Stream: t-compiler

Topic: #54818 weekly meeting 2019-02-14


pnkfelix (Feb 14 2019 at 14:42, on Zulip):

Hi @T-compiler/meeting ! We will starting our weekly meeting in this topic in about 18 minutes

pnkfelix (Feb 14 2019 at 14:42, on Zulip):

I am going to attempt to do the pretriage here within that time frame. (Wish me luck!)

pnkfelix (Feb 14 2019 at 14:43, on Zulip):

(As a reminder, the standing agenda is posted at #54818 )

pnkfelix (Feb 14 2019 at 14:43, on Zulip):

so, first, P-high

pnkfelix (Feb 14 2019 at 14:44, on Zulip):

I'll be skipping #46901, #56254, and #57374 because they are all NLL-specific things being handled by the NLL team

pnkfelix (Feb 14 2019 at 14:45, on Zulip):

Lets work bottom up today. We'll start with "PhantomData fields in repr(C) structs change ABI on aarch64" #56877

pnkfelix (Feb 14 2019 at 14:45, on Zulip):

Oh yes; as discussed two weeks ago (and noted in the last comment there) the immediate bug was addressed in PR #57645 but there were potentially still open questions.

nikomatsakis (Feb 14 2019 at 14:46, on Zulip):

hmm, I wonder if that still wants to be P-high

nikomatsakis (Feb 14 2019 at 14:46, on Zulip):

If so, I would prefer to find someone else to follow-up

pnkfelix (Feb 14 2019 at 14:46, on Zulip):

There hasn't been any response to my note about whether we should close this and open a fresh issue

pnkfelix (Feb 14 2019 at 14:46, on Zulip):

so I'll just plan on doing that

pnkfelix (Feb 14 2019 at 14:46, on Zulip):

(after the meeting is over)

pnkfelix (Feb 14 2019 at 14:46, on Zulip):

((unless someone wants to do it themselves in parallel while I continue with triage.))

pnkfelix (Feb 14 2019 at 14:47, on Zulip):

next: "Nightly rustc crashes with "unexpected region in query response"" #57464

pnkfelix (Feb 14 2019 at 14:47, on Zulip):

arg I have failed to investigate this

pnkfelix (Feb 14 2019 at 14:47, on Zulip):

my bad. Nothing to report at moment.

pnkfelix (Feb 14 2019 at 14:48, on Zulip):

So that's another action item for me: Investigate or delegate, but do something.

pnkfelix (Feb 14 2019 at 14:48, on Zulip):

next: "Regression in trait bounds that are redundant with associated type's HRTB" #57639

pnkfelix (Feb 14 2019 at 14:49, on Zulip):

In most recent comment, 21 days ago @nikomatsakis wrote "So, yes, let's say that a goal is to get a fix before the Rust 1.33 release on Thu Feb 28 2019"

pnkfelix (Feb 14 2019 at 14:49, on Zulip):

that is two weeks away

pnkfelix (Feb 14 2019 at 14:50, on Zulip):

@nikomatsakis How realistic is it to think that you might get around to this, with a beta-backportable fix, in that time frame?

pnkfelix (Feb 14 2019 at 14:50, on Zulip):

(I'll just leave a comment saying this on the issue and move along with triage)

nikomatsakis (Feb 14 2019 at 14:51, on Zulip):

I'll leave a note -- I am working with @Aaron Turon on this

nikomatsakis (Feb 14 2019 at 14:51, on Zulip):

In fact, I may re-assign it

pnkfelix (Feb 14 2019 at 14:51, on Zulip):

Next: "ICE on nightly when dereferencing boxed Iterator trait object" #57673

pnkfelix (Feb 14 2019 at 14:52, on Zulip):

ah the saga continues

pnkfelix (Feb 14 2019 at 14:52, on Zulip):

well PR #57885 was beta-nominated and beta-accepted. It presumably has not actually be backported yet.

