Stream: t-compiler

Topic: weekly meeting 2020-03-05 #54818


pnkfelix (Mar 04 2020 at 17:24, on Zulip):

Hi @T-compiler/meeting ; the triage meeting will be starting in 21 hours 37 minutes (i.e. its tomorrow). (I'm just starting the triage earlier than usual this week)

pnkfelix (Mar 04 2020 at 17:25, on Zulip):

I will be doing pre-triage in a parallel topic

pnkfelix (Mar 04 2020 at 17:26, on Zulip):

We'll be having a WG checkin from WG-meta tomorrow; @Santiago Pastorino , I assume you will be available to provide a checkin update?

pnkfelix (Mar 04 2020 at 17:27, on Zulip):

Likewise there is a scheduled checkin from WG-mir-opt; @oli do you think you will be around at the end of the meeting tomorrow?

Santiago Pastorino (Mar 04 2020 at 17:27, on Zulip):

I can do that, unsure if @nikomatsakis wants to mention something about it but I can check with him

pnkfelix (Mar 04 2020 at 18:14, on Zulip):

also: I've started filling out the agenda for this meeting: https://hackmd.io/Iv_CLF_iTVKtvspFS4ohkw?both

oli (Mar 05 2020 at 09:22, on Zulip):

I'll be around. I'll try to prep some notes in time

pnkfelix (Mar 05 2020 at 14:19, on Zulip):

(just another heads up to @T-compiler/meeting , since I started pre-triage so early yesterday; the meeting starts in 41 minutes.)

pnkfelix (Mar 05 2020 at 15:03, on Zulip):

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

pnkfelix (Mar 05 2020 at 15:04, on Zulip):

we'll start off with five minutes for ...

pnkfelix (Mar 05 2020 at 15:04, on Zulip):

Announcements

pnkfelix (Mar 05 2020 at 15:04, on Zulip):
pnkfelix (Mar 05 2020 at 15:04, on Zulip):
simulacrum (Mar 05 2020 at 15:05, on Zulip):

I just approved a iterative (improvement, hopefully) to our tidy checks (https://github.com/rust-lang/rust/pull/69688). I suspect it'll be a few more attempts before we find a happy balance here, so I'd appreciate feedback in a few weeks as to how people are feeling (e.g., tidy is too slow still)

pnkfelix (Mar 05 2020 at 15:08, on Zulip):

thanks @simulacrum !

pnkfelix (Mar 05 2020 at 15:08, on Zulip):

okay we have a lot of beta-noms today, so lets jump in

pnkfelix (Mar 05 2020 at 15:09, on Zulip):

(the agenda is here, by the way: https://hackmd.io/Iv_CLF_iTVKtvspFS4ohkw?both#Beta-nominations )

pnkfelix (Mar 05 2020 at 15:09, on Zulip):

beta-nom: "Fix a leak in DiagnosticBuilder::into_diagnostic." #69628

nikomatsakis (Mar 05 2020 at 15:11, on Zulip):

seems like a win

pnkfelix (Mar 05 2020 at 15:11, on Zulip):

heh. mem::forget for the loss

nikomatsakis (Mar 05 2020 at 15:12, on Zulip):

(pun intended)

pnkfelix (Mar 05 2020 at 15:12, on Zulip):

anyway, yeah, seems fine

pnkfelix (Mar 05 2020 at 15:12, on Zulip):

beta-accepted

pnkfelix (Mar 05 2020 at 15:12, on Zulip):

although ... why aren't we just letting this ride the trains, to be honest?

pnkfelix (Mar 05 2020 at 15:13, on Zulip):

ah never mind

pnkfelix (Mar 05 2020 at 15:13, on Zulip):

beta-accepted

pnkfelix (Mar 05 2020 at 15:13, on Zulip):

(the time for me to doubt has past)

pnkfelix (Mar 05 2020 at 15:13, on Zulip):

beta-nom: "stash API: remove panic to fix ICE." #69623

pnkfelix (Mar 05 2020 at 15:15, on Zulip):

what is a stashed diagnostic?

