Stream: t-compiler/wg-prioritization

Topic: pre-meeting triage 2020-04-09 #54818


Santiago Pastorino (Apr 08 2020 at 17:01, on Zulip):

@WG-prioritization the meeting will be starting in 1h

Santiago Pastorino (Apr 08 2020 at 17:16, on Zulip):

The @WG-prioritization will be doing pre-triage in this channel (in 45 minutes)

pnkfelix (Apr 08 2020 at 17:48, on Zulip):

this pre-triage session is for this T-compiler meeting

Santiago Pastorino (Apr 08 2020 at 18:00, on Zulip):

@WG-prioritization :wave:

Santiago Pastorino (Apr 08 2020 at 18:00, on Zulip):

I've done some of the boiler plate stuff and created the agenda and all that

Santiago Pastorino (Apr 08 2020 at 18:00, on Zulip):

so let's go straight to the meat :)

Santiago Pastorino (Apr 08 2020 at 18:01, on Zulip):

Unprioritized I-nominated

Santiago Pastorino (Apr 08 2020 at 18:01, on Zulip):

Go over unprioritized I-nominated issues. Nominations are sometimes an implicit request for prioritization.

Santiago Pastorino (Apr 08 2020 at 18:01, on Zulip):
  1. No team assigned
    - Add T-compiler tag when it corresponds.
Santiago Pastorino (Apr 08 2020 at 18:01, on Zulip):

empty

Santiago Pastorino (Apr 08 2020 at 18:01, on Zulip):
  1. T-compiler
    - Prioritize issues and remove nomination of the ones not worth discussing.
    - Tag regressions accordingly.
    - Ping appropriate people and/or ICE-breakers.
Santiago Pastorino (Apr 08 2020 at 18:02, on Zulip):

there are 6 of those

Santiago Pastorino (Apr 08 2020 at 18:02, on Zulip):

unpri nom 1/6: Compile regression "cannot infer an appropriate lifetime for lifetime parameter" #70917

Santiago Pastorino (Apr 08 2020 at 18:02, on Zulip):

this is a regression from stable to nightly

Santiago Pastorino (Apr 08 2020 at 18:03, on Zulip):

seems like @eddyb nominated to T-lang?

Santiago Pastorino (Apr 08 2020 at 18:03, on Zulip):

reading ...

eddyb (Apr 08 2020 at 18:03, on Zulip):

yeah I fixed something that looks like a soundness hole

eddyb (Apr 08 2020 at 18:03, on Zulip):

accidentally

eddyb (Apr 08 2020 at 18:04, on Zulip):

(but likely isn't a soundness hole, we're just inconsistent)

Santiago Pastorino (Apr 08 2020 at 18:04, on Zulip):

@eddyb so the nomination is just for T-lang discussion?

eddyb (Apr 08 2020 at 18:05, on Zulip):

I guess? there's not much compiler-y about it

eddyb (Apr 08 2020 at 18:05, on Zulip):

other than the compiler isn't inconsistent now :P

Santiago Pastorino (Apr 08 2020 at 18:05, on Zulip):

:+1:

eddyb (Apr 08 2020 at 18:05, on Zulip):

but I believe the question of strict vs relaxed is a lang team one

Santiago Pastorino (Apr 08 2020 at 18:05, on Zulip):

unpri nom 2/6: ICE on generators + arrays #70905

Santiago Pastorino (Apr 08 2020 at 18:06, on Zulip):

another regression from stable to nightly

LeSeulArtichaut (Apr 08 2020 at 18:06, on Zulip):

This still needs bisection

Santiago Pastorino (Apr 08 2020 at 18:06, on Zulip):

yeah

Santiago Pastorino (Apr 08 2020 at 18:06, on Zulip):

going to ping cleanup

LeSeulArtichaut (Apr 08 2020 at 18:06, on Zulip):

I'm going to run it in the background :slight_smile:

eddyb (Apr 08 2020 at 18:07, on Zulip):

this looks familiar

Santiago Pastorino (Apr 08 2020 at 18:07, on Zulip):

I really want to change the labels here :)

Santiago Pastorino (Apr 08 2020 at 18:08, on Zulip):

I wonder if the nomination was with the purpose of let's discuss or let's prioritize

Santiago Pastorino (Apr 08 2020 at 18:08, on Zulip):

anyway, let's prioritize this one

eddyb (Apr 08 2020 at 18:08, on Zulip):

@LeSeulArtichaut try the PR I linked

pnkfelix (Apr 08 2020 at 18:08, on Zulip):

its from jonas-schievink ; they usually are doing "let's prioritize"

pnkfelix (Apr 08 2020 at 18:09, on Zulip):

(has that team/wg been told about the switch to I-prioritize for that purpose?)

eddyb (Apr 08 2020 at 18:09, on Zulip):

@LeSeulArtichaut I've just had a crater run fail because almost 10% of the tested crates depended on rand which triggered a similar but different ICE for unrelated reasons :P

Santiago Pastorino (Apr 08 2020 at 18:09, on Zulip):

pnkfelix said:

(has that team/wg been told about the switch to I-prioritize for that purpose?)

cc @DPC, I don't remember what we have decided

DPC (Apr 08 2020 at 18:09, on Zulip):

which team?

DPC (Apr 08 2020 at 18:09, on Zulip):

will inform them

Santiago Pastorino (Apr 08 2020 at 18:09, on Zulip):

I guess we should post in IRLO, post in zulip t-compiler a new topic

Santiago Pastorino (Apr 08 2020 at 18:09, on Zulip):

and all that

Santiago Pastorino (Apr 08 2020 at 18:10, on Zulip):

anyway, I think this issue is bad

Santiago Pastorino (Apr 08 2020 at 18:10, on Zulip):

tempted to do P-high

Santiago Pastorino (Apr 08 2020 at 18:10, on Zulip):

thoughts?

Santiago Pastorino (Apr 08 2020 at 18:11, on Zulip):

it's like a simple async example that is ICEing

pnkfelix (Apr 08 2020 at 18:11, on Zulip):

sure, P-high sounds fine to me

Santiago Pastorino (Apr 08 2020 at 18:12, on Zulip):

done

Santiago Pastorino (Apr 08 2020 at 18:13, on Zulip):

unpri nom 3/6: PhantomData<T> no longer dropck? #70841

Santiago Pastorino (Apr 08 2020 at 18:13, on Zulip):

this is a stable to stable regression

Wesley Wiser (Apr 08 2020 at 18:16, on Zulip):

This seems critical

Santiago Pastorino (Apr 08 2020 at 18:16, on Zulip):

yes sounds bad

Wesley Wiser (Apr 08 2020 at 18:16, on Zulip):

Stable-to-stable + soundness implications

Santiago Pastorino (Apr 08 2020 at 18:16, on Zulip):

I'm tempted to go with P-high too

Santiago Pastorino (Apr 08 2020 at 18:16, on Zulip):

on the other hand this issue has been for a while there

Wesley Wiser (Apr 08 2020 at 18:16, on Zulip):

Why not P-critical?

