Stream: t-compiler

Topic: weekly meeting 2019-10-03 #54818


pnkfelix (Oct 03 2019 at 12:01, on Zulip):

Hi @T-compiler/meeting ; the triage meeting will be starting in 1 hours 59 minutes

pnkfelix (Oct 03 2019 at 12:02, on Zulip):

I will be doing pre-triage in a parallel topic

pnkfelix (Oct 03 2019 at 12:02, on Zulip):

For WG-checkin this week, we have Async/Await and Diagnostics scheduled

pnkfelix (Oct 03 2019 at 12:03, on Zulip):

@Taylor Cramer , are you around to give a checkin for WG-async-foundations?

pnkfelix (Oct 03 2019 at 12:03, on Zulip):

@oli or @Esteban Küber , are you around to give a checking for WG-diagnostics ?

pnkfelix (Oct 03 2019 at 14:01, on Zulip):

Hi @T-compiler/meeting now the meeting will start

pnkfelix (Oct 03 2019 at 14:01, on Zulip):

we will dedicate five minutes at the start to announcements; anyone can chime in

centril (Oct 03 2019 at 14:02, on Zulip):

EUV and rvalue promotion checks are being simplified following up on AST-borrowck removal: https://github.com/rust-lang/rust/pull/64874

pnkfelix (Oct 03 2019 at 14:02, on Zulip):

:exclamation: tomorrow's design meeting will be on the 2020 roadmap; see compiler-team#188

pnkfelix (Oct 03 2019 at 14:02, on Zulip):

(I had hoped to have written some text with thoughts on the roadmap by now. but it hasn't happened yet.)

nikomatsakis (Oct 03 2019 at 14:03, on Zulip):

Taylor Cramer , are you around to give a checkin for WG-async-foundations?

(I can)

pnkfelix (Oct 03 2019 at 14:06, on Zulip):

okay, so, as usual, I didn't get through all the P-high issues during triage this week

pnkfelix (Oct 03 2019 at 14:06, on Zulip):

but there were "only" 36

nikomatsakis (Oct 03 2019 at 14:06, on Zulip):

(I had hoped to have written some text with thoughts on the roadmap by now. but it hasn't happened yet.)

(maybe we can find some time to chat this over before meeting, @pnkfelix)

pnkfelix (Oct 03 2019 at 14:06, on Zulip):

which I think is less than last week. So, according to calculus, all is well.

pnkfelix (Oct 03 2019 at 14:06, on Zulip):

(because, you know, in the limit we have zero P-high bugs)

pnkfelix (Oct 03 2019 at 14:07, on Zulip):

um anyway lets go over the beta-nominations

pnkfelix (Oct 03 2019 at 14:07, on Zulip):

there are 8 nominated PR's (6 of which have landed on nightly)

pnkfelix (Oct 03 2019 at 14:08, on Zulip):

beta nom: "Always mark rust and rust-call abi's as unwind" #65020

nagisa (Oct 03 2019 at 14:08, on Zulip):

I’m surprised this managed to happen.

pnkfelix (Oct 03 2019 at 14:08, on Zulip):

as in, how we didn't see this bug crop up more often?

nagisa (Oct 03 2019 at 14:09, on Zulip):

no, how this happened in the first place :slight_smile:

pnkfelix (Oct 03 2019 at 14:09, on Zulip):

or as in, how our review process didn't manage to grep for the relevant extern "Rust" { block

pnkfelix (Oct 03 2019 at 14:09, on Zulip):

yeah I know

pnkfelix (Oct 03 2019 at 14:09, on Zulip):

its kind of bonkers

pnkfelix (Oct 03 2019 at 14:09, on Zulip):

anyway, looks uncontroversial (as it was deliberately authored to be beta backported)

pnkfelix (Oct 03 2019 at 14:09, on Zulip):

beta-accepted.

nagisa (Oct 03 2019 at 14:09, on Zulip):