nikomatsakis (Mar 05 2020 at 15:16, on Zulip):

Stashes diagnostic for possible later improvement in a different, later stage of the compiler. The diagnostic can be accessed with the provided span and key through .steal_diagnostic on Handler.

pnkfelix (Mar 05 2020 at 15:16, on Zulip):

and also, here "See the PR for discussion": which PR? PR #69623 ? there isn't much discussion there...

nikomatsakis (Mar 05 2020 at 15:16, on Zulip):

I think it's used when we generate the error in one place but can learn more later..? can't remember the exact case

nikomatsakis (Mar 05 2020 at 15:16, on Zulip):

I feel like this was added by @Esteban Küber I was wrong :)

simulacrum (Mar 05 2020 at 15:16, on Zulip):

yes

pnkfelix (Mar 05 2020 at 15:17, on Zulip):

okay.

simulacrum (Mar 05 2020 at 15:17, on Zulip):

discussion in https://github.com/rust-lang/rust/pull/69537#issuecomment-593143975

pnkfelix (Mar 05 2020 at 15:18, on Zulip):

okay. I wont block this on nits like "that comment isn't providing enough info for one to construct a link"

Esteban Küber (Mar 05 2020 at 15:18, on Zulip):

I think stash was added by centril sometime back

pnkfelix (Mar 05 2020 at 15:18, on Zulip):

just feedback

pnkfelix (Mar 05 2020 at 15:18, on Zulip):

beta-accepted

pnkfelix (Mar 05 2020 at 15:18, on Zulip):

(oh I see, the PR is referenced in the parentehical of the FIXME)

pnkfelix (Mar 05 2020 at 15:19, on Zulip):

okay then

pnkfelix (Mar 05 2020 at 15:19, on Zulip):

beta-nom: "Do not ICE on invalid type node after parse recovery" #69583

pnkfelix (Mar 05 2020 at 15:21, on Zulip):

okay looks good; beta-accepted.

pnkfelix (Mar 05 2020 at 15:22, on Zulip):

beta-nom: "error_derive_forbidden_on_non_adt: be more graceful" #69522

pnkfelix (Mar 05 2020 at 15:23, on Zulip):

looks good; beta-accepted

pnkfelix (Mar 05 2020 at 15:24, on Zulip):

beta-nom: "instantiate_value_path: on SelfCtor, avoid unconstrained tyvars" #69340

pnkfelix (Mar 05 2020 at 15:24, on Zulip):

(we touched n this last week but explicitly delayed decision to this week)

nikomatsakis (Mar 05 2020 at 15:25, on Zulip):

There was one reported regression so far: https://github.com/rust-lang/rust/issues/69611

pnkfelix (Mar 05 2020 at 15:26, on Zulip):

niko spelled out their thoughts on the PR itself

pnkfelix (Mar 05 2020 at 15:26, on Zulip):

do you interpret regression reports as incentive to backport? Or not to backport? :)

simulacrum (Mar 05 2020 at 15:26, on Zulip):

looks like Centril believes the regression to be good? i.e., expected

Wesley Wiser (Mar 05 2020 at 15:27, on Zulip):

https://github.com/rust-lang/rust/issues/69306#issuecomment-589141546 says this was introduced in 1.32. It doesn't seem particularly urgent to me to back port given how long this bug has existed.

pnkfelix (Mar 05 2020 at 15:27, on Zulip):

yes, as in "this code should never have compiled"

nikomatsakis (Mar 05 2020 at 15:27, on Zulip):

simulacrum said:

looks like Centril believes the regression to be good? i.e., expected

right, it's a bug fix

nikomatsakis (Mar 05 2020 at 15:27, on Zulip):

pnkfelix said:

do you interpret regression reports as incentive to backport? Or not to backport? :)

this is the question, isn't it =) I tend to think "not" -- give people more time to adapt

pnkfelix (Mar 05 2020 at 15:27, on Zulip):

the main argument for a backport, as I understand it, is to stop the pattern from spreading further

nikomatsakis (Mar 05 2020 at 15:27, on Zulip):