Santiago Pastorino (Apr 08 2020 at 18:17, on Zulip):

ahh right, we have the label created already

LeSeulArtichaut (Apr 08 2020 at 18:17, on Zulip):

We've got P-critical now :slight_smile:

Santiago Pastorino (Apr 08 2020 at 18:17, on Zulip):

well yeah, we would need to adapt our procedures to consider that

Wesley Wiser (Apr 08 2020 at 18:17, on Zulip):

I thought we've had that for a while? :confused:

Santiago Pastorino (Apr 08 2020 at 18:17, on Zulip):

but yes

LeSeulArtichaut (Apr 08 2020 at 18:17, on Zulip):

Since Monday I guess?

Wesley Wiser (Apr 08 2020 at 18:18, on Zulip):

My bad

Wesley Wiser (Apr 08 2020 at 18:18, on Zulip):

I thought we've had it far longer

Santiago Pastorino (Apr 08 2020 at 18:18, on Zulip):

I meant, this agenda will look for P-high issues so if we naively follow it we will be skipping all the P-critical checks

Santiago Pastorino (Apr 08 2020 at 18:18, on Zulip):

anyway, we can accomodate on the go

Santiago Pastorino (Apr 08 2020 at 18:18, on Zulip):

so P-critical then?

pnkfelix (Apr 08 2020 at 18:18, on Zulip):

is this actually P-critical? Is it a release blocker?

pnkfelix (Apr 08 2020 at 18:18, on Zulip):

I agree it sounds very bad

pnkfelix (Apr 08 2020 at 18:19, on Zulip):

but given that it is a stable-to-stable bug

LeSeulArtichaut (Apr 08 2020 at 18:19, on Zulip):

@Wesley Wiser https://rust-lang.zulipchat.com/#narrow/stream/227806-t-compiler.2Fwg-prioritization/topic/planning.20meeting.202020-04-06/near/193088404 :innocent:

pnkfelix (Apr 08 2020 at 18:19, on Zulip):

it doesn't make sense to me that it would be a release blocker

pnkfelix (Apr 08 2020 at 18:19, on Zulip):

(it perhaps would have been ...)

Santiago Pastorino (Apr 08 2020 at 18:19, on Zulip):

yeah I'm kind of not convinced

Santiago Pastorino (Apr 08 2020 at 18:19, on Zulip):

P-high definitely

Wesley Wiser (Apr 08 2020 at 18:19, on Zulip):

I guess you're saying that because it already shipped and it took this long for someone to notice?

pnkfelix (Apr 08 2020 at 18:19, on Zulip):

well, more that

pnkfelix (Apr 08 2020 at 18:19, on Zulip):

what good will it do to block a release on it?

Santiago Pastorino (Apr 08 2020 at 18:19, on Zulip):

but unsure if we should go P-critical

pnkfelix (Apr 08 2020 at 18:20, on Zulip):

I certainly agree it should be P-high. And the severity of this case, the fact that we really do want to fix it pronto

Wesley Wiser (Apr 08 2020 at 18:20, on Zulip):

I don't quite follow sorry

pnkfelix (Apr 08 2020 at 18:20, on Zulip):

@Wesley Wiser you don't follow my question?

Wesley Wiser (Apr 08 2020 at 18:20, on Zulip):

Yeah

pnkfelix (Apr 08 2020 at 18:20, on Zulip):

My interpretation of P-critical is that we use it for bugs so severe that we would block the release unless they are resolved

Wesley Wiser (Apr 08 2020 at 18:20, on Zulip):

Like what kind of issue would block a release if not a stable-to-stable regression with soundness implications?

pnkfelix (Apr 08 2020 at 18:21, on Zulip):

Right, but we already have the release out there

Wesley Wiser (Apr 08 2020 at 18:21, on Zulip):

I'm not trying to disagree, I just want to see your viewpoint

Wesley Wiser (Apr 08 2020 at 18:21, on Zulip):

Ah ok. So it's already in a regressed state

Santiago Pastorino (Apr 08 2020 at 18:21, on Zulip):

if this has been a stable to nightly or beta would have been P-critical

Wesley Wiser (Apr 08 2020 at 18:21, on Zulip):

So there's no point blocking the next release

pnkfelix (Apr 08 2020 at 18:22, on Zulip):

having said all that

Santiago Pastorino (Apr 08 2020 at 18:22, on Zulip):

:+1:

Santiago Pastorino (Apr 08 2020 at 18:22, on Zulip):

P-high

pnkfelix (Apr 08 2020 at 18:22, on Zulip):

I will admit that it does feel a bit weird

Santiago Pastorino (Apr 08 2020 at 18:22, on Zulip):

yeah

pnkfelix (Apr 08 2020 at 18:22, on Zulip):

having a time-dependent component to the priority assignment

pnkfelix (Apr 08 2020 at 18:22, on Zulip):

like, I can definitely see @Wesley Wiser 's point

Santiago Pastorino (Apr 08 2020 at 18:22, on Zulip):

we need another level P label, P-would-have-been-critical :P

pnkfelix (Apr 08 2020 at 18:22, on Zulip):

that since this bug would have been considered P-critical if we had caught it sooner

pnkfelix (Apr 08 2020 at 18:23, on Zulip):

we perhaps should still tag it as P-critical now

LeSeulArtichaut (Apr 08 2020 at 18:23, on Zulip):

This comes down to the meaning we give to the labels

pnkfelix (Apr 08 2020 at 18:23, on Zulip):

and given that we don't have an overload of P-critical bus right now

Wesley Wiser (Apr 08 2020 at 18:23, on Zulip):

I guess I'm coming around because we're already broken. Halting the current release doesn't prevent this bug from going out into the wild.

pnkfelix (Apr 08 2020 at 18:23, on Zulip):

I wouldn't object to using P-critical here

LeSeulArtichaut (Apr 08 2020 at 18:23, on Zulip):

If we tag it P-critical it will get more attention

Wesley Wiser (Apr 08 2020 at 18:23, on Zulip):

If it was the next release that was broken though, I definitely think it would be P-critical

pnkfelix (Apr 08 2020 at 18:23, on Zulip):

I just want to be clear that if someone is later asking "should we block the release on this" if it doesn't get fixed

Wesley Wiser (Apr 08 2020 at 18:23, on Zulip):

At least for this bug in particular.

pnkfelix (Apr 08 2020 at 18:23, on Zulip):

that I would then say "whoops we've had a process failure"

DPC (Apr 08 2020 at 18:24, on Zulip):

It's an important bug, but I don't think it is a release-blocker

Wesley Wiser (Apr 08 2020 at 18:24, on Zulip):

Yeah, I think I agree

Wesley Wiser (Apr 08 2020 at 18:24, on Zulip):

Sorry for dragging that out but I personally found it very helpful to talk that through @pnkfelix

pnkfelix (Apr 08 2020 at 18:24, on Zulip):

@DPC do you think it would have been a release blocker if we had caught it sooner?

pnkfelix (Apr 08 2020 at 18:24, on Zulip):

i.e. if this were a stable-to-beta regression?

