Stream: t-compiler/meetings

Topic: [weekly meeting] 2020-08-13 #54818


Santiago Pastorino (Aug 12 2020 at 18:18, on Zulip):

Hi @T-compiler/meeting; the triage meeting will happen tomorrow at 2pm UTC

Santiago Pastorino (Aug 12 2020 at 18:19, on Zulip):

The @WG-prioritization have done pre-triage in #t-compiler/wg-prioritization/alerts

Santiago Pastorino (Aug 12 2020 at 18:19, on Zulip):

@WG-prioritization have prepared the meeting agenda

Santiago Pastorino (Aug 12 2020 at 18:19, on Zulip):

We will have checkins from @WG-rls2.0 and @WG-self-profile

Santiago Pastorino (Aug 12 2020 at 18:20, on Zulip):

@matklad do you have something you want to share about @WG-rls2.0?

Santiago Pastorino (Aug 12 2020 at 18:20, on Zulip):

@Wesley Wiser do you have something you want to share about @WG-self-profile?

Santiago Pastorino (Aug 12 2020 at 18:23, on Zulip):

if someone has something they wanted to discuss it may be a good opportunity to nominate some issues

Santiago Pastorino (Aug 13 2020 at 13:08, on Zulip):

Hi @T-compiler/meeting, triage meeting will be starting in ~ 52 minutes

Santiago Pastorino (Aug 13 2020 at 13:09, on Zulip):

Check out the meeting agenda

Santiago Pastorino (Aug 13 2020 at 13:11, on Zulip):

@matklad @Wesley Wiser is there something to share about @WG-rls2.0 and @WG-self-profile?

Santiago Pastorino (Aug 13 2020 at 13:11, on Zulip):

if so, please fill checkins directly in the agenda

Wesley Wiser (Aug 13 2020 at 13:14, on Zulip):

Will do!

matklad (Aug 13 2020 at 13:25, on Zulip):

Ah, sorry, thanks for the reminder, I'll fill this up now!

matklad (Aug 13 2020 at 13:29, on Zulip):

@Santiago Pastorino i think I need edit rights for the note

Joshua Nelson (Aug 13 2020 at 13:32, on Zulip):

not sure if this is the right place to mention it, but I was interested in adding a new S-waiting-on-help tag, could that get added to the agenda? (near the end is fine, it's not high priority)

Joshua Nelson (Aug 13 2020 at 13:32, on Zulip):

https://rust-lang.zulipchat.com/#narrow/stream/185694-t-compiler.2Fwg-meta/topic/Add.20S-waiting-on-help.3F

Wesley Wiser (Aug 13 2020 at 13:32, on Zulip):

@matklad What's your HackMD username? I think I can add you

matklad (Aug 13 2020 at 13:33, on Zulip):

@matklad

Wesley Wiser (Aug 13 2020 at 13:35, on Zulip):

@matklad Done

Santiago Pastorino (Aug 13 2020 at 14:03, on Zulip):

Joshua Nelson said:

not sure if this is the right place to mention it, but I was interested in adding a new S-waiting-on-help tag, could that get added to the agenda? (near the end is fine, it's not high priority)

I guess after announcements maybe @pnkfelix can give some space to discuss this

pnkfelix (Aug 13 2020 at 14:04, on Zulip):

sorry all for delay; laptop. is being pokey

pnkfelix (Aug 13 2020 at 14:04, on Zulip):

(and/or github?)

pnkfelix (Aug 13 2020 at 14:05, on Zulip):

or maybe its finally time for me to switch to Chrome

pnkfelix (Aug 13 2020 at 14:05, on Zulip):

Hi @T-compiler/meeting! Add a :wave: emoji to show you're here :)

pnkfelix (Aug 13 2020 at 14:05, on Zulip):

we will start off with 5 minutes for ...

Announcements

pnkfelix (Aug 13 2020 at 14:06, on Zulip):
pnkfelix (Aug 13 2020 at 14:06, on Zulip):
pnkfelix (Aug 13 2020 at 14:06, on Zulip):
pnkfelix (Aug 13 2020 at 14:07, on Zulip):

heck i'm going to go second compiler-team#338 now

simulacrum (Aug 13 2020 at 14:07, on Zulip):

perf.rlo gained support for rendering flame graphs, e.g., https://perf.rust-lang.org/perf/processed-self-profile?commit=c2d1b0d9800d922b0451921d2ce17e6ae665d5b4&benchmark=serde-debug&run_name=full&type=flamegraph

this is a bit slow since we download and look at the raw self profile data each time

pnkfelix (Aug 13 2020 at 14:07, on Zulip):
pnkfelix (Aug 13 2020 at 14:07, on Zulip):
nikomatsakis (Aug 13 2020 at 14:08, on Zulip):

compiler-team#332 has actually been FCP merge'd

pnkfelix (Aug 13 2020 at 14:08, on Zulip):
pnkfelix (Aug 13 2020 at 14:08, on Zulip):
pnkfelix (Aug 13 2020 at 14:09, on Zulip):