(but yes, that)

nikomatsakis (Mar 05 2020 at 15:28, on Zulip):

this btw is a good test case for the triage guidelines (cc @Santiago Pastorino)

pnkfelix (Mar 05 2020 at 15:28, on Zulip):

but I think I'm basically in agreement with niko.

pnkfelix (Mar 05 2020 at 15:28, on Zulip):

lets let this ride the trains

simulacrum (Mar 05 2020 at 15:28, on Zulip):

one note is that we're a week out from a release

nikomatsakis (Mar 05 2020 at 15:28, on Zulip):

I think @Wesley Wiser's data point (how long it's been in the wild) is relevant too

simulacrum (Mar 05 2020 at 15:28, on Zulip):

so this would basically be almost a direct-to-stable

Esteban Küber (Mar 05 2020 at 15:28, on Zulip):

Could we backport a warning?

nikomatsakis (Mar 05 2020 at 15:28, on Zulip):

it'd be hard

pnkfelix (Mar 05 2020 at 15:28, on Zulip):

Esteban Küber said:

Could we backport a warning?

no, its not really feasible to just warn on this

nikomatsakis (Mar 05 2020 at 15:28, on Zulip):

and not worth it, I think, since failure to build on nightly is a kind of warning on itself

pnkfelix (Mar 05 2020 at 15:28, on Zulip):

otherwise I would have pushed for that path

pnkfelix (Mar 05 2020 at 15:29, on Zulip):

(as in, I would have pushed for this to be yet another future-compat lint)

nikomatsakis (Mar 05 2020 at 15:29, on Zulip):

agreed, that's the ideal path, but given the difficulty, not worth it

nikomatsakis (Mar 05 2020 at 15:29, on Zulip):

I vote for "ride the trains"

pnkfelix (Mar 05 2020 at 15:30, on Zulip):

simulacrum said:

so this would basically be almost a direct-to-stable

yeah, this will be in beta soon enough just by riding the trains

pnkfelix (Mar 05 2020 at 15:30, on Zulip):

okay, sounds great

pnkfelix (Mar 05 2020 at 15:31, on Zulip):

beta-nom: "Update compiler-builtins to 0.1.25" #69086

pnkfelix (Mar 05 2020 at 15:31, on Zulip):

and, I guess PR #69168 ... (someone else added that bullet to agenda item I think)

pnkfelix (Mar 05 2020 at 15:31, on Zulip):

(i.e the two are coupled)

simulacrum (Mar 05 2020 at 15:32, on Zulip):

yeah I added it

simulacrum (Mar 05 2020 at 15:32, on Zulip):

one is a change and the other's the test for that change

pnkfelix (Mar 05 2020 at 15:32, on Zulip):

but ... we don't typically require basckporting of tests ..?

pnkfelix (Mar 05 2020 at 15:32, on Zulip):

I don't object to it though. I assume we run the test suite as part of the release process?

simulacrum (Mar 05 2020 at 15:33, on Zulip):

Yes, as with any merge to any branch, tests are run.

pnkfelix (Mar 05 2020 at 15:33, on Zulip):

gotcha

simulacrum (Mar 05 2020 at 15:33, on Zulip):

We usually approve backport of tests implicitly, just this time it was in a separate PR

pnkfelix (Mar 05 2020 at 15:34, on Zulip):

the main effect of this change, as I understand it, is from this PR "Try setting __rust_probestack visibility back to hidden" #69045, right?

pnkfelix (Mar 05 2020 at 15:34, on Zulip):

sorry, the point there is the title of that PR

simulacrum (Mar 05 2020 at 15:34, on Zulip):

correct, fixing a regression

simulacrum (Mar 05 2020 at 15:34, on Zulip):

namely https://github.com/rust-lang/rust/issues/68794

pnkfelix (Mar 05 2020 at 15:34, on Zulip):