pnkfelix (Feb 14 2019 at 14:53, on Zulip):

so okay that's another action item for me: check in about the backport(s) relevant to #57673

pnkfelix (Feb 14 2019 at 14:54, on Zulip):

last P-high bug: "ICE !ty.needs_infer() && !ty.has_placeholders() from boxing closure of type dyn for<'a> _" #57843

nikomatsakis (Feb 14 2019 at 14:55, on Zulip):

Open PR https://github.com/rust-lang/rust/pull/58056

nikomatsakis (Feb 14 2019 at 14:55, on Zulip):

I've not checked in on that though

pnkfelix (Feb 14 2019 at 14:55, on Zulip):

this will get fixed by PR #58056, which is r? @pnkfelix. So that's a fourth action item for me. :)

nikomatsakis (Feb 14 2019 at 14:56, on Zulip):

AFAIK it should work

nikomatsakis (Feb 14 2019 at 14:56, on Zulip):

though there is a red X

nikomatsakis (Feb 14 2019 at 14:56, on Zulip):

I feel like tests passed locally, maybe a transient failure

pnkfelix (Feb 14 2019 at 14:56, on Zulip):

conflicts?

pnkfelix (Feb 14 2019 at 14:56, on Zulip):

or should that not cause problems for appveyor?

nikomatsakis (Feb 14 2019 at 14:56, on Zulip):

Could be

pnkfelix (Feb 14 2019 at 14:56, on Zulip):

anyway I'll look into it.

pnkfelix (Feb 14 2019 at 14:56, on Zulip):

as part of review

nikomatsakis (Feb 14 2019 at 14:56, on Zulip):

I can rebase

pnkfelix (Feb 14 2019 at 14:57, on Zulip):

sure, that'd be great

pnkfelix (Feb 14 2019 at 14:57, on Zulip):

okay so that's all the P-high issues and a nice pile of tasks for me.

pnkfelix (Feb 14 2019 at 14:57, on Zulip):

Next beta-nominations

pnkfelix (Feb 14 2019 at 14:58, on Zulip):

oh weird there's now a merged-by-bors label...

pnkfelix (Feb 14 2019 at 14:58, on Zulip):

well, first up is "Make intern_lazy_const actually intern its argument." #58207

pnkfelix (Feb 14 2019 at 14:59, on Zulip):

seems important to land some form of this, given its ~50% reduction in rss for some inputs

pnkfelix (Feb 14 2019 at 15:00, on Zulip):

I'll put up emojis; I'm inclined to beta-accept. We can discuss whether we'd prefer a narrower change without alpha-renames.

pnkfelix (Feb 14 2019 at 15:00, on Zulip):

next: "Lower constant patterns with ascribed types." #58161

pnkfelix (Feb 14 2019 at 15:01, on Zulip):

this is fixing a pretty serious miscompilation

pnkfelix (Feb 14 2019 at 15:01, on Zulip):

so I am definitely inclined to approve the backport.

pnkfelix (Feb 14 2019 at 15:02, on Zulip):

next: "resolve: Fix span arithmetics in the import conflict error" #57908

nagisa (Feb 14 2019 at 15:02, on Zulip):

it is now 15UTC

pnkfelix (Feb 14 2019 at 15:02, on Zulip):

thanks. (sigh.)

Wesley Wiser (Feb 14 2019 at 15:02, on Zulip):

:wave:

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

Okay I'll pause the traversal of the beta-nominations

pnkfelix (Feb 14 2019 at 15:04, on Zulip):

does anyone have any comments to inject regarding the P-high issues or the beta nominations?

nagisa (Feb 14 2019 at 15:04, on Zulip):

there are spurious failures on appveyor fwiw

nagisa (Feb 14 2019 at 15:05, on Zulip):

alex’s investigating it

pnkfelix (Feb 14 2019 at 15:05, on Zulip):

okay, is the usual response to do a @bors retry ?

pnkfelix (Feb 14 2019 at 15:05, on Zulip):

or does that not work?

davidtwco (Feb 14 2019 at 15:05, on Zulip):

Can I just ask a procedural question - does a PR author perform the backport if a nomination is accepted?

