Stream: t-compiler

Topic: pre-meeting triage 2019-08-15 #54818


nikomatsakis (Aug 15 2019 at 13:06, on Zulip):

I will be doing pre-triage in this channel.

nikomatsakis (Aug 15 2019 at 13:12, on Zulip):

The instructions on #54818 are a bit confusing but I'll start with the unprioritized nominated issues

nikomatsakis (Aug 15 2019 at 13:12, on Zulip):

miri no longer builds after rust-lang/rust#63575 #63581

nikomatsakis (Aug 15 2019 at 13:13, on Zulip):

Any update what's up with that, @centril @RalfJ or @oli ? :point_up:

nikomatsakis (Aug 15 2019 at 13:13, on Zulip):

er the "breaking" PR was a rollup, so I guess @centril is off the hook :)

centril (Aug 15 2019 at 13:13, on Zulip):

(I get cced on these because I make rollups and I really don't have time to figure out why they break... ^^)

nikomatsakis (Aug 15 2019 at 13:13, on Zulip):

looks like it's already assigned, I'm going to mark it as p-high; miri should build

centril (Aug 15 2019 at 13:14, on Zulip):

@nikomatsakis I just assume Ralf is already on that ;)

nikomatsakis (Aug 15 2019 at 13:14, on Zulip):

a fair assumption

nikomatsakis (Aug 15 2019 at 13:14, on Zulip):

run-make-fulldeps/linker-output-non-utf8 failure #63520

nikomatsakis (Aug 15 2019 at 13:15, on Zulip):

so I filed this, it's a bug I see locally, I'm not sure how to prioritize it, I'm going to say P-medium. I'm presuming it's something to do with my setup

centril (Aug 15 2019 at 13:15, on Zulip):

I could make an argument for P-high: Having this bug open means that all other bugs take more time to fix

centril (Aug 15 2019 at 13:15, on Zulip):

because it affects you ^^

nikomatsakis (Aug 15 2019 at 13:16, on Zulip):

heh, well, it comes relatively late in the test suite...

nikomatsakis (Aug 15 2019 at 13:16, on Zulip):

I'll ping a few folks

nikomatsakis (Aug 15 2019 at 13:16, on Zulip):

maybe run apt-get update :P

nikomatsakis (Aug 15 2019 at 13:16, on Zulip):

ICE when trying to reference an associated trait constant #63496

centril (Aug 15 2019 at 13:17, on Zulip):

Standard lazy norm issue it seems

nikomatsakis (Aug 15 2019 at 13:17, on Zulip):

yes, I was thinking something similar. going to mark as p-medium.

centril (Aug 15 2019 at 13:18, on Zulip):

This probably shouldn't be fixed (i.e. check-pass, ofc it should not ICE...) without signoff from T-lang or at the very least some tests re. associated constant default values.

centril (Aug 15 2019 at 13:18, on Zulip):

(similar to associated type defaults)

nikomatsakis (Aug 15 2019 at 13:18, on Zulip):

Regression: Matching on function pointers fails when one of the formal params implements custom PartialEq #63479

nikomatsakis (Aug 15 2019 at 13:20, on Zulip):

Hmm.

centril (Aug 15 2019 at 13:20, on Zulip):

This should not compile?

nikomatsakis (Aug 15 2019 at 13:20, on Zulip):

Not 100% clear to me. The constant is a fn() pointer, so it doesn't really matter what arguments it takes

nikomatsakis (Aug 15 2019 at 13:20, on Zulip):

but then the rules aren't 100% settled so it's hard to say

centril (Aug 15 2019 at 13:21, on Zulip):

I prefer to be conservative when that's the case -- much easier to become more liberal if you settle starting from a conservative position.

nikomatsakis (Aug 15 2019 at 13:21, on Zulip):