I wrote down a comment on the PR with my request for a codegen test

pnkfelix (Oct 03 2019 at 14:10, on Zulip):

okay i'll look into that

centril (Oct 03 2019 at 14:10, on Zulip):

shouldn't the PR be r-ed meanwhile?

centril (Oct 03 2019 at 14:10, on Zulip):

otherwise it will just be rolled up and merged

pnkfelix (Oct 03 2019 at 14:10, on Zulip):

becasue of nagisa's request?

centril (Oct 03 2019 at 14:10, on Zulip):

yeah

pnkfelix (Oct 03 2019 at 14:10, on Zulip):

I took that to mean it could be a follow-on work

centril (Oct 03 2019 at 14:11, on Zulip):

@nagisa clarify?

nagisa (Oct 03 2019 at 14:11, on Zulip):

Yeah could definitely be a follow-up PR

pnkfelix (Oct 03 2019 at 14:11, on Zulip):

okay

pnkfelix (Oct 03 2019 at 14:11, on Zulip):

beta-nom: " extract expected return type for async fn generators" #64999

centril (Oct 03 2019 at 14:12, on Zulip):

hasn't merged yet but in the pipeline

nagisa (Oct 03 2019 at 14:12, on Zulip):

looks fairly large to me, so shrug.

pnkfelix (Oct 03 2019 at 14:12, on Zulip):

does anyone want to argue it should wait until next week for consideration?

centril (Oct 03 2019 at 14:12, on Zulip):

I think we can prioritize some async PRs a bit more if they are not risky

pnkfelix (Oct 03 2019 at 14:12, on Zulip):

ah the size, I see

centril (Oct 03 2019 at 14:12, on Zulip):

fwiw, most of the size are tests :P

centril (Oct 03 2019 at 14:13, on Zulip):

(which is a good thing...)

nagisa (Oct 03 2019 at 14:13, on Zulip):

Even without tests this is fairly more sizable than the couple of lines for a fix I’m otherwise used to; but this being a fix to the typesystem, I guess it is warranted?

nagisa (Oct 03 2019 at 14:13, on Zulip):

also we don't want broken async in first stable release.

pnkfelix (Oct 03 2019 at 14:13, on Zulip):

it may not be a trivial backport given that it needed a rebase commit, not sure

nikomatsakis (Oct 03 2019 at 14:14, on Zulip):

I can squash, I think it would be an easy backport,

nikomatsakis (Oct 03 2019 at 14:14, on Zulip):

I do agree it's not entirely trivial though

centril (Oct 03 2019 at 14:14, on Zulip):

It's a "completeness fix" so to speak

centril (Oct 03 2019 at 14:14, on Zulip):

accepting code that was previously rejected

nikomatsakis (Oct 03 2019 at 14:14, on Zulip):

that said, I think there is some... mild backwards compat risk? like, we don't know of a case where there's a problem because of a missed coercion opportunit

eddyb (Oct 03 2019 at 14:14, on Zulip):

it's a bit sad that it has to know about Future at all, as opposed to deducing through the relevant impls

nikomatsakis (Oct 03 2019 at 14:14, on Zulip):

but I would be happier not to have the possibility

pnkfelix (Oct 03 2019 at 14:14, on Zulip):

doesn't seem like anyone's vetoing, but we could also wait a week safely, if I understand correctly

nikomatsakis (Oct 03 2019 at 14:15, on Zulip):

it's a bit sad that it has to know about Future at all, as opposed to deducing through the relevant impls

(yes, but that was by far the most practical way, and changing to do more general deduction would be a less targeted fix, too, that I'd less comfortable backporting)

centril (Oct 03 2019 at 14:15, on Zulip):

@nikomatsakis fwiw we don't guarantee stability for inference in that way at least "by the letter of the law"

pnkfelix (Oct 03 2019 at 14:15, on Zulip):

how about we just wait a week since niko themselves mentioned the backwards compat risk

nikomatsakis (Oct 03 2019 at 14:15, on Zulip):