So, a sort of non-standard announcement

pnkfelix (Aug 13 2020 at 14:09, on Zulip):

I'm sure you all are aware that Mozilla let go of a number of employees yesterday.

pnkfelix (Aug 13 2020 at 14:11, on Zulip):

I am one of those employees. More broadly, the set of people Mozilla let go basically represents a vast reduction in their investment in Rust

Santiago Pastorino (Aug 13 2020 at 14:11, on Zulip):

very sad :(

pnkfelix (Aug 13 2020 at 14:11, on Zulip):

(or rather, the staff reduction was announced on Tuesday, not yesterday)

nikomatsakis (Aug 13 2020 at 14:11, on Zulip):

I should say that I was not laid off, nor njn

pnkfelix (Aug 13 2020 at 14:11, on Zulip):

anyway, I've been thinking a lot about what all this means since then.

pnkfelix (Aug 13 2020 at 14:12, on Zulip):

For example: I've often said that the goals of Rust are very much aligned with. Mozilla's goals

pnkfelix (Aug 13 2020 at 14:12, on Zulip):

but now, for various reasons, Mozilla has decided they cannot afford to continue that investment

pnkfelix (Aug 13 2020 at 14:12, on Zulip):

in any case

pnkfelix (Aug 13 2020 at 14:13, on Zulip):

while I still think the goals of Rust were/are. aligned with the goals of Mozilla

pnkfelix (Aug 13 2020 at 14:13, on Zulip):

in the end, they are different missions

pnkfelix (Aug 13 2020 at 14:13, on Zulip):

and, in my semi-humble opinion

pnkfelix (Aug 13 2020 at 14:13, on Zulip):

I think Rust's mission is more important than Mozilla's.

pnkfelix (Aug 13 2020 at 14:14, on Zulip):

this is all a way to say that I'm planning to stick around, at least for the short term

pnkfelix (Aug 13 2020 at 14:14, on Zulip):

I'll continue on as T-compiler member, and co-lead, as long as I can

pnkfelix (Aug 13 2020 at 14:15, on Zulip):

(I'm still trying to figure out whether it makes sense for me to. continue in other roles, like T-lang. member. But that's less relevant to this meeting.)

pnkfelix (Aug 13 2020 at 14:15, on Zulip):

so, yeah. Not the typical announcement

pnkfelix (Aug 13 2020 at 14:16, on Zulip):

but the good news is that a lot of Rust's needs have been independent of Mozilla for a while

pnkfelix (Aug 13 2020 at 14:16, on Zulip):

and we'll have to see what we can do to. further decouple the two

pnkfelix (Aug 13 2020 at 14:17, on Zulip):

so lets see, were there other announcements?

pnkfelix (Aug 13 2020 at 14:17, on Zulip):

WG checkins

pnkfelix (Aug 13 2020 at 14:17, on Zulip):

@WG-rls2.0 checkin by @matklad:

nikomatsakis (Aug 13 2020 at 14:18, on Zulip):

Ooh, I hadn't seen the Microsoft FOSS announcement

pnkfelix (Aug 13 2020 at 14:19, on Zulip):

@WG-self-profile checkin by @wesleywiser:

pnkfelix (Aug 13 2020 at 14:19, on Zulip):

(omg on the rdpmc stuff)

eddyb (Aug 13 2020 at 14:19, on Zulip):

600x less noise than time measurements, baby!

eddyb (Aug 13 2020 at 14:20, on Zulip):

(OTOH, if you look at wall-time vs the other measurements on perf.r-l.o you can easily see why that would be the case)

pnkfelix (Aug 13 2020 at 14:20, on Zulip):

You can get things like cache miss counts from. rdpmc, right?

pnkfelix (Aug 13 2020 at 14:20, on Zulip):

e.g., pure instruction counts can sometimes not really reflect the reality of where time is actually spent.

eddyb (Aug 13 2020 at 14:20, on Zulip):

yeah, sadly it would probably be too expensive to record more than one thing, but we can experiment with it