Santiago Pastorino (Apr 08 2020 at 18:25, on Zulip):

Wesley Wiser said:

Sorry for dragging that out but I personally found it very helpful to talk that through pnkfelix

:+1:

DPC (Apr 08 2020 at 18:25, on Zulip):

probably yeah

pnkfelix (Apr 08 2020 at 18:25, on Zulip):

it is definitely tricky

pnkfelix (Apr 08 2020 at 18:25, on Zulip):

anyway

pnkfelix (Apr 08 2020 at 18:25, on Zulip):

what shall we do then?

pnkfelix (Apr 08 2020 at 18:25, on Zulip):

:)

Wesley Wiser (Apr 08 2020 at 18:25, on Zulip):

P-high :)

Santiago Pastorino (Apr 08 2020 at 18:25, on Zulip):

to be honest, as it has been discussed it could be any of those and depends on the definition of P-critical

Santiago Pastorino (Apr 08 2020 at 18:25, on Zulip):

Wesley Wiser said:

P-high :)

ok :+1:

pnkfelix (Apr 08 2020 at 18:25, on Zulip):

maybe this is even worth asking at tomorrow's meeting too

pnkfelix (Apr 08 2020 at 18:25, on Zulip):

get the opinion of the team about what the semantics of P-critical should be

Wesley Wiser (Apr 08 2020 at 18:26, on Zulip):

We should definitely highlight as it is a soundness issue.

Santiago Pastorino (Apr 08 2020 at 18:26, on Zulip):

we could tag with both labels :joy:

pnkfelix (Apr 08 2020 at 18:26, on Zulip):

right. I say: P-high, but also keep it I-nominated

pnkfelix (Apr 08 2020 at 18:26, on Zulip):

that way we will discuss it at the meeting tomorrow

LeSeulArtichaut (Apr 08 2020 at 18:26, on Zulip):

I think it depends on whether P-critical means "block the release" or "work on this before other P-high bugs"

Santiago Pastorino (Apr 08 2020 at 18:28, on Zulip):

LeSeulArtichaut said:

I think it depends on whether P-critical means "block the release" or "work on this before other P-high bugs"

definitely

Santiago Pastorino (Apr 08 2020 at 18:28, on Zulip):

anyway, done!

LeSeulArtichaut (Apr 08 2020 at 18:28, on Zulip):

@Santiago Pastorino Didn't you mean to keep I-nominated?

LeSeulArtichaut (Apr 08 2020 at 18:28, on Zulip):

:innocent:

Santiago Pastorino (Apr 08 2020 at 18:28, on Zulip):

yeah, I've removed and re-added it back

Santiago Pastorino (Apr 08 2020 at 18:29, on Zulip):

LeSeulArtichaut said:

I think it depends on whether P-critical means "block the release" or "work on this before other P-high bugs"

back to this for just a second

Santiago Pastorino (Apr 08 2020 at 18:29, on Zulip):

I guess P-critical could totally mean this is potentially a release blocker but then it can be discussed on a case by case basis

LeSeulArtichaut (Apr 08 2020 at 18:29, on Zulip):

Sorry to interrupt, but... Should we create a topic to discuss this so it can be accessed outside this meeting?

Santiago Pastorino (Apr 08 2020 at 18:30, on Zulip):

and by that definition this one could totally be P-critical, I'm not trying to re-discuss, we can just move to a different topic this discussion

Santiago Pastorino (Apr 08 2020 at 18:30, on Zulip):

LeSeulArtichaut said:

Sorry to interrupt, but... Should we create a topic to discuss this so it can be accessed outside this meeting?

yeah, exactly what I was going to say :)

Santiago Pastorino (Apr 08 2020 at 18:30, on Zulip):

let's move on

Santiago Pastorino (Apr 08 2020 at 18:30, on Zulip):

if someone can create the topic and move the discussion would be great

Santiago Pastorino (Apr 08 2020 at 18:30, on Zulip):

unpri nom 4/6: "-Funused" is overwritten by later "-Aunused" #70819

DPC (Apr 08 2020 at 18:31, on Zulip):

doing

Santiago Pastorino (Apr 08 2020 at 18:31, on Zulip):

stable to beta regression

Santiago Pastorino (Apr 08 2020 at 18:32, on Zulip):

pinged cleanup

Santiago Pastorino (Apr 08 2020 at 18:32, on Zulip):

anyway, I'd say this is P-medium

LeSeulArtichaut (Apr 08 2020 at 18:33, on Zulip):

This is not P-high nor P-low, so I think P-medium is fine

Santiago Pastorino (Apr 08 2020 at 18:33, on Zulip):

any other thought or argument about this?

Santiago Pastorino (Apr 08 2020 at 18:34, on Zulip):

ok, will tag as P-medium if someone disagrees we can go back to this one

Wesley Wiser (Apr 08 2020 at 18:34, on Zulip):

I'm not entirely sure this isn't expected behavior

pnkfelix (Apr 08 2020 at 18:34, on Zulip):

note that it is a regression

pnkfelix (Apr 08 2020 at 18:35, on Zulip):

so even if its not that important from our POV

pnkfelix (Apr 08 2020 at 18:35, on Zulip):

it still may be worth trying to address. Unless it is indeed a fix

Wesley Wiser (Apr 08 2020 at 18:36, on Zulip):

Ah ok. This is specifically about forbid lints not deny lints.

Wesley Wiser (Apr 08 2020 at 18:36, on Zulip):

P-medium seems good

pnkfelix (Apr 08 2020 at 18:36, on Zulip):

it also may need to be T-lang

Santiago Pastorino (Apr 08 2020 at 18:36, on Zulip):

pnkfelix said:

it still may be worth trying to address. Unless it is indeed a fix

what do you mean exactly? that it's not P-medium?

pnkfelix (Apr 08 2020 at 18:36, on Zulip):

given that Ralf has noted a similar problem with attributes in the source code itself

pnkfelix (Apr 08 2020 at 18:36, on Zulip):

that might make it a language issue

pnkfelix (Apr 08 2020 at 18:37, on Zulip):

Santiago Pastorino said:

what do you mean exactly? that it's not P-medium?

that is what I was wondering, yeah.

Santiago Pastorino (Apr 08 2020 at 18:38, on Zulip):

so, if we were discussing this fact as a generalized thing I'd say there are a lot of regressions that are P-medium

Santiago Pastorino (Apr 08 2020 at 18:38, on Zulip):

but still regressions

pnkfelix (Apr 08 2020 at 18:38, on Zulip):

true

Santiago Pastorino (Apr 08 2020 at 18:38, on Zulip):

if we were discussing this particular issue you may be right that's more important than P-medium

Wesley Wiser (Apr 08 2020 at 18:38, on Zulip):

At least it seems the command line is consistent now with source level attributes.

Santiago Pastorino (Apr 08 2020 at 18:38, on Zulip):

it's another interesting discussion, what a stable to beta P-medium regression means?

RalfJ (Apr 08 2020 at 18:39, on Zulip):

Wesley Wiser said:

Ah ok. This is specifically about forbid lints not deny lints.

indeed with deny this is expected behavior, but with forbid it is explicitly not -- that is the only reason forbid exists, as far as I understand.