I know, i'm not overly worried about it

nikomatsakis (Oct 03 2019 at 14:16, on Zulip):

this is why I didn't try to block stabilization on it :)

pnkfelix (Oct 03 2019 at 14:16, on Zulip):

(as in, leave the beta-nominated tag for now)

pnkfelix (Oct 03 2019 at 14:16, on Zulip):

sound okay?

nikomatsakis (Oct 03 2019 at 14:16, on Zulip):

that seems ok

centril (Oct 03 2019 at 14:16, on Zulip):

seems fair

pnkfelix (Oct 03 2019 at 14:16, on Zulip):

beta-nom: "Fix ICE #64964" #64989

centril (Oct 03 2019 at 14:16, on Zulip):

This is linked to https://github.com/rust-lang/rust/pull/64895

nikomatsakis (Oct 03 2019 at 14:16, on Zulip):

this should be considered together with https://github.com/rust-lang/rust/pull/64895

pnkfelix (Oct 03 2019 at 14:16, on Zulip):

i guess this was also an await related bug?

centril (Oct 03 2019 at 14:16, on Zulip):

yep

eddyb (Oct 03 2019 at 14:16, on Zulip):

methinks we should have a policy of editing PR titles so they include information

centril (Oct 03 2019 at 14:17, on Zulip):

either we backport both or none

nikomatsakis (Oct 03 2019 at 14:17, on Zulip):

in particular, #64989 is a fix to #64895, and it makes no sense to backport one without the other

nikomatsakis (Oct 03 2019 at 14:17, on Zulip):

/me stops repeating everything centril says

pnkfelix (Oct 03 2019 at 14:17, on Zulip):

okay lets treat them as coupled, we'll get to #64895 in a moment then

pnkfelix (Oct 03 2019 at 14:18, on Zulip):

beta-nom: "Silence unreachable code lint from await desugaring" #64930

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

I'm of two minds about this one, as the fix is not entirely trivial, and it's "just" diagnostics, but it's also a fairly major diagnostics fail (as I noted in my comment on #64895)

pnkfelix (Oct 03 2019 at 14:18, on Zulip):

so much await stuff!

centril (Oct 03 2019 at 14:19, on Zulip):

One could argue that because a fix was needed to the first PR that it is a bit risky?

centril (Oct 03 2019 at 14:19, on Zulip):

on the other hand we want less serious diagnostics fails

pnkfelix (Oct 03 2019 at 14:20, on Zulip):

well it seems like #64930 is okay, in any case?

pnkfelix (Oct 03 2019 at 14:20, on Zulip):

beta accepted

centril (Oct 03 2019 at 14:20, on Zulip):

agree

pnkfelix (Oct 03 2019 at 14:20, on Zulip):

okay now what you've all been chomping at the bit for

pnkfelix (Oct 03 2019 at 14:20, on Zulip):

beta-nom: "async/await: improve not-send errors" #64895

centril (Oct 03 2019 at 14:21, on Zulip):

it's a non-trivial amount of code, that needed a fix, but it's mostly error code

centril (Oct 03 2019 at 14:21, on Zulip):

difficult decision :P

centril (Oct 03 2019 at 14:22, on Zulip):

we could wait a week... but testing it sooner rather than later is also beneficial

nikomatsakis (Oct 03 2019 at 14:22, on Zulip):

yeah, that about sums it up

centril (Oct 03 2019 at 14:22, on Zulip):

(I'll stop arguing against myself back and forth...)

nikomatsakis (Oct 03 2019 at 14:22, on Zulip):

I guess waiting a week will also give us more testing (since more async-await folks are on nightly anyway, afaik?)

pnkfelix (Oct 03 2019 at 14:23, on Zulip):

the bulk of the PR is indeed diagnostics ... I don't know if "boilerplate" is a fair term here, but its definitely less risky than some other kinds of code

pnkfelix (Oct 03 2019 at 14:23, on Zulip):

I guess waiting a week will also give us more testing (since more async-await folks are on nightly anyway, afaik?)

this is my big question, indeed: How many people are both on beta and testing async-await ...?