I've tried not to mess with the measureme event format but at least now we have a precise way to quantify the overhead of profiling (e.g. I now roughly know that getting one value out of rdpmc, compared to a hardcoded 0, is about 20 instructions)

pnkfelix (Aug 13 2020 at 14:21, on Zulip):

okay next. up

pnkfelix (Aug 13 2020 at 14:21, on Zulip):

Beta-nominations

T-compiler

pnkfelix (Aug 13 2020 at 14:21, on Zulip):
pnkfelix (Aug 13 2020 at 14:23, on Zulip):

yeah this sounds right

pnkfelix (Aug 13 2020 at 14:23, on Zulip):

its. basically just adding a new type inference variable, or at least turning a T == V into T <: V (or maybe other way around)

nikomatsakis (Aug 13 2020 at 14:24, on Zulip):

yeah the change looked pretty harmless

pnkfelix (Aug 13 2020 at 14:24, on Zulip):

that, or maybe harmful !

nikomatsakis (Aug 13 2020 at 14:24, on Zulip):

basically if it was typed as T: Index<U>, we convert to T: IndexMut<U> with the same U

pnkfelix (Aug 13 2020 at 14:24, on Zulip):

I mean, we've definitely had cases where. you needed invariance

nikomatsakis (Aug 13 2020 at 14:24, on Zulip):

unless I am confused

nikomatsakis (Aug 13 2020 at 14:24, on Zulip):

I don't think this is adding a new variable

pnkfelix (Aug 13 2020 at 14:25, on Zulip):

oh right, its not

nikomatsakis (Aug 13 2020 at 14:25, on Zulip):

more the opposite probably

pnkfelix (Aug 13 2020 at 14:25, on Zulip):

I missed the note that it was looking it up somewhere

pnkfelix (Aug 13 2020 at 14:25, on Zulip):

okay then, this is probably just a bug fix then

pnkfelix (Aug 13 2020 at 14:25, on Zulip):

I get it now

pnkfelix (Aug 13 2020 at 14:25, on Zulip):

anyway, beta-accepted

pnkfelix (Aug 13 2020 at 14:26, on Zulip):
eddyb (Aug 13 2020 at 14:26, on Zulip):

ugh speaking of things assigned to me, please PM me for reviews, I'm so hopelessly behind everything

nikomatsakis (Aug 13 2020 at 14:26, on Zulip):

/me hides

nikomatsakis (Aug 13 2020 at 14:27, on Zulip):

I was hoping to catch up this week on notifications and reviews :)

pnkfelix (Aug 13 2020 at 14:27, on Zulip):

the bug fixed by #75319 is not a beta. regression, right?

pnkfelix (Aug 13 2020 at 14:28, on Zulip):

or its regression status is not tagged. on the issue.../.]

pnkfelix (Aug 13 2020 at 14:28, on Zulip):

description indicates it is stable-to-stable regression

Santiago Pastorino (Aug 13 2020 at 14:28, on Zulip):

it seems a stable regression

eddyb (Aug 13 2020 at 14:28, on Zulip):

it would have to be, yeah, test ICEs on stable

pnkfelix (Aug 13 2020 at 14:28, on Zulip):

I'm sort of "shrug" on this one

pnkfelix (Aug 13 2020 at 14:28, on Zulip):

it could ride the trains

nikomatsakis (Aug 13 2020 at 14:29, on Zulip):

is this an ICE any time that there are many {}?

nikomatsakis (Aug 13 2020 at 14:29, on Zulip):

like, extra ones?

nikomatsakis (Aug 13 2020 at 14:29, on Zulip):

because that's... fairly common for me, anyway

nikomatsakis (Aug 13 2020 at 14:29, on Zulip):

I'd be inclined to at least beta backport

nikomatsakis (Aug 13 2020 at 14:30, on Zulip):

(usually occurs when I modify the format string)

pnkfelix (Aug 13 2020 at 14:30, on Zulip):

the minized verrsion seems. way more subtle: https://github.com/rust-lang/rust/issues/75307#issuecomment-670978358

pnkfelix (Aug 13 2020 at 14:30, on Zulip):

. Reduced format invocation:

rust format!(r"{}{}{}", named_arg=1);

All of using an r"" literal, having a named argument, and referencing at least 3 positional arguments seem necessary.

nikomatsakis (Aug 13 2020 at 14:31, on Zulip):

ok ok

nikomatsakis (Aug 13 2020 at 14:31, on Zulip):

then never mind, I agree :)

pnkfelix (Aug 13 2020 at 14:31, on Zulip):

if there were evidence that it didn't require all those compoennts...

pnkfelix (Aug 13 2020 at 14:31, on Zulip):

but yeah, lets let this ride the trains then.

