Stream: wg-async-foundations

Topic: meeting 2019.12.03


nikomatsakis (Dec 03 2019 at 18:00, on Zulip):

Hey @WG-async-foundations, sync meeting?

nikomatsakis (Dec 03 2019 at 18:00, on Zulip):

I'm just back from vacation last week

nikomatsakis (Dec 03 2019 at 18:04, on Zulip):

Uncategorized issues

nikomatsakis (Dec 03 2019 at 18:04, on Zulip):

ICE on rustc 1.41.0-nightly (25d8a9494 2019-11-29) running on x86_64-apple-darwin #66958

nikomatsakis (Dec 03 2019 at 18:04, on Zulip):

narrowed down to https://github.com/rust-lang/rust/pull/66567

nikomatsakis (Dec 03 2019 at 18:05, on Zulip):

@Esteban Küber is on it, marking it as "triaged"

nikomatsakis (Dec 03 2019 at 18:05, on Zulip):

Slightly incorrect help message with mismatched typs in async code #66910

nikomatsakis (Dec 03 2019 at 18:05, on Zulip):

Very interesting

nikomatsakis (Dec 03 2019 at 18:06, on Zulip):

@Esteban Küber -- if you're around -- do you know if we do any suggestions that advise you to add .await? is there an issue on that?

nikomatsakis (Dec 03 2019 at 18:06, on Zulip):

I'm going to go ahead and tag this as "on deck", seems like a good change to make

nikomatsakis (Dec 03 2019 at 18:07, on Zulip):

async_await RFC is not fully implemented, but tracking issue is closed #66909

nikomatsakis (Dec 03 2019 at 18:07, on Zulip):

specifically, it seems like there is missing docs for the keywords

Esteban Küber (Dec 03 2019 at 18:08, on Zulip):

It is quite a corner case but I think it makes sense to address. No need for me to fast track it, is there?
There is a ticket for suggesting await that is partly covered

nikomatsakis (Dec 03 2019 at 18:08, on Zulip):

@Esteban Küber I'm interested more in the general case of suggesting await

nikomatsakis (Dec 03 2019 at 18:08, on Zulip):

then in this specific case

Esteban Küber (Dec 03 2019 at 18:08, on Zulip):

The more complex cases like these will be trickier to handle, my worry is for foo.bar.future.baz

Esteban Küber (Dec 03 2019 at 18:08, on Zulip):

That was the reason to keep the original ticket open

Esteban Küber (Dec 03 2019 at 18:09, on Zulip):

I'm on mobile now, but I can fish it out later

nikomatsakis (Dec 03 2019 at 18:09, on Zulip):

I imagined we could check for any case where you have a type error between impl Future<Output = T> and T

nikomatsakis (Dec 03 2019 at 18:09, on Zulip):

ok

Esteban Küber (Dec 03 2019 at 18:09, on Zulip):

And link the tickets

nikomatsakis (Dec 03 2019 at 18:09, on Zulip):

thanks

Esteban Küber (Dec 03 2019 at 18:09, on Zulip):

Yeah

Esteban Küber (Dec 03 2019 at 18:09, on Zulip):

One thing to decide, right now we o ly suggest on 2018 and don't mention async await in 2015

Esteban Küber (Dec 03 2019 at 18:09, on Zulip):

Should we?

nikomatsakis (Dec 03 2019 at 18:10, on Zulip):

not sure, good question

nikomatsakis (Dec 03 2019 at 18:10, on Zulip):

I'd be inclined to say "probably yes"

nikomatsakis (Dec 03 2019 at 18:10, on Zulip):

but I could go either way

centril (Dec 03 2019 at 18:10, on Zulip):

iirc we do suggest it in 2015 somewhere; iirc I added that

centril (Dec 03 2019 at 18:11, on Zulip):

https://github.com/rust-lang/rust/pull/63539

nikomatsakis (Dec 03 2019 at 18:12, on Zulip):

that looks perfect, yeah

nikomatsakis (Dec 03 2019 at 18:13, on Zulip):