LeSeulArtichaut (Apr 08 2020 at 18:39, on Zulip):

I don’t think it would be hard to fix when we get to the regressed commit

RalfJ (Apr 08 2020 at 18:39, on Zulip):

LeSeulArtichaut said:

I don’t think it would be hard to fix when we get to the regressed commit

I am not sure actually

Santiago Pastorino (Apr 08 2020 at 18:39, on Zulip):

do people want to go to P-high then?

RalfJ (Apr 08 2020 at 18:39, on Zulip):

the "regressed" commit made the CLI arg behavior consistent with the embedded attributes

pnkfelix (Apr 08 2020 at 18:40, on Zulip):

I think P-medium will be okay, as long as we also nominate it for T-lang discussion

RalfJ (Apr 08 2020 at 18:40, on Zulip):

oh you mean the attribute regressed commit

RalfJ (Apr 08 2020 at 18:40, on Zulip):

sure, that's just forever ago^^

pnkfelix (Apr 08 2020 at 18:40, on Zulip):

because I'm pretty sure the lang team will figure out what we actually want here

pnkfelix (Apr 08 2020 at 18:40, on Zulip):

(which I suspect will match the semantics RalfJ outlined)

Santiago Pastorino (Apr 08 2020 at 18:41, on Zulip):

ok, P-medium and nominating it again adding t-lang too

pnkfelix (Apr 08 2020 at 18:42, on Zulip):

(do make sure to note that your nominating for T-lang discussion. Or at least, I assume we don't want to force T-compiler to discuss tomorrow?)

Santiago Pastorino (Apr 08 2020 at 18:42, on Zulip):

yeah, have made that clear

Santiago Pastorino (Apr 08 2020 at 18:42, on Zulip):

unpri nom 5/6: "Aborted (core dumped)" on typo in attribute #70763

centril (Apr 08 2020 at 18:43, on Zulip):

My opinion on #70764: It's not clear this is a bug

Santiago Pastorino (Apr 08 2020 at 18:44, on Zulip):

centril said:

My opinion on #70764: It's not clear this is a bug

you meant #70763 ?

centril (Apr 08 2020 at 18:44, on Zulip):

yea that one :D

Santiago Pastorino (Apr 08 2020 at 18:44, on Zulip):

I was going to say that's unimportant in my opinion

Santiago Pastorino (Apr 08 2020 at 18:44, on Zulip):

without even considering if it's a bug or not

centril (Apr 08 2020 at 18:44, on Zulip):

It arises due to an unfortunate interaction with a process::abort due to a panic in the closure of atake_mut. The panic happens due to FatalError.raise()

Santiago Pastorino (Apr 08 2020 at 18:45, on Zulip):

I guess having the thing segfaulting makes it a bug?

centril (Apr 08 2020 at 18:45, on Zulip):

It's not a segfault I think

centril (Apr 08 2020 at 18:45, on Zulip):

Just a regular process::abort

Santiago Pastorino (Apr 08 2020 at 18:46, on Zulip):

Aborted (core dumped)

centril (Apr 08 2020 at 18:46, on Zulip):

Removing FatalError.raise() would introduce an ICE, removing the abort is not an option as it would be unsound

Santiago Pastorino (Apr 08 2020 at 18:46, on Zulip):

ohh I read that and thought it was

pnkfelix (Apr 08 2020 at 18:46, on Zulip):

FatalError.raise() causes aborts rather than ICE's ?

centril (Apr 08 2020 at 18:46, on Zulip):

not using visit_clobber would require a deeper rewrite of expansion to avoid &mut T; I believe @eddyb would be happy with that

centril (Apr 08 2020 at 18:47, on Zulip):

@pnkfelix it unwinds the thread which results in the compiler stopping eventually

Santiago Pastorino (Apr 08 2020 at 18:47, on Zulip):

meanwhile we figure out if this is a bug or not, should we go with P-low?

centril (Apr 08 2020 at 18:47, on Zulip):

except in this case, unwinding ends up in visit_clobber's catch_unwind, which results in process::abort();

eddyb (Apr 08 2020 at 18:47, on Zulip):

oh dear

eddyb (Apr 08 2020 at 18:47, on Zulip):

I wasn't sure what visit_clobber was, but I can tell by the failure mode

centril (Apr 08 2020 at 18:48, on Zulip):

@eddyb its take_mut with a different name

eddyb (Apr 08 2020 at 18:48, on Zulip):

(yes, that's what I mean, only take_mut fails in that specific way)

pnkfelix (Apr 08 2020 at 18:48, on Zulip):

okay i see the comment thread now talks about visit_clobber

centril (Apr 08 2020 at 18:48, on Zulip):

@Santiago Pastorino Yes, P-low sgtm; -- I tried fixing this by avoiding self.expect(...) in the list parser, but it regressed diagnostics imo

centril (Apr 08 2020 at 18:49, on Zulip):

so I think I'm dropping that

eddyb (Apr 08 2020 at 18:49, on Zulip):

/me is not a big fan of MutVisitor and learning that it was happening/it happened ruined some plans for advancing macro expansion integration with incremental and/or solving the problem where expanded ASTs require pretty-printing to be fed to proc macros

eddyb (Apr 08 2020 at 18:50, on Zulip):

but that's all in the past and it was a temporary perf improvement

eddyb (Apr 08 2020 at 18:50, on Zulip):

(temporary because we can't keep it forever :P)

DPC (Apr 08 2020 at 18:50, on Zulip):

maybe we could discuss that issue separately? :P

Santiago Pastorino (Apr 08 2020 at 18:50, on Zulip):

added P-low and removed nomination

eddyb (Apr 08 2020 at 18:50, on Zulip):

@DPC ofc just wanted to point it out (but it is offtopic)

Santiago Pastorino (Apr 08 2020 at 18:51, on Zulip):

DPC said:

maybe we could discuss that issue separately? :P

I'm very happy that this kind of discussions are happening to be very honest :)

Santiago Pastorino (Apr 08 2020 at 18:51, on Zulip):

and feel free to keep discussing :heart:

Santiago Pastorino (Apr 08 2020 at 18:51, on Zulip):

also remember that's very cheap to open a new topic :slight_smile:

DPC (Apr 08 2020 at 18:51, on Zulip):

i understand. just time-checking :slight_smile:

Santiago Pastorino (Apr 08 2020 at 18:51, on Zulip):

unpri nom 6/6: resolving bounds after type-checking: predict trait for std::ops::FnMut #70746

centril (Apr 08 2020 at 18:52, on Zulip):

error: internal compiler error: src/librustc_trait_selection/traits/codegen/mod.rs:107: Encountered errors [FulfillmentError(Obligation(predicate=Binder(TraitPredicate(<dyn Callback<T1, Output = ()> as std::ops::FnMut<(usize,)>>)), depth=1),Unimplemented)] resolving bounds after type-checking

centril (Apr 08 2020 at 18:52, on Zulip):

Add it to the list of duplicates that are not getting fixed anytime soon :slight_smile:

Santiago Pastorino (Apr 08 2020 at 18:52, on Zulip):

hehe

Santiago Pastorino (Apr 08 2020 at 18:52, on Zulip):

say more

Santiago Pastorino (Apr 08 2020 at 18:52, on Zulip):

I'm just starting to read it :)

centril (Apr 08 2020 at 18:53, on Zulip):

Assuming it is a dupe of that one @pnkfelix ?