pnkfelix (Aug 13 2020 at 14:31, on Zulip):

libs-impl

T-rustdoc

pnkfelix (Aug 13 2020 at 14:32, on Zulip):

Stable-nominations

T-compiler

libs-impl

T-rustdoc

pnkfelix (Aug 13 2020 at 14:32, on Zulip):

PR's S-waiting-on-team

T-compiler

libs-impl

pnkfelix (Aug 13 2020 at 14:32, on Zulip):

Issues of Note

Short Summary

pnkfelix (Aug 13 2020 at 14:32, on Zulip):

P-critical

T-compiler

pnkfelix (Aug 13 2020 at 14:32, on Zulip):
pnkfelix (Aug 13 2020 at 14:33, on Zulip):

I haven't had chance to even experiment to see what impact (positive or negative) would be. of changing the range.rs code

pnkfelix (Aug 13 2020 at 14:34, on Zulip):

and I suspect the best answer continues to be to fix LLVM upstream

pnkfelix (Aug 13 2020 at 14:34, on Zulip):

but I myself may have less time to investigate ....

pnkfelix (Aug 13 2020 at 14:34, on Zulip):

well, I can leave it assigned to me for one more week.

nikomatsakis (Aug 13 2020 at 14:35, on Zulip):

hmm

pnkfelix (Aug 13 2020 at 14:36, on Zulip):
pnkfelix (Aug 13 2020 at 14:36, on Zulip):

/me wonders why the reference to rust#74960 was added on there at the end

pnkfelix (Aug 13 2020 at 14:37, on Zulip):

(cut-and-paste error. along the way?)

Santiago Pastorino (Aug 13 2020 at 14:37, on Zulip):

right

nikomatsakis (Aug 13 2020 at 14:37, on Zulip):

looks like this could just be assigned to @Matthew Jasper

pnkfelix (Aug 13 2020 at 14:38, on Zulip):

it is in fact (self-)assigned to them

nikomatsakis (Aug 13 2020 at 14:38, on Zulip):

or if they are short of time maybe they can identify the commit that requires backport

nikomatsakis (Aug 13 2020 at 14:38, on Zulip):

/me is excited for https://github.com/rust-lang/rust/pull/73905 to land, on a somewhat related note

pnkfelix (Aug 13 2020 at 14:39, on Zulip):

I wonder if its one of the commits. that says "avoid cycle[s] ..."

pnkfelix (Aug 13 2020 at 14:39, on Zulip):

anyway

pnkfelix (Aug 13 2020 at 14:39, on Zulip):

when is the next release? How much time do we have here?

lcnr (Aug 13 2020 at 14:40, on Zulip):

14 days

pnkfelix (Aug 13 2020 at 14:40, on Zulip):

okay

lcnr (Aug 13 2020 at 14:40, on Zulip):

27.08

nikomatsakis (Aug 13 2020 at 14:40, on Zulip):

pnkfelix said:

I wonder if its one of the commits. that says "avoid cycle[s] ..."

lol sounds...likely

pnkfelix (Aug 13 2020 at 14:40, on Zulip):

probably be good to make sure this is actually resolved by next week's meeting then

Santiago Pastorino (Aug 13 2020 at 14:40, on Zulip):

pnkfelix said:

when is the next release? How much time do we have here?

btw @XAMPPRocky has created a calendar to track Rust releases, I should have said that in announcements

simulacrum (Aug 13 2020 at 14:40, on Zulip):

more like, 10 days, really

Matthew Jasper (Aug 13 2020 at 14:41, on Zulip):

I'll try to get a PR for this up today.

pnkfelix (Aug 13 2020 at 14:41, on Zulip):

Thanks @Matthew Jasper :heart:

pnkfelix (Aug 13 2020 at 14:41, on Zulip):

last one

pnkfelix (Aug 13 2020 at 14:41, on Zulip):
Santiago Pastorino (Aug 13 2020 at 14:41, on Zulip):

Santiago Pastorino said:

pnkfelix said:

when is the next release? How much time do we have here?

btw XAMPPRocky has created a calendar to track Rust releases, I should have said that in announcements

https://calendar.google.com/calendar?cid=bDFiMWdrcXZmYmd1bmpzMThuZW1xNGM1ODBAZ3JvdXAuY2FsZW5kYXIuZ29vZ2xlLmNvbQ

lcnr (Aug 13 2020 at 14:42, on Zulip):

there is a potentially cleaner fix for this in #75443

pnkfelix (Aug 13 2020 at 14:42, on Zulip):

sounds like some great team work between @eddyb and @RalfJ

pnkfelix (Aug 13 2020 at 14:42, on Zulip):