centril (Oct 03 2019 at 14:23, on Zulip):

the blog post might have sent some folks to beta

pnkfelix (Oct 03 2019 at 14:24, on Zulip):

I'm inclined to go for it

nikomatsakis (Oct 03 2019 at 14:24, on Zulip):

I feel like most people test on beta in CI, nightly or stable in their day to day, but I guess I don't know for sure :)

centril (Oct 03 2019 at 14:24, on Zulip):

that's probably right ^--

pnkfelix (Oct 03 2019 at 14:24, on Zulip):

(as in, backport both #64895 and #64989 )

centril (Oct 03 2019 at 14:24, on Zulip):

I think I'm mildly in favor of backport as well

nikomatsakis (Oct 03 2019 at 14:25, on Zulip):

one other relevant factor: we expect to do a bit of follow-up work; I think I'm leaning mildly towards backport this work but not the follow-up

nikomatsakis (Oct 03 2019 at 14:25, on Zulip):

just because it'll have less testing, and this is the most important part

nikomatsakis (Oct 03 2019 at 14:25, on Zulip):

plus that way we get some improvements to deliver in 1.40 :P

pnkfelix (Oct 03 2019 at 14:25, on Zulip):

apart from follow-up work along the same lines as PR #64989, right?

nikomatsakis (Oct 03 2019 at 14:25, on Zulip):

right, of course

pnkfelix (Oct 03 2019 at 14:26, on Zulip):

what's the follow-up work?

nikomatsakis (Oct 03 2019 at 14:26, on Zulip):

I mean purely "aesthetic" improvements, not ICEs etc

pnkfelix (Oct 03 2019 at 14:26, on Zulip):

or maybe that's better discussed at a different time

nikomatsakis (Oct 03 2019 at 14:26, on Zulip):

basically just making the message less noisy, tweaking some of the wording

pnkfelix (Oct 03 2019 at 14:26, on Zulip):

oh yeah, okay

pnkfelix (Oct 03 2019 at 14:26, on Zulip):

cool.

pnkfelix (Oct 03 2019 at 14:26, on Zulip):

beta-accepted then.

pnkfelix (Oct 03 2019 at 14:26, on Zulip):

beta-nom: "syntax: fix dropping of attribute on first param of non-method assocated fn" #64894

centril (Oct 03 2019 at 14:27, on Zulip):

this is a fix to a newly stabilized feature as well (in 1.39)

centril (Oct 03 2019 at 14:27, on Zulip):

(which production users have hit)

pnkfelix (Oct 03 2019 at 14:28, on Zulip):

I'm a little surprised the fix was this big

pnkfelix (Oct 03 2019 at 14:28, on Zulip):

from the bug I had assumed this was a like a fencepost error or something. :wink:

pnkfelix (Oct 03 2019 at 14:28, on Zulip):

having said that, I don't object. It does seem like a real serious bug

nikomatsakis (Oct 03 2019 at 14:29, on Zulip):

it is a bit more involved than I expected, but we gotta fix that

centril (Oct 03 2019 at 14:29, on Zulip):

@pnkfelix the actual fix is in parser.rs -- most of the rest is just a whole lot of tests

pnkfelix (Oct 03 2019 at 14:29, on Zulip):

I'm reading over parse/parser/item.rs and just surprised

pnkfelix (Oct 03 2019 at 14:29, on Zulip):

did you refactor or something? add new recovery code?

centril (Oct 03 2019 at 14:30, on Zulip):

@pnkfelix yes I baked parsing of self params into parse_param_general because the old approach was fundamentally broken

centril (Oct 03 2019 at 14:30, on Zulip):

(and this is more "clearly right")

pnkfelix (Oct 03 2019 at 14:30, on Zulip):

okay I'll assume that it was more trouble than it would have been worth to try to make a more targetted (as in minimal diff) fix

centril (Oct 03 2019 at 14:30, on Zulip):

yep

centril (Oct 03 2019 at 14:31, on Zulip):