Wesley Wiser (Apr 08 2020 at 18:53, on Zulip):

Seems to have regressed in 1.32

pnkfelix (Apr 08 2020 at 18:53, on Zulip):

there are a slew of FulfillmentError ICE's

Wesley Wiser (Apr 08 2020 at 18:53, on Zulip):

(but was an error previously)

centril (Apr 08 2020 at 18:54, on Zulip):

Seems similar to https://github.com/rust-lang/rust/issues/69892

centril (Apr 08 2020 at 18:54, on Zulip):

(which also was not prioritized)

Santiago Pastorino (Apr 08 2020 at 18:55, on Zulip):

yeah

pnkfelix (Apr 08 2020 at 18:55, on Zulip):

though #69892 needs an MCVE and uses a #![feature]

pnkfelix (Apr 08 2020 at 18:55, on Zulip):

#70746 uses no features and is pretty minimal already.

Santiago Pastorino (Apr 08 2020 at 18:56, on Zulip):

true

pnkfelix (Apr 08 2020 at 18:56, on Zulip):

anyway I'd be willing to call this P-medium

pnkfelix (Apr 08 2020 at 18:56, on Zulip):

it would be nice to figure out which other bug to tag it as a duplicate of

Santiago Pastorino (Apr 08 2020 at 18:57, on Zulip):

going with P-medium then

Santiago Pastorino (Apr 08 2020 at 18:57, on Zulip):

commented all that

pnkfelix (Apr 08 2020 at 18:57, on Zulip):

does it look anything like #27675 ?

pnkfelix (Apr 08 2020 at 18:58, on Zulip):

(omg look at how low that issue number is... its from 2015 ...)

Santiago Pastorino (Apr 08 2020 at 18:58, on Zulip):

yeah

Santiago Pastorino (Apr 08 2020 at 18:58, on Zulip):

that's similar and very old

Santiago Pastorino (Apr 08 2020 at 18:59, on Zulip):

should we ping any ICE breaking group for this search?

Santiago Pastorino (Apr 08 2020 at 18:59, on Zulip):

to look for duplicates I meant

pnkfelix (Apr 08 2020 at 18:59, on Zulip):

hmm

pnkfelix (Apr 08 2020 at 18:59, on Zulip):

we don't really have any groups dedicated to that

pnkfelix (Apr 08 2020 at 19:00, on Zulip):

talk about a thankless pursuit... :sad:

Santiago Pastorino (Apr 08 2020 at 19:00, on Zulip):

isn't cleanup crew also for that?

pnkfelix (Apr 08 2020 at 19:00, on Zulip):

are they? maybe

DPC (Apr 08 2020 at 19:00, on Zulip):

could be considered part of triage as well :P

pnkfelix (Apr 08 2020 at 19:00, on Zulip):

at this point, if we don't already have an umbrella issue dedicated to this class of ICE's, we probably should

Santiago Pastorino (Apr 08 2020 at 19:00, on Zulip):

pnkfelix said:

are they? maybe

a minimal, standalone example that shows the problem
links to duplicates or related bugs
if the bug is a regression (something that used to work, but no longer does), then a bisection to the PR or nightly that caused the regression
Santiago Pastorino (Apr 08 2020 at 19:01, on Zulip):

links to duplicates or related bugs

pnkfelix (Apr 08 2020 at 19:01, on Zulip):

Fair enough

Santiago Pastorino (Apr 08 2020 at 19:01, on Zulip):

going to ping cleanup crew for this

Santiago Pastorino (Apr 08 2020 at 19:01, on Zulip):

ok, let's move on

Santiago Pastorino (Apr 08 2020 at 19:01, on Zulip):

we are 1h in and we still are in the first step :S

Santiago Pastorino (Apr 08 2020 at 19:02, on Zulip):
  1. All teams
    - No action required. It's nice to look at this for comparison.
Santiago Pastorino (Apr 08 2020 at 19:02, on Zulip):

this is for comparison purposes but let's move on this time because this will take ages

Santiago Pastorino (Apr 08 2020 at 19:02, on Zulip):
  1. Unnominate leftover I-nominated
    - Remove I-nominated tag from issues discussed on the last meeting.
Santiago Pastorino (Apr 08 2020 at 19:03, on Zulip):

we have 8 nominations

Santiago Pastorino (Apr 08 2020 at 19:03, on Zulip):

let's see what we should remove

Santiago Pastorino (Apr 08 2020 at 19:03, on Zulip):

“Box<dyn FnOnce> doesn’t respect self alignment” #68304
“internal compiler error: cannot relate region: LUB(ReErased, ReErased)” #70608
“regression: assertion failed: data.is_empty()” #70445
“internal compiler error: no type for local variable” #70594
“Should enum discriminants have generics in scope?” #70453

Santiago Pastorino (Apr 08 2020 at 19:03, on Zulip):

that was last meeting nominations

centril (Apr 08 2020 at 19:04, on Zulip):

#70608 is closed

pnkfelix (Apr 08 2020 at 19:04, on Zulip):

there has been progress on the comment thread for #68304

Santiago Pastorino (Apr 08 2020 at 19:05, on Zulip):

we can remove nomination from #68304

pnkfelix (Apr 08 2020 at 19:05, on Zulip):

so we could either leave it I-nominated (with the intention of discussing its status with the team) or remove the nomination (under the assumption that its under control)

Santiago Pastorino (Apr 08 2020 at 19:05, on Zulip):

I'm going to work in that one and we were discussing with @eddyb about it

pnkfelix (Apr 08 2020 at 19:05, on Zulip):

I am indeed inclined to remove the nomination

Santiago Pastorino (Apr 08 2020 at 19:05, on Zulip):

yeah

centril (Apr 08 2020 at 19:06, on Zulip):

sounds reasonable; although I'm a bit concerned that if we don't keep it nominated, it will slide out of control

Santiago Pastorino (Apr 08 2020 at 19:07, on Zulip):

#70453 was also discussed in last meeting, right?

pnkfelix (Apr 08 2020 at 19:07, on Zulip):

yeah, and it is more directed at T-lang, I think?

centril (Apr 08 2020 at 19:07, on Zulip):

I think there's an emerging T-Lang consensus on that one

centril (Apr 08 2020 at 19:08, on Zulip):

well; from the ones who have commented

Santiago Pastorino (Apr 08 2020 at 19:08, on Zulip):

cc @eddyb does #70453 still needs nomination?

centril (Apr 08 2020 at 19:08, on Zulip):

and eddyb has a PR

centril (Apr 08 2020 at 19:08, on Zulip):

so I think nomination can be removed

pnkfelix (Apr 08 2020 at 19:08, on Zulip):

lets leave it nominated but for T-lang. that is, we won't put it on T-compiler agenda