More friendly error msg when await on NONE ASYNC fn/block or return a obj that implements deprecated Future #66731

nikomatsakis (Dec 03 2019 at 18:13, on Zulip):

very confusing issue title

nikomatsakis (Dec 03 2019 at 18:13, on Zulip):

but basically the reverse problem

nikomatsakis (Dec 03 2019 at 18:14, on Zulip):

seems like a good suggestion, I will also "on deck" this one

centril (Dec 03 2019 at 18:15, on Zulip):

hah; interesting -- so basically the suggestion is "add async before fn"

nikomatsakis (Dec 03 2019 at 18:15, on Zulip):

Error in async function/block produces errors for valid unrelated expressions #66634

nikomatsakis (Dec 03 2019 at 18:16, on Zulip):

this does seem annoying, but I'm not going to bump the priority I don't think

centril (Dec 03 2019 at 18:16, on Zulip):

hah; interesting -- so basically the suggestion is "add async before fn"

seems like it could be a fairly targeted fix in typeck

nikomatsakis (Dec 03 2019 at 18:16, on Zulip):

yeah, that's what I was thinking too

nikomatsakis (Dec 03 2019 at 18:16, on Zulip):

I should leave a comment to that effect

nikomatsakis (Dec 03 2019 at 18:16, on Zulip):

I'm not sure if "on-unimplemented" is necessarily the right vector here

centril (Dec 03 2019 at 18:18, on Zulip):

could be worth a try to see if that mechanism is cheaper before trying something custom

nikomatsakis (Dec 03 2019 at 18:18, on Zulip):

on-unimplemented would certainly be an easy way, not clear how important this case is

nikomatsakis (Dec 03 2019 at 18:19, on Zulip):

you won't get the nice suggestion

nikomatsakis (Dec 03 2019 at 18:19, on Zulip):

well anyway

nikomatsakis (Dec 03 2019 at 18:20, on Zulip):

rustc crash on 1.39.0 stable with combination of async and .. #66618

nikomatsakis (Dec 03 2019 at 18:21, on Zulip):

"crash" here means ICE

centril (Dec 03 2019 at 18:21, on Zulip):

renamed ^^

nikomatsakis (Dec 03 2019 at 18:21, on Zulip):

looks like there is a fix already

nikomatsakis (Dec 03 2019 at 18:21, on Zulip):

and it was beta-nominated etc

nikomatsakis (Dec 03 2019 at 18:22, on Zulip):

I'll mark as triaged, move on

nikomatsakis (Dec 03 2019 at 18:22, on Zulip):

RFC: process-handle for future async child-processes-term-handling #2817

Esteban Küber (Dec 03 2019 at 18:22, on Zulip):

It's needstest only

nikomatsakis (Dec 03 2019 at 18:22, on Zulip):

er, I guess that's a side-effect of using the "org-wide search" :)

centril (Dec 03 2019 at 18:23, on Zulip):

If only async/await was stable in 2012 :D

nikomatsakis (Dec 03 2019 at 18:23, on Zulip):

I'm not really sure why that is tagged A-async-await

nikomatsakis (Dec 03 2019 at 18:23, on Zulip):

but anyway

nikomatsakis (Dec 03 2019 at 18:23, on Zulip):

Symbols in optimized async programs are often not useful #65978

nikomatsakis (Dec 03 2019 at 18:24, on Zulip):

I guess I'll mark this as "triaged"

nikomatsakis (Dec 03 2019 at 18:25, on Zulip):

seems like a "nic to have"

centril (Dec 03 2019 at 18:25, on Zulip):

@Steven Fackler landed a PR relevant to this one (which I reviewed)

centril (Dec 03 2019 at 18:25, on Zulip):

(https://github.com/rust-lang/rust/pull/66398)

nikomatsakis (Dec 03 2019 at 18:27, on Zulip):

cool

nikomatsakis (Dec 03 2019 at 18:27, on Zulip):

ok, that's all the untriaged issue, and..that's the whole 30 minutes

nikomatsakis (Dec 03 2019 at 18:28, on Zulip):

I'll observe that we've lost some momentum, but, hey, it's the holidays. Seems ok. :)

centril (Dec 03 2019 at 18:28, on Zulip):

:slight_smile:

nikomatsakis (Dec 03 2019 at 18:28, on Zulip):

(We do need to figure out how we're going to organize this group in a bigger sense, I think, I'm still wondering where the "leadership energy" will come from. Anyway who thinks they'd like to play more of a leadership and organizational role should ping me...)

Last update: Jun 07 2020 at 10:40UTC