/me wonders if @RalfJ should be subscribed to this stream. Probably, right?

lcnr (Aug 13 2020 at 14:43, on Zulip):

which needs signoff by either @nikomatsakis or @Matthew Jasper afaict

nikomatsakis (Aug 13 2020 at 14:43, on Zulip):

lcnr said:

there is a potentially cleaner fix for this in #75443

ah, that also fixes the regression?

eddyb (Aug 13 2020 at 14:43, on Zulip):

we likely don't want to land #75419, it's a hacky workaround for the bug #75443 fixes directly

eddyb (Aug 13 2020 at 14:43, on Zulip):

@nikomatsakis yeah I uhhh screwed up a long time ago

pnkfelix (Aug 13 2020 at 14:43, on Zulip):

yes it sounds like @eddyb prefers PR #75443

eddyb (Aug 13 2020 at 14:43, on Zulip):

somehow this hasn't bitten us until recently

nikomatsakis (Aug 13 2020 at 14:43, on Zulip):

that...

pnkfelix (Aug 13 2020 at 14:43, on Zulip):

@lcnr can we assign #75313 to you ?

nikomatsakis (Aug 13 2020 at 14:43, on Zulip):

seems like a fairly safe PR

lcnr (Aug 13 2020 at 14:43, on Zulip):

yeah

pnkfelix (Aug 13 2020 at 14:44, on Zulip):

great

pnkfelix (Aug 13 2020 at 14:44, on Zulip):

in terms of timing

nikomatsakis (Aug 13 2020 at 14:44, on Zulip):

I can r+ that now I think

pnkfelix (Aug 13 2020 at 14:44, on Zulip):

we can afford. to wait a week before discussing a beta-nomination of PR #75313 ?

pnkfelix (Aug 13 2020 at 14:44, on Zulip):

or should we just quickly beta-nominate it now?

pnkfelix (Aug 13 2020 at 14:45, on Zulip):

@simulacrum ^ ?

simulacrum (Aug 13 2020 at 14:45, on Zulip):

up to you

pnkfelix (Aug 13 2020 at 14:45, on Zulip):

letswait a week

pnkfelix (Aug 13 2020 at 14:45, on Zulip):

let it land on. nightly

simulacrum (Aug 13 2020 at 14:45, on Zulip):

there won't be more beta crater or whatever

pnkfelix (Aug 13 2020 at 14:45, on Zulip):

okay. thanks

simulacrum (Aug 13 2020 at 14:45, on Zulip):

and we can backport tonight/tomorrow or 7 days from now

pnkfelix (Aug 13 2020 at 14:45, on Zulip):

libs-impl

T-rustdoc

pnkfelix (Aug 13 2020 at 14:45, on Zulip):

Unassigned P-high regressions

Beta regressions

pnkfelix (Aug 13 2020 at 14:46, on Zulip):
nikomatsakis (Aug 13 2020 at 14:48, on Zulip):

huh

nikomatsakis (Aug 13 2020 at 14:48, on Zulip):

that bisection is interesting

pnkfelix (Aug 13 2020 at 14:48, on Zulip):

should we reach out more directly to @Petr Hosek and @tmandry ?

pnkfelix (Aug 13 2020 at 14:49, on Zulip):

this note in description on rust#72746 is interesting: "This matches the CMake build of libunwind. "

pnkfelix (Aug 13 2020 at 14:49, on Zulip):

does that imply that CMake's build would suffer from same problem?

pnkfelix (Aug 13 2020 at 14:49, on Zulip):

seems like. there's. lots of investigate here

pnkfelix (Aug 13 2020 at 14:50, on Zulip):

but not a terrible amount of time

pnkfelix (Aug 13 2020 at 14:50, on Zulip):

what tier is this platform? does this only happen under QEMU armv7, not hardware armv7 ?

nikomatsakis (Aug 13 2020 at 14:50, on Zulip):

that's a good question

pnkfelix (Aug 13 2020 at 14:51, on Zulip):

ah niko's comments on issue point out

pnkfelix (Aug 13 2020 at 14:51, on Zulip):

this is due to lock poisoning ?

pnkfelix (Aug 13 2020 at 14:51, on Zulip):

i.e. it could be some sort of race condition issue?

nikomatsakis (Aug 13 2020 at 14:51, on Zulip):

PoisonError { inner: ..

nikomatsakis (Aug 13 2020 at 14:51, on Zulip):