(did consider it)

pnkfelix (Oct 03 2019 at 14:31, on Zulip):

beta-accepted then, unless I hear a veto in the meantime

pnkfelix (Oct 03 2019 at 14:31, on Zulip):

beta-nom: "Upgrade async/await to "used" keywords." #64875

pnkfelix (Oct 03 2019 at 14:31, on Zulip):

I guess this is fine?

centril (Oct 03 2019 at 14:32, on Zulip):

not a big deal but improves some diagnostics

centril (Oct 03 2019 at 14:32, on Zulip):

small code changes also

pnkfelix (Oct 03 2019 at 14:32, on Zulip):

(sorry I keep forgetting to add the emoji votes)

pnkfelix (Oct 03 2019 at 14:32, on Zulip):

yep, only Q I had was whether it was worth the hassle

pnkfelix (Oct 03 2019 at 14:32, on Zulip):

it probably is

pnkfelix (Oct 03 2019 at 14:32, on Zulip):

since patch is relatively trivial

pnkfelix (Oct 03 2019 at 14:33, on Zulip):

okay i'll call it beta-accepted unless someone vetos in the meantime

pnkfelix (Oct 03 2019 at 14:33, on Zulip):

last one

pnkfelix (Oct 03 2019 at 14:33, on Zulip):

beta-nom: "Fix redundant semicolon lint interaction with proc macro attributes" #64387

pnkfelix (Oct 03 2019 at 14:34, on Zulip):

oh is this really just changing the pretty-print behavior?

pnkfelix (Oct 03 2019 at 14:35, on Zulip):

/me skims

centril (Oct 03 2019 at 14:35, on Zulip):

yea but macros...

nikomatsakis (Oct 03 2019 at 14:35, on Zulip):

(sorry I keep forgetting to add the emoji votes)

we should double check whether poll works on mobile or whatever now

Vadim Petrochenkov (Oct 03 2019 at 14:36, on Zulip):

#64387 and the previous semicolon PRs walk on such a minefield and don't even realize it.

Vadim Petrochenkov (Oct 03 2019 at 14:36, on Zulip):

I didn't block them only because I was going to rewrite all that later anyway.

pnkfelix (Oct 03 2019 at 14:36, on Zulip):

yea but macros...

I'm sorry, do macros use the pretty-printer?

centril (Oct 03 2019 at 14:36, on Zulip):

oh yeah

nikomatsakis (Oct 03 2019 at 14:36, on Zulip):

/me quickly puts his hands over his ears, "not listening!"

pnkfelix (Oct 03 2019 at 14:36, on Zulip):

I know they generate token streams, for some reason I assumed those token-streams were generated through other means

pnkfelix (Oct 03 2019 at 14:37, on Zulip):

I don't know why I thought this

nikomatsakis (Oct 03 2019 at 14:37, on Zulip):

heh yes it's a dark secret :)

pnkfelix (Oct 03 2019 at 14:37, on Zulip):

goodness gracious

nikomatsakis (Oct 03 2019 at 14:37, on Zulip):

we... should probably try to change that at some point, before we get really stuck stuck

pnkfelix (Oct 03 2019 at 14:37, on Zulip):

roadmap 2020/2021!

pnkfelix (Oct 03 2019 at 14:38, on Zulip):

okay so, @Vadim Petrochenkov , what do you think, is this something you would backport?

eddyb (Oct 03 2019 at 14:38, on Zulip):

I wanted to help with that earlier this year but muffled screams got sidetracked

pnkfelix (Oct 03 2019 at 14:38, on Zulip):

I'm personally inclined to say no, but mostly because I don't want to believe we're in this situation in the first place.

eddyb (Oct 03 2019 at 14:39, on Zulip):

it really has to be a main priority to get fixed (since we need to redesign a few things around ASTs, macro expansion, etc.)

eddyb (Oct 03 2019 at 14:39, on Zulip):

the bad news is that it's hard/tricky the good news is that getting it right would also help incremental :P

centril (Oct 03 2019 at 14:39, on Zulip):