I'm trying to understand/summarize the effect of the upgrade in question, as to assess the risks

pnkfelix (Mar 05 2020 at 15:35, on Zulip):

yeah I guess I meant that I wanted to know if there are other changes to compiler-builtins that are being coupled here

simulacrum (Mar 05 2020 at 15:35, on Zulip):

afaik, this is basically a revert to a previous state

simulacrum (Mar 05 2020 at 15:35, on Zulip):

hm, I can check, but we can also approve only if there aren't

pnkfelix (Mar 05 2020 at 15:36, on Zulip):

I'm skimming now

nikomatsakis (Mar 05 2020 at 15:36, on Zulip):

I did say :back: but I admit I was wondering same thing :)

pnkfelix (Mar 05 2020 at 15:36, on Zulip):

heh. the logs state where bump to 0.1.23 happened, and the bump to 0.1.25

simulacrum (Mar 05 2020 at 15:36, on Zulip):

looks like no

pnkfelix (Mar 05 2020 at 15:37, on Zulip):

but I don't see 0.1.24 in my skim of the commits

simulacrum (Mar 05 2020 at 15:37, on Zulip):

beta already has 0.1.24

pnkfelix (Mar 05 2020 at 15:37, on Zulip):

right that's my point

pnkfelix (Mar 05 2020 at 15:37, on Zulip):

I'm trying to see, from https://github.com/rust-lang/compiler-builtins/commits/master

pnkfelix (Mar 05 2020 at 15:37, on Zulip):

what is actually the delta from 0.1.24

pnkfelix (Mar 05 2020 at 15:38, on Zulip):

okay 0.1.24 was here: https://github.com/rust-lang/compiler-builtins/commit/0a15c9bdfe97b76f61fdc06626cecc0573aa8269

simulacrum (Mar 05 2020 at 15:38, on Zulip):

https://github.com/rust-lang/compiler-builtins/compare/0.1.24..0.1.25

simulacrum (Mar 05 2020 at 15:38, on Zulip):

just probestack

pnkfelix (Mar 05 2020 at 15:38, on Zulip):

aka "Don't modify the intrinsic abi for aarch64 windows (#337) "

pnkfelix (Mar 05 2020 at 15:38, on Zulip):

simulacrum said:

https://github.com/rust-lang/compiler-builtins/compare/0.1.24..0.1.25

ah good call. :)

pnkfelix (Mar 05 2020 at 15:38, on Zulip):

great great, okay. beta-accepted whole-heartedly then

pnkfelix (Mar 05 2020 at 15:39, on Zulip):

(and obviously, if there are any problems with that test, since run-make tests can be finicky, then just don't include it.)

simulacrum (Mar 05 2020 at 15:40, on Zulip):

:thumbs_up:

pnkfelix (Mar 05 2020 at 15:40, on Zulip):

okay, that's all the beta-noms

pnkfelix (Mar 05 2020 at 15:40, on Zulip):

luckily we have no stable-noms

pnkfelix (Mar 05 2020 at 15:40, on Zulip):

nor any S-waiting-on-team

pnkfelix (Mar 05 2020 at 15:40, on Zulip):

there are four nominated issues I was hoping to discuss

pnkfelix (Mar 05 2020 at 15:41, on Zulip):

but largely to raise awareness

pnkfelix (Mar 05 2020 at 15:41, on Zulip):

I-nom: “can not compile in release mode in raspberry pi zero” #69420

pnkfelix (Mar 05 2020 at 15:41, on Zulip):

ARM is tier 2

pnkfelix (Mar 05 2020 at 15:41, on Zulip):

but this seems really bad

pnkfelix (Mar 05 2020 at 15:41, on Zulip):

like, compiling Hello World causes rustc to SIGSEGV

pnkfelix (Mar 05 2020 at 15:42, on Zulip):

so I categorized it as P-high.