Santiago Pastorino (Apr 08 2020 at 19:08, on Zulip):

yeah I guess it doesn't need nomination

pnkfelix (Apr 08 2020 at 19:08, on Zulip):

okay, we can also remove nomination entirely

centril (Apr 08 2020 at 19:08, on Zulip):

let's do what @pnkfelix said

centril (Apr 08 2020 at 19:08, on Zulip):

better to discuss at T-Lang

eddyb (Apr 08 2020 at 19:08, on Zulip):

if you remove the nomination, @nikomatsakis' approval of option 2 is implicitly accepted :P

Santiago Pastorino (Apr 08 2020 at 19:09, on Zulip):

ok, left as is and commented

centril (Apr 08 2020 at 19:09, on Zulip):

@eddyb you know other people also approved, right? :slight_smile:

Santiago Pastorino (Apr 08 2020 at 19:09, on Zulip):

let's keep it going

Santiago Pastorino (Apr 08 2020 at 19:09, on Zulip):

Regressions

Santiago Pastorino (Apr 08 2020 at 19:09, on Zulip):
  1. Beta regressions without P-label
    - Prioritize.
    - Ping appropriate people and/or ICE-breakers.
    - Assign if possible; if it remains unassigned, add it to agenda so we can assign during the meeting.
Santiago Pastorino (Apr 08 2020 at 19:09, on Zulip):

empty

Santiago Pastorino (Apr 08 2020 at 19:10, on Zulip):
  1. Nightly regressions without P-label
    - Prioritize.
    - Ping appropriate people and/or ICE-breakers.
    - Assign if possible; if it remains unassigned, add it to agenda so we can assign during the meeting.
Santiago Pastorino (Apr 08 2020 at 19:10, on Zulip):

there's one

Santiago Pastorino (Apr 08 2020 at 19:10, on Zulip):

beta regression 1/1: Compile regression "cannot infer an appropriate lifetime for lifetime parameter" #70917

eddyb (Apr 08 2020 at 19:10, on Zulip):

@centril sorry, I forgot that, so 29% of the lang team :P

centril (Apr 08 2020 at 19:11, on Zulip):

beta regression 1/1: Compile regression "cannot infer an appropriate lifetime for lifetime parameter" #70917

(it's a nightly regression)

Santiago Pastorino (Apr 08 2020 at 19:11, on Zulip):

yeah we should prioritize and assign (if possible)

eddyb (Apr 08 2020 at 19:12, on Zulip):

blocked on lang team decision IMO

centril (Apr 08 2020 at 19:12, on Zulip):

I believe the main question is: "did @eddyb introduce a fix or a bug?"

centril (Apr 08 2020 at 19:12, on Zulip):

If they did a fix, I will issue a point release to proptest; although other crates could have regressed

DPC (Apr 08 2020 at 19:13, on Zulip):

maybe a crater run?

Santiago Pastorino (Apr 08 2020 at 19:13, on Zulip):

I wonder if we should prioritize this anyway or what

pnkfelix (Apr 08 2020 at 19:13, on Zulip):

I agree with @eddyb that this seems like T-lang's call

pnkfelix (Apr 08 2020 at 19:13, on Zulip):

lets prioritize as P-high

Santiago Pastorino (Apr 08 2020 at 19:13, on Zulip):

would be like prioritizing a discussion :)

Santiago Pastorino (Apr 08 2020 at 19:13, on Zulip):

yeah, P-high

centril (Apr 08 2020 at 19:13, on Zulip):

I agree with everyone :slight_smile:

eddyb (Apr 08 2020 at 19:13, on Zulip):

if someone pings me slightly ahead of the lang team meeting tomorrow I can try to attend

pnkfelix (Apr 08 2020 at 19:14, on Zulip):

@eddyb like 23 hours ahead? <ping>

centril (Apr 08 2020 at 19:14, on Zulip):

@eddyb le sigh, but also, I can do that

eddyb (Apr 08 2020 at 19:14, on Zulip):

well for starters, I don't think I know how to/can join anymore :P

Santiago Pastorino (Apr 08 2020 at 19:14, on Zulip):

pnkfelix said:

eddyb like 23 hours ahead? <ping>

LOL

pnkfelix (Apr 08 2020 at 19:14, on Zulip):

we'll get it sorted out

Santiago Pastorino (Apr 08 2020 at 19:15, on Zulip):

commented and added p-high

Santiago Pastorino (Apr 08 2020 at 19:15, on Zulip):

Beta nominations

Santiago Pastorino (Apr 08 2020 at 19:15, on Zulip):
  1. No team assigned
    - Add T-compiler tag when it corresponds.
centril (Apr 08 2020 at 19:15, on Zulip):

@eddyb 1. You have Zoom installed? Yes? ==> 2. Use https://mozilla.zoom.us/j/547421668, No? Install Zoom and jump to 1.

Santiago Pastorino (Apr 08 2020 at 19:15, on Zulip):

empty

Santiago Pastorino (Apr 08 2020 at 19:15, on Zulip):
  1. All teams
    - Add T-compiler tag when it corresponds.
eddyb (Apr 08 2020 at 19:15, on Zulip):

ah I wasn't sure if the link was private, thanks @centril :D

Santiago Pastorino (Apr 08 2020 at 19:16, on Zulip):

nothing to do on 2

Santiago Pastorino (Apr 08 2020 at 19:16, on Zulip):
  1. T-compiler
    - Add these issues to the meeting agenda.
pnkfelix (Apr 08 2020 at 19:16, on Zulip):

whelp if it was, its not anymore.

centril (Apr 08 2020 at 19:16, on Zulip):

nah; anyone can join t-lang meetings

pnkfelix (Apr 08 2020 at 19:16, on Zulip):

(that was a joke; I knew/know its public)

Santiago Pastorino (Apr 08 2020 at 19:16, on Zulip):

Santiago Pastorino said:

  1. T-compiler
    - Add these issues to the meeting agenda.

there's one beta accepted and one that's not, adding it to the agenda

Santiago Pastorino (Apr 08 2020 at 19:18, on Zulip):

done

Santiago Pastorino (Apr 08 2020 at 19:18, on Zulip):

Stable nominations

Santiago Pastorino (Apr 08 2020 at 19:18, on Zulip):
  1. No team assigned
    - Add T-compiler tag when it corresponds.
Santiago Pastorino (Apr 08 2020 at 19:18, on Zulip):

empty

Santiago Pastorino (Apr 08 2020 at 19:18, on Zulip):
  1. All teams
    - Add T-compiler tag when it corresponds.
Santiago Pastorino (Apr 08 2020 at 19:18, on Zulip):

there's one but already stable-accepted

Santiago Pastorino (Apr 08 2020 at 19:19, on Zulip):
  1. T-compiler
    - Add these issues to the meeting agenda.
Santiago Pastorino (Apr 08 2020 at 19:19, on Zulip):

well already covered :)

Santiago Pastorino (Apr 08 2020 at 19:19, on Zulip):

PR's waiting for our team

Santiago Pastorino (Apr 08 2020 at 19:19, on Zulip):
Santiago Pastorino (Apr 08 2020 at 19:19, on Zulip):

there are 7!!!