Denial is always a strategy that... works?

nikomatsakis (Oct 03 2019 at 14:39, on Zulip):

okay so, Vadim Petrochenkov , what do you think, is this something you would backport?

you're referring here to #64387, @pnkfelix ?

pnkfelix (Oct 03 2019 at 14:39, on Zulip):

or maybe it never works and could never possibly work?

nikomatsakis (Oct 03 2019 at 14:40, on Zulip):

it looks pretty small

pnkfelix (Oct 03 2019 at 14:40, on Zulip):

okay so, Vadim Petrochenkov , what do you think, is this something you would backport?

you're referring here to #64387, pnkfelix ?

yes, PR #64387

pnkfelix (Oct 03 2019 at 14:40, on Zulip):

#64387 and the previous semicolon PRs walk on such a minefield and don't even realize it.

^ this is what I was referencing when I tried to pull @Vadim Petrochenkov back in

Vadim Petrochenkov (Oct 03 2019 at 14:41, on Zulip):

If redundant_semicolon which introduced the issue is already on beta, then we probably have to backport.

centril (Oct 03 2019 at 14:41, on Zulip):

it should be on beta

nikomatsakis (Oct 03 2019 at 14:42, on Zulip):

ok. Yeah, I had assumed it would only make sense to skip #64387 if we also skipped whatever it's fixing.

nikomatsakis (Oct 03 2019 at 14:42, on Zulip):

(which I don't think is on the table...)

centril (Oct 03 2019 at 14:42, on Zulip):

confirm; I checked playground

pnkfelix (Oct 03 2019 at 14:42, on Zulip):

okay then

pnkfelix (Oct 03 2019 at 14:42, on Zulip):

beta-accepted

pnkfelix (Oct 03 2019 at 14:42, on Zulip):

what else what esle

pnkfelix (Oct 03 2019 at 14:42, on Zulip):

nominations!

pnkfelix (Oct 03 2019 at 14:42, on Zulip):

we had two nominations, then I did triage and decided that two was not a large enough number

pnkfelix (Oct 03 2019 at 14:43, on Zulip):

so now we have four nominatins

pnkfelix (Oct 03 2019 at 14:43, on Zulip):

nominations

pnkfelix (Oct 03 2019 at 14:43, on Zulip):

I-nom: "Add an unstable 'extra randomization' compile-time flag for hashers" #65042

pnkfelix (Oct 03 2019 at 14:44, on Zulip):

basically someone wants a flag to allow us to increase randomization in how hash table walk, for improving test coverage of otherwise hidden sources of under-determinism

centril (Oct 03 2019 at 14:44, on Zulip):

I think this feature as-is isn't tenable, it's just too problematic in practice. You'll probably never finish a compiler build if it does have high overhead, and I suspect it will.

I don't get this... would someone like to elaborate, e.g. @simulacrum ?

centril (Oct 03 2019 at 14:44, on Zulip):

to be clear, this would only be for local usage, not on CI

centril (Oct 03 2019 at 14:44, on Zulip):

with like... fast machines :slight_smile:

pnkfelix (Oct 03 2019 at 14:45, on Zulip):

well, I wanted to bring it up here to see if anyone thought this was worth discussing

centril (Oct 03 2019 at 14:45, on Zulip):

imo the non-determinism in diagnostics is a fairly annoying problem when it shows up

pnkfelix (Oct 03 2019 at 14:45, on Zulip):

in which case I would want to perhaps discuss it in a design meeting

nikomatsakis (Oct 03 2019 at 14:45, on Zulip):

this seems like it could be pretty useful

pnkfelix (Oct 03 2019 at 14:45, on Zulip):

not here

pnkfelix (Oct 03 2019 at 14:45, on Zulip):

it seems like @centril is interested in it. Is anyone else?

nikomatsakis (Oct 03 2019 at 14:46, on Zulip):

I feel like we don't need a design meeting per se

nikomatsakis (Oct 03 2019 at 14:46, on Zulip):