(from the output)

simulacrum (Aug 13 2020 at 14:51, on Zulip):

yeah something is getting poisoned it seems, based on the backtrace in the description (need to open a details toggle)

nikomatsakis (Aug 13 2020 at 14:52, on Zulip):

oh, I missed the backtrace toggle actually

nikomatsakis (Aug 13 2020 at 14:52, on Zulip):

but it's not that interesting

pnkfelix (Aug 13 2020 at 14:52, on Zulip):

well the. most immediate. question is

pnkfelix (Aug 13 2020 at 14:52, on Zulip):

can. anyone take charge. on this

pnkfelix (Aug 13 2020 at 14:53, on Zulip):

I don't think I have time in next week to relearn how to setup and properly use QEMU here

nikomatsakis (Aug 13 2020 at 14:53, on Zulip):

I wonder if @Robin Randhawa might have a suggestion for someone

pnkfelix (Aug 13 2020 at 14:53, on Zulip):

but if this is tier1 platform (still unclear) then we should prioritize it

nikomatsakis (Aug 13 2020 at 14:53, on Zulip):

/me hopes he got the right Robin

pnkfelix (Aug 13 2020 at 14:53, on Zulip):

if

Robin Randhawa (Aug 13 2020 at 14:53, on Zulip):

Hi! You did! :)

Robin Randhawa (Aug 13 2020 at 14:54, on Zulip):

Just building some context!

simulacrum (Aug 13 2020 at 14:54, on Zulip):

it is not tier 1

simulacrum (Aug 13 2020 at 14:54, on Zulip):

we don't have tier 1 arm yet I think

nikomatsakis (Aug 13 2020 at 14:54, on Zulip):

Not yet

pnkfelix (Aug 13 2020 at 14:54, on Zulip):

oh right

nikomatsakis (Aug 13 2020 at 14:54, on Zulip):

Though we have a pending RFC for aarch64 at least :)

pnkfelix (Aug 13 2020 at 14:54, on Zulip):

/me just checked against https://forge.rust-lang.org/release/platform-support.html

pnkfelix (Aug 13 2020 at 14:54, on Zulip):

okay then

pnkfelix (Aug 13 2020 at 14:54, on Zulip):

so its not end-of-world for this to hit stable

pnkfelix (Aug 13 2020 at 14:55, on Zulip):

so okay, leave it as P-high, and unassigned, I guess.

nikomatsakis (Aug 13 2020 at 14:55, on Zulip):

Anyway @Robin Randhawa if nothing else it'd be great to know whether #74820 is specific to QEMU

Robin Randhawa (Aug 13 2020 at 14:55, on Zulip):

Here's the current RFC for AArch64 Tier-1 promotion: https://github.com/rust-lang/rfcs/pull/2959

pnkfelix (Aug 13 2020 at 14:55, on Zulip):

Nightly regressions

Robin Randhawa (Aug 13 2020 at 14:55, on Zulip):

Will check!

Robin Randhawa (Aug 13 2020 at 14:56, on Zulip):

Hmm. On the face of it it may not be a qemu problem....

Robin Randhawa (Aug 13 2020 at 14:56, on Zulip):

I'll set some time aside to dig a bit deeper!

pnkfelix (Aug 13 2020 at 14:56, on Zulip):

pnkfelix said:

seems to me like this is maybe not so high priority?

pnkfelix (Aug 13 2020 at 14:57, on Zulip):

its mostly an issue that one error has been replaced with a worse (less helpful and/or specific) error, right?

nikomatsakis (Aug 13 2020 at 14:57, on Zulip):

yeah

nikomatsakis (Aug 13 2020 at 14:57, on Zulip):

I remember seeing that go by

pnkfelix (Aug 13 2020 at 14:57, on Zulip):

should it remain P-high?

nikomatsakis (Aug 13 2020 at 14:57, on Zulip):

trying to remember if I had formed an opinion

pnkfelix (Aug 13 2020 at 14:57, on Zulip):

I'm tempted to downgrade to P-medium, even though it is a regression

nikomatsakis (Aug 13 2020 at 14:57, on Zulip):

oh right because it was bisected to me

nikomatsakis (Aug 13 2020 at 14:57, on Zulip):

:)

pnkfelix (Aug 13 2020 at 14:58, on Zulip):

anyway I want to make fwd progress on myg

pnkfelix (Aug 13 2020 at 14:58, on Zulip):

mtg

pnkfelix (Aug 13 2020 at 14:58, on Zulip):

so I won't decide. now!

pnkfelix (Aug 13 2020 at 14:58, on Zulip):

Performance logs

Triage done by simulacrum.
Revision range: 19cefa68640843956eedd86227ddc1d35dbc6754..8b84156c6ee2b4e707dc32f5a516e3143ab924d3

1 regression, 1 improvement, 1 of them on rollups.
0 outstanding nags from last week.