Santiago Pastorino (Apr 08 2020 at 19:19, on Zulip):

:scream:

Santiago Pastorino (Apr 08 2020 at 19:20, on Zulip):

going to add those to the meeting agenda

Santiago Pastorino (Apr 08 2020 at 19:20, on Zulip):

but we should nominate the ones we think could be fixed quickly

Santiago Pastorino (Apr 08 2020 at 19:24, on Zulip):

added all this PRs to the agenda

Wesley Wiser (Apr 08 2020 at 19:24, on Zulip):

#60900 seems like it's more waiting on a review or maybe the author to fix CI?

Santiago Pastorino (Apr 08 2020 at 19:24, on Zulip):

any worth discusseing?

Wesley Wiser (Apr 08 2020 at 19:24, on Zulip):

I'm not sure so I don't want to adjust the labels.

centril (Apr 08 2020 at 19:24, on Zulip):

hmm; I'm not sure why it's waiting on team

centril (Apr 08 2020 at 19:24, on Zulip):

seems like its waiting on author and reviewer

DPC (Apr 08 2020 at 19:25, on Zulip):

@Wesley Wiser yes it should be waiting on author

Santiago Pastorino (Apr 08 2020 at 19:25, on Zulip):

@DPC can you fix the labels?

centril (Apr 08 2020 at 19:25, on Zulip):

https://github.com/rust-lang/rust/pull/70680 seems like it should be i-nominated for t-compiler btw

DPC (Apr 08 2020 at 19:25, on Zulip):

was already on it :D

Santiago Pastorino (Apr 08 2020 at 19:26, on Zulip):

centril said:

https://github.com/rust-lang/rust/pull/70680 seems like it should be i-nominated for t-compiler btw

let's do it :)

centril (Apr 08 2020 at 19:26, on Zulip):

#70680 probably needs some discussion re. the approach, whether we want to do this, stability (and t-lang might be involved there), etc.

Santiago Pastorino (Apr 08 2020 at 19:26, on Zulip):

@centril can you?

centril (Apr 08 2020 at 19:26, on Zulip):

and @eddyb had concerns

centril (Apr 08 2020 at 19:26, on Zulip):

done

Santiago Pastorino (Apr 08 2020 at 19:26, on Zulip):

on both?

centril (Apr 08 2020 at 19:26, on Zulip):

both?

eddyb (Apr 08 2020 at 19:27, on Zulip):

I probably want to stay out of it, as long as the addition is not too big

eddyb (Apr 08 2020 at 19:27, on Zulip):

/me shrugs

Santiago Pastorino (Apr 08 2020 at 19:27, on Zulip):

centril said:

both?

sorry got confused

centril (Apr 08 2020 at 19:27, on Zulip):

@eddyb I personally feel uncomfortable by you staying out of this tbh

eddyb (Apr 08 2020 at 19:27, on Zulip):

I have better things to do than block toy features

centril (Apr 08 2020 at 19:27, on Zulip):

like we should be hearing about your concerns

centril (Apr 08 2020 at 19:27, on Zulip):

fair point though

Santiago Pastorino (Apr 08 2020 at 19:28, on Zulip):

let's continue then

Santiago Pastorino (Apr 08 2020 at 19:28, on Zulip):

High priority issues

Santiago Pastorino (Apr 08 2020 at 19:28, on Zulip):
Santiago Pastorino (Apr 08 2020 at 19:31, on Zulip):

added all that to the agenda

Santiago Pastorino (Apr 08 2020 at 19:31, on Zulip):

there are 3 regressions that are for T-rustdoc and have no priorities

Santiago Pastorino (Apr 08 2020 at 19:31, on Zulip):

unsure what should we do with those

Santiago Pastorino (Apr 08 2020 at 19:32, on Zulip):

cc @pnkfelix

Santiago Pastorino (Apr 08 2020 at 19:32, on Zulip):

moving on ...

Santiago Pastorino (Apr 08 2020 at 19:32, on Zulip):

Stable to beta regressions

Santiago Pastorino (Apr 08 2020 at 19:32, on Zulip):
centril (Apr 08 2020 at 19:32, on Zulip):

The inbetween is a bit confusing; perhaps t-compiler should take over rustdoc entirely or something, or not at all

Santiago Pastorino (Apr 08 2020 at 19:33, on Zulip):

centril said:

The inbetween is a bit confusing; perhaps t-compiler should take over rustdoc entirely or something, or not at all

yeah unsure

Santiago Pastorino (Apr 08 2020 at 19:33, on Zulip):

Santiago Pastorino said:

do we want to raise awareness of any of those

Santiago Pastorino (Apr 08 2020 at 19:33, on Zulip):

I'd say if any the only P-high we have

Santiago Pastorino (Apr 08 2020 at 19:34, on Zulip):

that one has a PR I think, checking ....

Santiago Pastorino (Apr 08 2020 at 19:35, on Zulip):

btw, was talking about this one #70445

Santiago Pastorino (Apr 08 2020 at 19:35, on Zulip):

but it was also beta-accepted

Santiago Pastorino (Apr 08 2020 at 19:35, on Zulip):

so I guess there's nothing to raise awareness

Santiago Pastorino (Apr 08 2020 at 19:36, on Zulip):

if somebody finds something let me know

Santiago Pastorino (Apr 08 2020 at 19:36, on Zulip):

let's move on

Santiago Pastorino (Apr 08 2020 at 19:36, on Zulip):

Stable to nightly regressions

Santiago Pastorino (Apr 08 2020 at 19:36, on Zulip):
Santiago Pastorino (Apr 08 2020 at 19:37, on Zulip):

most important ones are moving forward

Santiago Pastorino (Apr 08 2020 at 19:37, on Zulip):

Stable to stable regressions

Santiago Pastorino (Apr 08 2020 at 19:37, on Zulip):
Santiago Pastorino (Apr 08 2020 at 19:38, on Zulip):

71 stable to stable regressions

Santiago Pastorino (Apr 08 2020 at 19:38, on Zulip):

just 27 prioritized

Santiago Pastorino (Apr 08 2020 at 19:38, on Zulip):

:S

Santiago Pastorino (Apr 08 2020 at 19:39, on Zulip):

I think at some point we should have a dedicated meeting to prioritize stable to stable regressions

Santiago Pastorino (Apr 08 2020 at 19:39, on Zulip):

I-nominated T-compiler

Santiago Pastorino (Apr 08 2020 at 19:39, on Zulip):
Santiago Pastorino (Apr 08 2020 at 19:40, on Zulip):

there are 8 but some are for t-lang

Santiago Pastorino (Apr 08 2020 at 19:40, on Zulip):

going to figure this out

Santiago Pastorino (Apr 08 2020 at 19:48, on Zulip):

ok done

Santiago Pastorino (Apr 08 2020 at 19:48, on Zulip):

Toolstate

Santiago Pastorino (Apr 08 2020 at 19:49, on Zulip):

Check toolstate for outstanding tool breakage.

Santiago Pastorino (Apr 08 2020 at 19:49, on Zulip):

clippy-driver and miri are failing

Santiago Pastorino (Apr 08 2020 at 19:49, on Zulip):

anyway, we have time for the release still

Santiago Pastorino (Apr 08 2020 at 19:49, on Zulip):

Performance regressions

Santiago Pastorino (Apr 08 2020 at 19:50, on Zulip):

Check perf regressions.

