May 28 2019 at 17:40, on Zulip):

I think the low level book would be interesting to people writing executors or async bindings, whereas most people would want to read an async cookbook with more practical examples. We're ready for a low-level book now, but the ecosystem is still not mature enough for a cookbook (unfortunately).

boats (May 28 2019 at 17:40, on Zulip):

@Florian Gilcher Well I'm a little confused by your comment because I would describe "how futures work" as the low level book.

boats (May 28 2019 at 17:41, on Zulip):

@Florian Gilcher Obviously people also want to be productive and ideally we'd be ready for it, but with most of the ecosystem not even using the compatible futures API yet, the ecosystem doesn't feel stable enough for us to push that documentation in my opinion.

Florian Gilcher (May 28 2019 at 17:41, on Zulip):

Maybe I'm not sure what the exact scope of such a book would be? But we definitely need something that goes into "how do I build an application out of that"?

nikomatsakis (May 28 2019 at 17:41, on Zulip):

Don't we ship the rust book with the release, and shouldn't there be docs in there?

You mean https://doc.rust-lang.org/book/, @boats? I don't think this tries to be complete, but it does feel like it'd be good for it to at least link to other documentation. I'll reach out to carol10cents perhaps and get her take.

davidtwco (May 28 2019 at 17:43, on Zulip):

I don't follow this part of the ecosystem that closely, so this might be a silly question, but how usable (in terms of, what practical things can I build with this now w/out writing everything myself) is async/await w/out the ecosystem being more mature? Is that something we'd want to mention in the release post?

nikomatsakis (May 28 2019 at 17:43, on Zulip):

Maybe I'm not sure what the exact scope of such a book would be? But we definitely need something that goes into "how do I build an application out of that"?

I am pretty sympathetic with this, but I wonder what the timeline should be on it -- i.e., if this stabilization is framed as "stabilizing the building blocks", we can perhaps explicitly scope such docs as coming later, or as a "work in progress"

boats (May 28 2019 at 17:44, on Zulip):

Yea I think the main thing we need is to communicate clearly about the lack of such a book and how we are pushing that forward

Florian Gilcher (May 28 2019 at 17:45, on Zulip):

I'm not happy with that. That also means that we keep the knowledge about how such things should be built in a very small circle.

Florian Gilcher (May 28 2019 at 17:45, on Zulip):

IMHO, The mechanics of how async applications should be constructed around futures stay largely the same with async/await, with a couple of patterns becoming drastically easier.

Florian Gilcher (May 28 2019 at 17:46, on Zulip):

But e.g. even simple questions like "how do I test a future?" are a question of digging through current implementations.

nikomatsakis (May 28 2019 at 17:47, on Zulip):

in practice, one has to combine async-await with some runtime, a la runtime or tokio, correct?

nikomatsakis (May 28 2019 at 17:47, on Zulip):

how much is true "across" those projects?

centril (May 28 2019 at 17:47, on Zulip):

But e.g. even simple questions like "how do I test a future?" are a question of digging through current implementations.

Hmm... I would have guessed "the same way you test any trait, by testing out .poll(..)"

Yoshua Wuyts (May 28 2019 at 17:47, on Zulip):

@Florian Gilcher if I'm hearing you correctly, what you're saying is that in order for async/await to be useful to people on release, we should at least describe the basic patterns of how to use them?

Florian Gilcher (May 28 2019 at 17:47, on Zulip):

@Yoshua Wuyts yes

Yoshua Wuyts (May 28 2019 at 17:48, on Zulip):

@Florian Gilcher I'm very sympathetic to that position -- I believe most people will agree with you here

boats (May 28 2019 at 17:48, on Zulip):

But what I think we can't expect is answers to questions like.. how do I make an HTTP request? the reality is the ecosystem doesn't know yet

nikomatsakis (May 28 2019 at 17:48, on Zulip):

I know that if I were looking for docs, I would not care so much about the Future trait, and would be looking for something more like https://github.com/rustasync/runtime/tree/master/examples =)

centril (May 28 2019 at 17:48, on Zulip):

The point of stabilizing futures in 1.36.0 was to let the ecosystem important crates catch up... how much of this has happened?

nikomatsakis (May 28 2019 at 17:48, on Zulip):

I feel like we could enumerate some of the options -- or link to docs from within the projects themselves

Yoshua Wuyts (May 28 2019 at 17:49, on Zulip):

@Florian Gilcher the way I'm interpreting what's being said so far is a matter of prioritization -- in the next 3 months, while more parts of the ecosystem is being developed, we first focus on documenting the underlying layers. Once those are done, we move up the stack.