Santiago Pastorino (Aug 13 2020 at 14:58, on Zulip):

https://rust-lang.zulipchat.com/#narrow/stream/245100-t-compiler.2Fwg-prioritization.2Falerts/topic/I-prioritize.20.2374400.20Type.20mismatch.20in.20function.20arguments.20E063

pnkfelix (Aug 13 2020 at 14:58, on Zulip):

Regressions

pnkfelix (Aug 13 2020 at 14:58, on Zulip):

Improvements

pnkfelix (Aug 13 2020 at 14:58, on Zulip):

New nags to follow up on

Santiago Pastorino (Aug 13 2020 at 14:58, on Zulip):

Santiago Pastorino said:

https://rust-lang.zulipchat.com/#narrow/stream/245100-t-compiler.2Fwg-prioritization.2Falerts/topic/I-prioritize.20.2374400.20Type.20mismatch.20in.20function.20arguments.20E063

it was prioritized as P-high as a mejor diagnostic regression

pnkfelix (Aug 13 2020 at 14:59, on Zulip):

Santiago Pastorino said:

Santiago Pastorino said:

https://rust-lang.zulipchat.com/#narrow/stream/245100-t-compiler.2Fwg-prioritization.2Falerts/topic/I-prioritize.20.2374400.20Type.20mismatch.20in.20function.20arguments.20E063

it was prioritized as P-high as a mejor diagnostic regression

fair enough

Santiago Pastorino (Aug 13 2020 at 14:59, on Zulip):

although it can be changed, but we can continue discussing on that thread if you want :)

pnkfelix (Aug 13 2020 at 14:59, on Zulip):

seems like there was some follow-up on-issue on #75174 ?

pnkfelix (Aug 13 2020 at 15:00, on Zulip):

performance review seems okay to me, in any case

pnkfelix (Aug 13 2020 at 15:00, on Zulip):

Nominated Issues

T-compiler

libs-impl

pnkfelix (Aug 13 2020 at 15:00, on Zulip):

and with that

pnkfelix (Aug 13 2020 at 15:00, on Zulip):

we are at the end. of the hour !

pnkfelix (Aug 13 2020 at 15:00, on Zulip):

thanks to everyone in @T-compiler/meeting for attending!

Santiago Pastorino (Aug 13 2020 at 15:00, on Zulip):

one little thing

pnkfelix (Aug 13 2020 at 15:00, on Zulip):

and as a reminder: no design meeting tomorrow

simulacrum (Aug 13 2020 at 15:00, on Zulip):

yeah the triage logs are updated not too often; the nag will be checked at next triage

Santiago Pastorino (Aug 13 2020 at 15:00, on Zulip):

@Joshua Nelson wanted to discuss something about a new label

nikomatsakis (Aug 13 2020 at 15:00, on Zulip):

@Santiago Pastorino what was prioritized as "P-high"?

Joshua Nelson (Aug 13 2020 at 15:01, on Zulip):

S-waiting-on-help, yeah

pnkfelix (Aug 13 2020 at 15:01, on Zulip):

@nikomatsakis rust#74400

Joshua Nelson (Aug 13 2020 at 15:01, on Zulip):

https://rust-lang.zulipchat.com/#narrow/stream/185694-t-compiler.2Fwg-meta/topic/Add.20S-waiting-on-help.3F

eddyb (Aug 13 2020 at 15:01, on Zulip):

you can shorten that to S-welp

Santiago Pastorino (Aug 13 2020 at 15:01, on Zulip):

nikomatsakis said:

Santiago Pastorino what was prioritized as "P-high"?

check https://rust-lang.zulipchat.com/#narrow/stream/245100-t-compiler.2Fwg-prioritization.2Falerts/topic/I-prioritize.20.2374400.20Type.20mismatch.20in.20function.20arguments.20E063

nikomatsakis (Aug 13 2020 at 15:02, on Zulip):

I see now

Joshua Nelson (Aug 13 2020 at 15:02, on Zulip):

maybe S-needs-help if you want it to be short

Joshua Nelson (Aug 13 2020 at 15:02, on Zulip):

I'm open to bikeshedding

Joshua Nelson (Aug 13 2020 at 15:02, on Zulip):

but the reason I thought of it is https://github.com/rust-lang/rust/pull/72412#issuecomment-667217945 wasn't handled very well IMO

Joshua Nelson (Aug 13 2020 at 15:02, on Zulip):

(no fault of the triager)