simulacrum (Mar 05 2020 at 15:42, on Zulip):

https://github.com/rust-lang/rust/issues/69420#issuecomment-593654973 suggests LLVM is the root

pnkfelix (Mar 05 2020 at 15:42, on Zulip):

If you object (as in, "this is P-medium at best", I welcome feedback on the bug itself.

pnkfelix (Mar 05 2020 at 15:42, on Zulip):

and the only other question I would have is whether this would be P-critical under our new scheme.

pnkfelix (Mar 05 2020 at 15:43, on Zulip):

I.e. would we block a release on this

nikomatsakis (Mar 05 2020 at 15:43, on Zulip):

I'm not entirely sure what P-high means :)

pnkfelix (Mar 05 2020 at 15:43, on Zulip):

and my suspicion is: No, we would not block a reelase

simulacrum (Mar 05 2020 at 15:43, on Zulip):

as tier 2 it can't be critical I think

nikomatsakis (Mar 05 2020 at 15:43, on Zulip):

But I don't think I feel this is critical

simulacrum (Mar 05 2020 at 15:43, on Zulip):

if it was critical then it would basically say that this is tier 1

pnkfelix (Mar 05 2020 at 15:43, on Zulip):

this is always the question, isn't it.

nikomatsakis (Mar 05 2020 at 15:43, on Zulip):

Maybe good to dig into why a bit -- tier 2 is definitely big part of that judgement

pnkfelix (Mar 05 2020 at 15:43, on Zulip):

simulacrum said:

if it was critical then it would basically say that this is tier 1

okay that is in line with my expectations

nikomatsakis (Mar 05 2020 at 15:44, on Zulip):

something this makes me think though

pnkfelix (Mar 05 2020 at 15:44, on Zulip):

as for what P-high means to me, in a case like this, is: Should we be checkin on this every week

nikomatsakis (Mar 05 2020 at 15:44, on Zulip):

I saw you pinged LLVM ICE-breakers

nikomatsakis (Mar 05 2020 at 15:44, on Zulip):

we've often wondered about "less well supported" tiers

nikomatsakis (Mar 05 2020 at 15:44, on Zulip):

maybe we should have groups we can ping for them too

nikomatsakis (Mar 05 2020 at 15:44, on Zulip):

that folks can add themselves too

pnkfelix (Mar 05 2020 at 15:44, on Zulip):

hmm, ARM-breakers

simulacrum (Mar 05 2020 at 15:44, on Zulip):

that would be amazing

nikomatsakis (Mar 05 2020 at 15:44, on Zulip):

eg., it'd be great to be able to ping "arm folks"

pnkfelix (Mar 05 2020 at 15:45, on Zulip):

"nice compiler you have here. shame if something happened to it, like a SIGSEGV"

pnkfelix (Mar 05 2020 at 15:45, on Zulip):

okay anyway that's all I wanted for that ticket I think

pnkfelix (Mar 05 2020 at 15:46, on Zulip):

I-nominated: “Box<dyn FnOnce> doesn’t respect self alignment” #68304

pnkfelix (Mar 05 2020 at 15:46, on Zulip):

oh I think this has seen activity since I put it on the list yesterday

pnkfelix (Mar 05 2020 at 15:47, on Zulip):

yeah I think this is now getting attention, thank to @eddyb and @RalfJ

pnkfelix (Mar 05 2020 at 15:47, on Zulip):

so we don't need to discuss here

pnkfelix (Mar 05 2020 at 15:47, on Zulip):

I-nominated: “Use wmemchr for searching NULL in [u16] on Windows” #67705

pnkfelix (Mar 05 2020 at 15:47, on Zulip):

this is a Windows bug. Clearly we also need a Window-breakers

nikomatsakis (Mar 05 2020 at 15:48, on Zulip):

certainly, though I think windows is also tier 1

eddyb (Mar 05 2020 at 15:48, on Zulip):

anarchy!