DPC (Apr 08 2020 at 19:50, on Zulip):

miri & clippy are in the queue and will be updated by tomorrow end

Santiago Pastorino (Apr 08 2020 at 19:51, on Zulip):

image.png

Santiago Pastorino (Apr 08 2020 at 19:51, on Zulip):

wondering what happened there, it was "fixed" but not to the exact previous perf

Santiago Pastorino (Apr 08 2020 at 19:51, on Zulip):

cc @centril

centril (Apr 08 2020 at 19:52, on Zulip):

@eddyb ^---

Santiago Pastorino (Apr 08 2020 at 19:52, on Zulip):

starting here https://perf.rust-lang.org/compare.html?start=9e55101bb681010c82c3c827305e2665fc8f2aa0&end=6050e523bae6de61de4e060facc43dc512adaccd&stat=instructions:u

centril (Apr 08 2020 at 19:52, on Zulip):

the hash related regressions were not entirely fixed?

Santiago Pastorino (Apr 08 2020 at 19:52, on Zulip):

something very bad happened

eddyb (Apr 08 2020 at 19:53, on Zulip):

it was explained though?

eddyb (Apr 08 2020 at 19:53, on Zulip):

I saw a comment go by

eddyb (Apr 08 2020 at 19:53, on Zulip):

there is one point just before the actual regression

simulacrum (Apr 08 2020 at 19:53, on Zulip):

clap sets codegen-units to 4

eddyb (Apr 08 2020 at 19:53, on Zulip):

which raises the graph to the point it is after the regression is fixed

simulacrum (Apr 08 2020 at 19:53, on Zulip):
[profile.dev]
opt-level = 0
debug = true
rpath = false
lto = false
debug-assertions = true
codegen-units = 4
DPC (Apr 08 2020 at 19:54, on Zulip):

where is that from?

simulacrum (Apr 08 2020 at 19:54, on Zulip):

so the incremental patch regressed it, because it started respecting that

eddyb (Apr 08 2020 at 19:54, on Zulip):

image.png

simulacrum (Apr 08 2020 at 19:54, on Zulip):

https://github.com/rust-lang/rustc-perf/blob/master/collector/benchmarks/clap-rs/Cargo.toml#L60

eddyb (Apr 08 2020 at 19:54, on Zulip):

(note the circled triangle)

centril (Apr 08 2020 at 19:54, on Zulip):

Okay so it seems everything here is under control and expected

simulacrum (Apr 08 2020 at 19:55, on Zulip):

and indeed this is https://github.com/rust-lang/rust/commit/9e55101bb681010c82c3c827305e2665fc8f2aa0 that eddyb points out

simulacrum (Apr 08 2020 at 19:55, on Zulip):

which was expected by mw, and deemed acceptable (I agree)

eddyb (Apr 08 2020 at 19:55, on Zulip):

https://github.com/rust-lang/rust/pull/70803#issuecomment-610749611

eddyb (Apr 08 2020 at 19:55, on Zulip):

this is what I should've linked but I didn't have it on hand

centril (Apr 08 2020 at 19:55, on Zulip):

Good good; as long as everything is expected and everyone is up to speed

eddyb (Apr 08 2020 at 19:55, on Zulip):

(i.e. this discussion happened already :P)

eddyb (Apr 08 2020 at 19:56, on Zulip):

(but the people involved aren't here right now AFAIK)

centril (Apr 08 2020 at 19:56, on Zulip):

@eddyb yea we're just dispersing the knowledge wider :slight_smile:

eddyb (Apr 08 2020 at 19:57, on Zulip):

maybe I should have a firefox contextual menu thing that starts KolourPaint on an image :P

eddyb (Apr 08 2020 at 19:57, on Zulip):

sometimes it's much easier to edit something to point at a detail but it's effort

eddyb (Apr 08 2020 at 19:58, on Zulip):

also note the inclusive/exclusive comparisons I do https://github.com/rust-lang/rust/issues/70785#issuecomment-609438109

eddyb (Apr 08 2020 at 19:59, on Zulip):

[regression, fix] and (regression, fix)

eddyb (Apr 08 2020 at 19:59, on Zulip):

just to catch stuff like this where the fix may not be complete (it is here)

Santiago Pastorino (Apr 08 2020 at 20:05, on Zulip):

ok, just in case this is over :) :wave:

Santiago Pastorino (Apr 08 2020 at 20:05, on Zulip):

going to tweak a bit the agenda

DPC (Apr 08 2020 at 20:09, on Zulip):

:wave:

Santiago Pastorino (Apr 08 2020 at 21:13, on Zulip):

BTW, a new issue showed up #70934

Santiago Pastorino (Apr 08 2020 at 21:13, on Zulip):

we should prioritize this one

LeSeulArtichaut (Apr 08 2020 at 21:16, on Zulip):

(deleted)

LeSeulArtichaut (Apr 08 2020 at 21:17, on Zulip):

It should be labeled regression-from-stable-to-stable too

Santiago Pastorino (Apr 08 2020 at 21:24, on Zulip):

right, added the label

LeSeulArtichaut (Apr 08 2020 at 21:43, on Zulip):

Thanks

eddyb (Apr 09 2020 at 12:24, on Zulip):

@Santiago Pastorino if I just nominated #70953, will it make it into the meeting?

LeSeulArtichaut (Apr 09 2020 at 12:27, on Zulip):

I guess it should be added to the agenda?

eddyb (Apr 09 2020 at 12:27, on Zulip):

yeah idk how the new procedures work

LeSeulArtichaut (Apr 09 2020 at 12:28, on Zulip):

Pre-triage procedure for I-nominated is:

(From the procedure document)

eddyb (Apr 09 2020 at 12:32, on Zulip):

yeah but where's the agenda :)

eddyb (Apr 09 2020 at 12:32, on Zulip):

oh I should check the real meeting topic lmao

LeSeulArtichaut (Apr 09 2020 at 12:32, on Zulip):

It's here

LeSeulArtichaut (Apr 09 2020 at 12:33, on Zulip):

Yeah it's in the meeting channel

Santiago Pastorino (Apr 09 2020 at 12:33, on Zulip):

@eddyb yeah, let's add to the agenda

Santiago Pastorino (Apr 09 2020 at 12:34, on Zulip):

btw, if you feel like, we can also prioritize about the discussion items

Santiago Pastorino (Apr 09 2020 at 12:34, on Zulip):

because the agenda feels packed and it may be good if we start with the most important stuff

eddyb (Apr 09 2020 at 12:34, on Zulip):

I added it at the end :P

eddyb (Apr 09 2020 at 12:35, on Zulip):

lemme reorder the items a bit

eddyb (Apr 09 2020 at 12:35, on Zulip):

regressions should come first IMO

eddyb (Apr 09 2020 at 12:36, on Zulip):

actually two of the nominated things are PRs that can wait for a while

eddyb (Apr 09 2020 at 12:37, on Zulip):

@Santiago Pastorino lmk if you take issue with my changes

Santiago Pastorino (Apr 09 2020 at 12:42, on Zulip):

I was going to check the agenda again

Last update: Jun 05 2020 at 23:00UTC