Joshua Nelson (Aug 13 2020 at 15:04, on Zulip):

cc @Yuki Okushi who was interested

nikomatsakis (Aug 13 2020 at 15:04, on Zulip):

/me clicks

Esteban Küber (Aug 13 2020 at 15:04, on Zulip):

I thought this case was covered by waiting in review but I'm fine with a new label

Joshua Nelson (Aug 13 2020 at 15:05, on Zulip):

well it's not really ready for review

eddyb (Aug 13 2020 at 15:05, on Zulip):

(that's the kind of accident that would feel less strange coming from a bot)

nikomatsakis (Aug 13 2020 at 15:05, on Zulip):

I think I feel the same as @Esteban Küber

simulacrum (Aug 13 2020 at 15:05, on Zulip):

it seems like it is waiting for review, in the sense that the reviewer is the one to dispatch requests from help generally

nikomatsakis (Aug 13 2020 at 15:05, on Zulip):

i.e., I would have thought of it as "waiting for review", but maybe it's helpful to distinguish

simulacrum (Aug 13 2020 at 15:05, on Zulip):

er, requests for help

eddyb (Aug 13 2020 at 15:06, on Zulip):

if it was help from someone else, the label would make more sense hmm yeah

Joshua Nelson (Aug 13 2020 at 15:06, on Zulip):

hmm so 'waiting for review' includes PRs that aren't ready to merge?

Joshua Nelson (Aug 13 2020 at 15:06, on Zulip):

I guess that sort of makes sense

simulacrum (Aug 13 2020 at 15:06, on Zulip):

I will note that I filter my assigned PRs by waiting for review, so if that's not the label then I'd just not see it

nikomatsakis (Aug 13 2020 at 15:06, on Zulip):

to be quite honest, I find that unless the setting is waiting-for-bors or team, the labels betwen "waiting on author" and "waiting on review" etc is often either outdated or just kind of "hard to summarize"

Joshua Nelson (Aug 13 2020 at 15:06, on Zulip):

can we ask triage to be less eager to switch to waiting-on-author then?

nikomatsakis (Aug 13 2020 at 15:06, on Zulip):

maybe we need a S-unclear :)

eddyb (Aug 13 2020 at 15:07, on Zulip):

S-uperposition

eddyb (Aug 13 2020 at 15:07, on Zulip):

or we can just have multiple labels at the same time shrug

simulacrum (Aug 13 2020 at 15:07, on Zulip):

@Joshua Nelson sure, if you have concerns over labeling we can take those to #t-release/triage, I'm guessing there's adjustments that can be made

nikomatsakis (Aug 13 2020 at 15:07, on Zulip):

that said I think that it probably is useful for reviews to use the labels to signal when they are waiting for feedback (and it might be useful to help e.g. me when going over my PR queue)

nikomatsakis (Aug 13 2020 at 15:07, on Zulip):

I think my concern with filtering is usually that I worry the label is out of date

Joshua Nelson (Aug 13 2020 at 15:07, on Zulip):

another place this came up was for https://github.com/rust-lang/rust/pull/75127

Joshua Nelson (Aug 13 2020 at 15:08, on Zulip):

where I wasn't sure the right approach to take

Joshua Nelson (Aug 13 2020 at 15:08, on Zulip):

so there was nothing to review

simulacrum (Aug 13 2020 at 15:08, on Zulip):

that also seems like a review thing though?

simulacrum (Aug 13 2020 at 15:08, on Zulip):

just a different kind of review

Joshua Nelson (Aug 13 2020 at 15:08, on Zulip):

hmm I guess so

Joshua Nelson (Aug 13 2020 at 15:09, on Zulip):

ok I'll take this to wg-triage then :)

simulacrum (Aug 13 2020 at 15:09, on Zulip):

nikomatsakis said:

I think my concern with filtering is usually that I worry the label is out of date

This is true, but I guess I just hope that triage assigns the right label "soon enough" or that the author pings me and I see that

Joshua Nelson (Aug 13 2020 at 15:11, on Zulip):

oh another suggested label was S-waiting-on-fcp to go with S-waiting-on-mcp

Joshua Nelson (Aug 13 2020 at 15:12, on Zulip):

but I don't feel as strongly about that one

Joshua Nelson (Aug 13 2020 at 15:12, on Zulip):

cc @LeSeulArtichaut

DPC (Aug 13 2020 at 15:28, on Zulip):

@Joshua Nelson we already have S-waiting-on-fcp

Joshua Nelson (Aug 13 2020 at 15:29, on Zulip):

other way around, sorry, S-waiting-on-mcp

Last update: Nov 25 2020 at 02:30UTC