pnkfelix (Mar 05 2020 at 15:48, on Zulip):

but seriously, this is seeking a Windows dev to benchmark as part of evaluation

nikomatsakis (Mar 05 2020 at 15:48, on Zulip):

this is a PR, though

pnkfelix (Mar 05 2020 at 15:49, on Zulip):

it is also adding a pub fn to std::sys::windows

eddyb (Mar 05 2020 at 15:49, on Zulip):

isn't std::sys private?

pnkfelix (Mar 05 2020 at 15:49, on Zulip):

oh is it?

pnkfelix (Mar 05 2020 at 15:50, on Zulip):

yes, yes it is private

pnkfelix (Mar 05 2020 at 15:50, on Zulip):

okay never mind that then

simulacrum (Mar 05 2020 at 15:50, on Zulip):

Should be I think

Wesley Wiser (Mar 05 2020 at 15:50, on Zulip):

I have a Windows box

pnkfelix (Mar 05 2020 at 15:50, on Zulip):

Okay we just found our volunteer

eddyb (Mar 05 2020 at 15:50, on Zulip):

we should transition to crate visibility inside std::sys

pnkfelix (Mar 05 2020 at 15:50, on Zulip):

thanks @Wesley Wiser !

Wesley Wiser (Mar 05 2020 at 15:50, on Zulip):

Is this just a matter of running a script to test the performance or do I need to come up with the benchmarks myself?

eddyb (Mar 05 2020 at 15:51, on Zulip):

before opening the PR I was going to bring up @BurntSushi, seems he was assigned the PR last year but hasn't commented on it, anyone know if he's busy or on vacation etc.?

pnkfelix (Mar 05 2020 at 15:52, on Zulip):

Wesley Wiser said:

Is this just a matter of running a script to test the performance or do I need to come up with the benchmarks myself?

what, is there some problem with running JS code via rustc? ;)

pnkfelix (Mar 05 2020 at 15:53, on Zulip):

its a good point though: should we be spending time on this at all?

pnkfelix (Mar 05 2020 at 15:53, on Zulip):

the fact that the developer proposing it does not even have windows machine makes me say ... .What is going on?

Wesley Wiser (Mar 05 2020 at 15:54, on Zulip):

If it's just a matter of making a build and running cargo bench or something, I can likely do it today or tomorrow. If I have to come up with the benchmark myself, this probably won't get done for a while honestly...

pnkfelix (Mar 05 2020 at 15:54, on Zulip):

I'm sort of switching my mind here to thinking we may want to just close. On the one hand, it is a shame to throw away potentially good work.

pnkfelix (Mar 05 2020 at 15:54, on Zulip):

but we have to choose where we invest effort

pnkfelix (Mar 05 2020 at 15:55, on Zulip):

@Wesley Wiser i say: give it a whirl in cargo bench

eddyb (Mar 05 2020 at 15:55, on Zulip):

opened #69737 btw for the visibility thing

pnkfelix (Mar 05 2020 at 15:55, on Zulip):

don't spend time trying to make your own benchmakr

pnkfelix (Mar 05 2020 at 15:56, on Zulip):

if this change regresses, or even if it just doesn't show any improvement

pnkfelix (Mar 05 2020 at 15:56, on Zulip):

then we throw it away

pnkfelix (Mar 05 2020 at 15:56, on Zulip):

Sound good?

eddyb (Mar 05 2020 at 15:56, on Zulip):

wow this is 6 years old https://bugzilla.mozilla.org/show_bug.cgi?id=1003726#c5 (nevermind, I see, that's C++ code and we might have a similar situation in Rust code but a different usecase)

pnkfelix (Mar 05 2020 at 15:56, on Zulip):

(that is: I don' want to throw it away without at least doing some basic evaluation)

pnkfelix (Mar 05 2020 at 15:56, on Zulip):

okay, next

pnkfelix (Mar 05 2020 at 15:56, on Zulip):