Nemo157 (May 28 2019 at 17:49, on Zulip):

But e.g. even simple questions like "how do I test a future?" are a question of digging through current implementations.

That's the sort of thing that is still missing from the ecosystem, I have been trying to contribute to the futures-test crate utilities for this as I need them, but there's still a lot missing for testing a lot of usecases

Taylor Cramer (May 28 2019 at 17:49, on Zulip):

@nikomatsakis IMO if you start typing examples like that for more than 10 minutes without reading about the Future trait, you're going to be vv confused

Yoshua Wuyts (May 28 2019 at 17:49, on Zulip):

@Florian Gilcher with the goal of covering both before the stabilization

Taylor Cramer (May 28 2019 at 17:50, on Zulip):

because you're going to misunderstand something fundamental about how the system works that isn't solved by just typing .await

Yoshua Wuyts (May 28 2019 at 17:50, on Zulip):

(At least that's how I'm thinking about approaching documentation. Do people agree with this?)

nikomatsakis (May 28 2019 at 17:51, on Zulip):

I agree @Yoshua Wuyts I'm just not sure about the timeline. I guess you mean: we should be aiming to have that stuff done by mid-August, when async-await syntax hits stable?

Yoshua Wuyts (May 28 2019 at 17:52, on Zulip):

@nikomatsakis I think that would make a good goal, yes

davidtwco (May 28 2019 at 17:53, on Zulip):

(At least that's how I'm thinking about approaching documentation. Do people agree with this?)

I think this makes a lot of sense, we just need to make sure that it is properly communicated in any announcements/release notes.

centril (May 28 2019 at 17:55, on Zulip):

I guess you mean: we should be aiming to have that stuff done by mid-August, when async-await syntax hits stable?

@nikomatsakis It would be good to have people assigned for this work.

centril (May 28 2019 at 17:55, on Zulip):

so we don't just say a bunch of nice things... :slight_smile:

Yoshua Wuyts (May 28 2019 at 17:55, on Zulip):

@Taylor Cramer given it's your work, would you be okay if we started to move the async book forward again?

nikomatsakis (May 28 2019 at 17:55, on Zulip):

That's what I'm thinking about indeed, @centril.

Florian Gilcher (May 28 2019 at 17:55, on Zulip):

(At least that's how I'm thinking about approaching documentation. Do people agree with this?)

I'm fine with any approach, but have the clear wish that there's docs and a roadmap for early adopter on the _users_ side. I appreciate that it can't be completely covering.

Taylor Cramer (May 28 2019 at 17:56, on Zulip):

@Yoshua Wuyts I'd love that! I'd be happy to help, as well

Yoshua Wuyts (May 28 2019 at 17:56, on Zulip):

@Taylor Cramer fantastic!

Taylor Cramer (May 28 2019 at 17:57, on Zulip):

@Yoshua Wuyts should we start a separate stream to discuss book plans inside this zulip, or elsewhere?

Yoshua Wuyts (May 28 2019 at 17:57, on Zulip):

@Taylor Cramer zulip sounds good -- can you start one?

nikomatsakis (May 28 2019 at 17:57, on Zulip):

so we're coming up on an hour. I'm going to try and condense this into notes, I think @Florian Gilcher is definitely raising a good requirement, and I think we should think about how we'll make sure we have enough. Not sure if we're going to work it all out now. :)

centril (May 28 2019 at 17:58, on Zulip):

(Would be good to have everything in one stream I think)

centril (May 28 2019 at 18:00, on Zulip):

so... do I have any more volunteers for #Test coverage? :slight_smile:

nikomatsakis (May 28 2019 at 18:00, on Zulip):

can y'all take a look at the work items list here?

nikomatsakis (May 28 2019 at 18:00, on Zulip):

just let me know if you see anything major missing

nikomatsakis (May 28 2019 at 18:01, on Zulip):

I'm going to look into figuring out how to make sure this stuff is happening :)

Last update: Nov 18 2019 at 01:30UTC