I'd be ok with it, esp as it could be helpful if we want to push on deterministic builds

pnkfelix (Oct 03 2019 at 14:46, on Zulip):

(I think it could tie into e.g. deterministic builds, maybe)

centril (Oct 03 2019 at 14:46, on Zulip):

I'm guessing the actual PR won't be too involved in terms of diff

simulacrum (Oct 03 2019 at 14:46, on Zulip):

I think it's not solving the problem the right way, but am unopposed otherwise

pnkfelix (Oct 03 2019 at 14:46, on Zulip):

okay I think there may be general interest in talking about determinism

pnkfelix (Oct 03 2019 at 14:47, on Zulip):

and this would be one facet of that.

pnkfelix (Oct 03 2019 at 14:47, on Zulip):

is that fair?

pnkfelix (Oct 03 2019 at 14:47, on Zulip):

I think that's fair

nikomatsakis (Oct 03 2019 at 14:47, on Zulip):

(that said, I agree with @simulacrum that we should experiment with just using IndexMap and getting a determinsitic order all the time :)

pnkfelix (Oct 03 2019 at 14:47, on Zulip):

let's move along

pnkfelix (Oct 03 2019 at 14:47, on Zulip):

(as in, we can talk determinism out side of this meeting. and I will not close the issue.)

pnkfelix (Oct 03 2019 at 14:47, on Zulip):

I-nom: " ICE building Apache Arrow PR in librustc_traits/normalize_erasing_regions.rs:42" #64908

pnkfelix (Oct 03 2019 at 14:48, on Zulip):

why did I leave this nominated

centril (Oct 03 2019 at 14:49, on Zulip):

let's sincerely hope this isn't a "mir-opt triggered long-standing lazy-norm-issue" :slight_smile:

pnkfelix (Oct 03 2019 at 14:49, on Zulip):

I think that was a mistake on my part; at least, I dont think there's much to discuss here. (This may be a bug that needs lazy-normalization. Or it may have another solution.)

centril (Oct 03 2019 at 14:49, on Zulip):

(cause then we are royally screwed)

pnkfelix (Oct 03 2019 at 14:49, on Zulip):

next

pnkfelix (Oct 03 2019 at 14:49, on Zulip):

I-nom: "ThinLTO and -o compiler option can lead to duplicate object file inclusion in staticlib" #64153

pnkfelix (Oct 03 2019 at 14:49, on Zulip):

:construction_worker: someone volunteer to take the above, please!

pnkfelix (Oct 03 2019 at 14:50, on Zulip):

(its assigned to @mw to delegate but I think they did not see that.)

pnkfelix (Oct 03 2019 at 14:50, on Zulip):

anyway that's enough begging out of me ... or is it ...

pnkfelix (Oct 03 2019 at 14:50, on Zulip):

I-nom: "regression: rustdoc resolution changes" #61560

pnkfelix (Oct 03 2019 at 14:50, on Zulip):

:construction_worker: someone volunteer to take the above, please!

pnkfelix (Oct 03 2019 at 14:51, on Zulip):

Seriously this one may be a matter of just "can you even replicate this? If you can, can you bisect it?"

pnkfelix (Oct 03 2019 at 14:51, on Zulip):

I just haven't had time to do it myself.

pnkfelix (Oct 03 2019 at 14:51, on Zulip):

Okay now that's enough begging out of me.

nikomatsakis (Oct 03 2019 at 14:51, on Zulip):

OK so

pnkfelix (Oct 03 2019 at 14:51, on Zulip):

Maybe I'll tweet out some more begging tonight

nikomatsakis (Oct 03 2019 at 14:52, on Zulip):

once we launch the LLVM ICE-breakers, we need a similar "Bisecters and minimizers" group, don't we?

nikomatsakis (Oct 03 2019 at 14:52, on Zulip):

ooh yeah we should also have a compiler twitter account that automatically tweets any issue identified as such

nikomatsakis (Oct 03 2019 at 14:52, on Zulip):