I-nominated: “Consider handling signals to catch hard crashes of rustc to provide nice ICE message” #16238

pnkfelix (Mar 05 2020 at 15:56, on Zulip):

I nominated this because I want to close it

pnkfelix (Mar 05 2020 at 15:57, on Zulip):

I think we don't have much incentive to catch signals ourselves

eddyb (Mar 05 2020 at 15:57, on Zulip):

Ctrl+C would be great

eddyb (Mar 05 2020 at 15:57, on Zulip):

(but probably orthogonal to anything like crash reporting)

pnkfelix (Mar 05 2020 at 15:57, on Zulip):

its a complex enough topic, and one that rust doesn't support well today, at least for SIGSEGV (to my knowledge)

pnkfelix (Mar 05 2020 at 15:58, on Zulip):

so, that's all. I wanted to draw attention to it, now I have

eddyb (Mar 05 2020 at 15:58, on Zulip):

also, IIRC debuggers pause the program before the signal handler runs

pnkfelix (Mar 05 2020 at 15:58, on Zulip):

if any of you want to fight to keep it open, leave a comment on the ticket.

eddyb (Mar 05 2020 at 15:58, on Zulip):

(and you can customize it too, on top of that. or if I'm wrong, you can get the desired behavior)

pnkfelix (Mar 05 2020 at 15:58, on Zulip):

eddyb said:

also, IIRC debuggers pause the program before the signal handler runs

yeah, or maybe its a setting you can control from the debugger, that sounds right.

eddyb (Mar 05 2020 at 15:59, on Zulip):

I think the only relevant crash here, that isn't inside LLVM itself, would be a stack overflow. we could try to make those nicer (if we don't already)

pnkfelix (Mar 05 2020 at 16:00, on Zulip):

okay that's all of the nominations I put on the agenda.

pnkfelix (Mar 05 2020 at 16:00, on Zulip):

and unfortunately that's our hour

pnkfelix (Mar 05 2020 at 16:00, on Zulip):

I have a hard out right now. @Santiago Pastorino and @oli , please do post a WG- checkin

pnkfelix (Mar 05 2020 at 16:00, on Zulip):

but adios to @T-compiler/meeting from me!

Santiago Pastorino (Mar 05 2020 at 16:01, on Zulip):

WG-Meta

Santiago Pastorino (Mar 05 2020 at 16:01, on Zulip):

So remember that when we need more information about an issue, like:

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

we can just do @rustbot ping icebreakers-cleanup-crew and have some people helping us.

This was leadup to the design meeting last week. General conclusion was that we should try to create a working group for triage, and spastorino and pnkfelix started talking about that this week.

This has stalled out a bit, we need to talk through how to finalize the process and ensure that things are actually moving forward.

We are starting to have a "Standard checklist" for new ICE-breakers groups!

In general, we probably want to think about the working groups we'll be pushing on this year, and also a bit about different "categories" of working groups:

This is mostly taking details from the compiler contributors RFC and putting them in a central place (forge).

oli (Mar 05 2020 at 16:05, on Zulip):

wg-mir-opt

Santiago Pastorino (Mar 05 2020 at 16:15, on Zulip):

I'm glad about all the effort I've put into making MIR a not experimental project anymore :joy:

nikomatsakis (Mar 05 2020 at 19:02, on Zulip):

@oli the wg-mir-opt github team, is that synced with the team repo?

nikomatsakis (Mar 05 2020 at 19:02, on Zulip):

also, that sounds amazing

oli (Mar 05 2020 at 19:08, on Zulip):

Probably not, I'll look into it

oli (Mar 06 2020 at 08:25, on Zulip):

done https://github.com/rust-lang/team/pull/290

nikomatsakis (Mar 06 2020 at 15:06, on Zulip):

thanks @oli

Last update: May 29 2020 at 17:55UTC