I'd like to assign to @pnkfelix but of course they're not going to be back until early Sep

nikomatsakis (Aug 15 2019 at 13:23, on Zulip):

ok I assigned to @pnkfelix but I'm going to leave it nominated for another week to see what the aswer is

nikomatsakis (Aug 15 2019 at 13:23, on Zulip):

segfault with target-feature=+soft-float #63466

centril (Aug 15 2019 at 13:24, on Zulip):

(Overall, I think we need a more holistic thinking re. structural match and think about const generics / dependent typing/ definitional equality / ... etc. etc.)

nikomatsakis (Aug 15 2019 at 13:25, on Zulip):

welp, that seems ungreat.

nikomatsakis (Aug 15 2019 at 13:25, on Zulip):

I have no idea what soft-float is

nikomatsakis (Aug 15 2019 at 13:25, on Zulip):

(presumably floating point emulation?)

nikomatsakis (Aug 15 2019 at 13:25, on Zulip):

maybe I'll leave this one for the meeting to cal attention to it and see if anybody has any thoughts

centril (Aug 15 2019 at 13:26, on Zulip):

probably yeah;
(I add I-unsound when I see "segfault")

centril (Aug 15 2019 at 13:26, on Zulip):

P-high then?

nikomatsakis (Aug 15 2019 at 13:27, on Zulip):

guess so

nikomatsakis (Aug 15 2019 at 13:27, on Zulip):

Trait bounds are not checked on type aliases until they are used #51626

nikomatsakis (Aug 15 2019 at 13:28, on Zulip):

This was nominated by @Esteban K├╝ber but it's a rather old and long standing problem

nikomatsakis (Aug 15 2019 at 13:28, on Zulip):

I'm inclined to mark it as p-medium, afaik there's been no new development?

centril (Aug 15 2019 at 13:28, on Zulip):

yeah, standard "not checking WF"

centril (Aug 15 2019 at 13:29, on Zulip):

actually this is in defaults...

centril (Aug 15 2019 at 13:29, on Zulip):

this is possibly not a bug?

nikomatsakis (Aug 15 2019 at 13:29, on Zulip):

Well, it all sort of depends I guess :)

nikomatsakis (Aug 15 2019 at 13:29, on Zulip):

the default meets the bounds on the type alias

nikomatsakis (Aug 15 2019 at 13:29, on Zulip):

(which are none)

nikomatsakis (Aug 15 2019 at 13:30, on Zulip):

it doesn't meet the bounds from the value

centril (Aug 15 2019 at 13:30, on Zulip):

right, we discussed defaults and bounds not too long ago

nikomatsakis (Aug 15 2019 at 13:30, on Zulip):

don't we have some other issue (or ten) sort of tracking this? I'd be inclined to cc that and add this example to keep it in mind

nikomatsakis (Aug 15 2019 at 13:31, on Zulip):

Beta regressions: None

nikomatsakis (Aug 15 2019 at 13:31, on Zulip):

Nightly regressions:

centril (Aug 15 2019 at 13:31, on Zulip):

@nikomatsakis possibly https://github.com/rust-lang-nursery/reference/issues/636 ?

nikomatsakis (Aug 15 2019 at 13:31, on Zulip):

well, sort of, but that's not the one I was thinking of

nikomatsakis (Aug 15 2019 at 13:32, on Zulip):

https://github.com/rust-lang/rust/issues/55222

nikomatsakis (Aug 15 2019 at 13:32, on Zulip):

maybe that, though the name is...not very good:)

nikomatsakis (Aug 15 2019 at 13:33, on Zulip):

(but that's the comments I was thinking of)

nikomatsakis (Aug 15 2019 at 13:33, on Zulip):

we could maybe repurpose that as a tracking issue

nikomatsakis (Aug 15 2019 at 13:33, on Zulip):

Regression: Matching on function pointers fails when one of the formal params implements custom PartialEq #63479

nikomatsakis (Aug 15 2019 at 13:33, on Zulip):

ah, we covered this one already

nikomatsakis (Aug 15 2019 at 13:33, on Zulip):