I'm going to add that to the compiler-team#174 :)

pnkfelix (Oct 03 2019 at 14:52, on Zulip):

and maybe tweets again if they go unassigned lose their assignment

pnkfelix (Oct 03 2019 at 14:53, on Zulip):

though honestly I have not looked into how issue-assignments work in the ICE-breakers system

pnkfelix (Oct 03 2019 at 14:53, on Zulip):

okay okay

nagisa (Oct 03 2019 at 14:53, on Zulip):

FWIW you got 8 mins remaining and the check-ins on the table

pnkfelix (Oct 03 2019 at 14:53, on Zulip):

that's all the nominations

pnkfelix (Oct 03 2019 at 14:53, on Zulip):

yeah

pnkfelix (Oct 03 2019 at 14:53, on Zulip):

thanks @nagisa

pnkfelix (Oct 03 2019 at 14:53, on Zulip):

SO

pnkfelix (Oct 03 2019 at 14:54, on Zulip):

WG-async-await

pnkfelix (Oct 03 2019 at 14:54, on Zulip):

@nikomatsakis you think you can fit it into 4 minutes? We've already covered some stuff (with the beta-nom discussion, that is)

nikomatsakis (Oct 03 2019 at 14:54, on Zulip):

Yeah I'll be fast

nikomatsakis (Oct 03 2019 at 14:54, on Zulip):

TL;DR is this:

nikomatsakis (Oct 03 2019 at 14:54, on Zulip):
nikomatsakis (Oct 03 2019 at 14:54, on Zulip):
nikomatsakis (Oct 03 2019 at 14:54, on Zulip):
nikomatsakis (Oct 03 2019 at 14:55, on Zulip):

Basically we have a big pool of bugs and it's hard to differentiate

nikomatsakis (Oct 03 2019 at 14:55, on Zulip):

So in our meeting, we triage out uncategorized ones, then we check in on focus bugs, and if we need to create more work for people, we pull from "on deck" and replace as needed from "other"

nikomatsakis (Oct 03 2019 at 14:55, on Zulip):

so, if you have some async-await bug you think should be higher priority than the rest, go ahead and leave a note in #wg-async-foundations -- or maybe we should create some nomination scheme (maybe just add the OnDeck label...)

nikomatsakis (Oct 03 2019 at 14:56, on Zulip):

That said, we've already done some good polish work since shipping:

nikomatsakis (Oct 03 2019 at 14:56, on Zulip):

I didn't really prep a list :P

pnkfelix (Oct 03 2019 at 14:56, on Zulip):

Sounds great

pnkfelix (Oct 03 2019 at 14:57, on Zulip):

Might be good at some point to try to collect/compare these different triaging systems

pnkfelix (Oct 03 2019 at 14:57, on Zulip):

but not now

pnkfelix (Oct 03 2019 at 14:57, on Zulip):

because ...

pnkfelix (Oct 03 2019 at 14:57, on Zulip):

well I don't actually know

pnkfelix (Oct 03 2019 at 14:57, on Zulip):

is @oli or @Esteban Küber around ?

pnkfelix (Oct 03 2019 at 14:57, on Zulip):

for a 3-minute checkin from WG-diagnostics ?

pnkfelix (Oct 03 2019 at 14:57, on Zulip):

I'm suspecting the answer is "no"

pnkfelix (Oct 03 2019 at 14:59, on Zulip):

okay then everyone

pnkfelix (Oct 03 2019 at 14:59, on Zulip):

Good meeting all!

pnkfelix (Oct 03 2019 at 14:59, on Zulip):

thanks for coming @T-compiler/meeting

nikomatsakis (Oct 03 2019 at 14:59, on Zulip):

:wave:

nikomatsakis (Oct 03 2019 at 14:59, on Zulip):

thanks @pnkfelix!

lqd (Oct 03 2019 at 16:32, on Zulip):

I-nom: "regression: rustdoc resolution changes" #61560

@pnkfelix I'll take it

Last update: Nov 16 2019 at 01:00UTC