OK we have 39 P-high issues :scared:

nikomatsakis (Aug 15 2019 at 13:34, on Zulip):

oh which 17 are unassigned

nikomatsakis (Aug 15 2019 at 13:34, on Zulip):

I guess the question is whether to choose some to highlight

nikomatsakis (Aug 15 2019 at 13:34, on Zulip):

let's start by skimming for updates maybe..

nikomatsakis (Aug 15 2019 at 13:36, on Zulip):

tracking issue for lifetime inference error work (E0495) #42516

nikomatsakis (Aug 15 2019 at 13:36, on Zulip):

I feel like this could be downgraded -- or maybe even closed -- after the NLL work.

nikomatsakis (Aug 15 2019 at 13:38, on Zulip):

Let's start by looking at I-unsound P-high issues

nikomatsakis (Aug 15 2019 at 13:40, on Zulip):

Weird compiler behavior with in-lining of a function -> constant value arbitrary changed #63516

nikomatsakis (Aug 15 2019 at 13:40, on Zulip):

this seems like an LLVM bug

nikomatsakis (Aug 15 2019 at 13:41, on Zulip):

also perhaps specific to arm or something like that?

nikomatsakis (Aug 15 2019 at 13:45, on Zulip):

Coherence can be bypassed by an indirect impl for a trait object #57893

nikomatsakis (Aug 15 2019 at 13:45, on Zulip):

no updates here since our lang team discussion

nikomatsakis (Aug 15 2019 at 13:45, on Zulip):

I have to bring this back in cache, I don't recall what the next actionable steps were going to be

centril (Aug 15 2019 at 13:46, on Zulip):

trying out some good solutions and trying crater maybe?

centril (Aug 15 2019 at 13:46, on Zulip):

and then modelling in chalk possibly

nikomatsakis (Aug 15 2019 at 13:46, on Zulip):

Yeah, probably.

nikomatsakis (Aug 15 2019 at 13:46, on Zulip):

I'm looking over nominated issues presently, trying to decide if some are worth calling out at meeting

centril (Aug 15 2019 at 13:46, on Zulip):

seems like something for wg-traits in general

nikomatsakis (Aug 15 2019 at 13:47, on Zulip):

Use dangling for default slices (like vec and box do) #62487

nikomatsakis (Aug 15 2019 at 13:47, on Zulip):

seems like this one -- we are debating about doing a more general fix -- @eddyb or @oli, would one of you care to be assigned?

eddyb (Aug 15 2019 at 13:48, on Zulip):

to help with, or actually go ahead and do it :P?

nikomatsakis (Aug 15 2019 at 13:48, on Zulip):

really either :)

nikomatsakis (Aug 15 2019 at 13:48, on Zulip):

I doubt @scottmcm would object

eddyb (Aug 15 2019 at 13:48, on Zulip):

idk how much time oli has, feel free to assign me or w/e

oli (Aug 15 2019 at 13:50, on Zulip):

I can review and help with figuring out things

oli (Aug 15 2019 at 13:50, on Zulip):

it shouldn't be too much work I think

oli (Aug 15 2019 at 13:50, on Zulip):

or yea, just grab the PR and do the changes

nikomatsakis (Aug 15 2019 at 13:50, on Zulip):

ICE: Generic type alias to invalid type crashes during type check on latest stable #62742

I'm going to assign this to myself -- but I need to formulate a plan to either delegate or solve some of these ICEs and things assigned to me.

centril (Aug 15 2019 at 13:51, on Zulip):

Would be good to minimize that if possible

scottmcm (Aug 15 2019 at 18:52, on Zulip):

I'm happy either way -- I can try to figure it out and change the PR, or if it's trivial for someone with context they can do it and I'll just close my PR.

RalfJ (Aug 16 2019 at 06:25, on Zulip):

nikomatsakis I just assume Ralf is already on that ;)

it was already fixed in Miri when you were speaking ;) the Miri PR was sitting there patiently for 2 weeks waiting for its rustc brother to land.

Last update: Nov 22 2019 at 04:30UTC