pnkfelix (Feb 14 2019 at 15:06, on Zulip):

@davidtwco in recent memory that has not been the case

pnkfelix (Feb 14 2019 at 15:06, on Zulip):

or at least, the release time team has tended to get to it before anyone, from what I've seen

davidtwco (Feb 14 2019 at 15:06, on Zulip):

Just checking in case I'd have had to do something.

pnkfelix (Feb 14 2019 at 15:06, on Zulip):

it is a good idea, however, to keep your eye on it. :smile:

nagisa (Feb 14 2019 at 15:06, on Zulip):

@pnkfelix AFAIK retry works

mw (Feb 14 2019 at 15:06, on Zulip):

@davidtwco I asked the same question recently. the answer was that the author does not have to do it

mw (Feb 14 2019 at 15:07, on Zulip):

the release team likes to batch up backports for efficiency reasons

pnkfelix (Feb 14 2019 at 15:07, on Zulip):

regarding PR #57908, it fixes a bunch of ICE's and ... is probably low risk...?

pnkfelix (Feb 14 2019 at 15:07, on Zulip):

so I'm inclined to backport, and it looks from the emojis like a bunch of you all are also inclined to backport

pnkfelix (Feb 14 2019 at 15:08, on Zulip):

next: "typeck: remove leaky nested probe during trait object method resolution" #57835

pnkfelix (Feb 14 2019 at 15:08, on Zulip):

oh right we already talked about this

pnkfelix (Feb 14 2019 at 15:08, on Zulip):

I think its implicitly beta-accepted since we beta-accepted #57885

pnkfelix (Feb 14 2019 at 15:11, on Zulip):

okay that's all the beta-nominations

pnkfelix (Feb 14 2019 at 15:11, on Zulip):

no new stable-nominations to look at

pnkfelix (Feb 14 2019 at 15:12, on Zulip):

next are the stable-to-beta regressions

pnkfelix (Feb 14 2019 at 15:12, on Zulip):

some of these are not yet marked P-high so I'll focus on those

pnkfelix (Feb 14 2019 at 15:12, on Zulip):

"ICE related to lifetimes and traits" #58451

pnkfelix (Feb 14 2019 at 15:13, on Zulip):

looks like this arose from one or more of PR #55517 #56507 #57282

nagisa (Feb 14 2019 at 15:13, on Zulip):

My guess would be universes.

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

Seems pretty clear

pnkfelix (Feb 14 2019 at 15:13, on Zulip):

which are universes, polonius, and wellformed-return-ty

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

Doesn't look too serious

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

That is, I have some ideas what caused it

pnkfelix (Feb 14 2019 at 15:13, on Zulip):

you willing to take it, @nikomatsakis ?

pnkfelix (Feb 14 2019 at 15:13, on Zulip):

time is short

nikomatsakis (Feb 14 2019 at 15:14, on Zulip):

I'll take it and try to find someone to work with on it

nagisa (Feb 14 2019 at 15:14, on Zulip):

@nikomatsakis how related to the fn(&T) + trait impls is it?

pnkfelix (Feb 14 2019 at 15:14, on Zulip):

okay that sounds good

nikomatsakis (Feb 14 2019 at 15:14, on Zulip):

Perhaps @lqd, if they have time

nikomatsakis (Feb 14 2019 at 15:14, on Zulip):

@nikomatsakis how related to the fn(&T) + trait impls is it?

unclear

nikomatsakis (Feb 14 2019 at 15:14, on Zulip):

well, not at all

nikomatsakis (Feb 14 2019 at 15:14, on Zulip):

apart from being universes

pnkfelix (Feb 14 2019 at 15:14, on Zulip):

is it P-high? I assume it is, since it is a regression

nagisa (Feb 14 2019 at 15:14, on Zulip):

okay.

nikomatsakis (Feb 14 2019 at 15:14, on Zulip):

I guess so

nikomatsakis (Feb 14 2019 at 15:14, on Zulip):

I'll try to get some mentoring instructions up ASAP

nikomatsakis (Feb 14 2019 at 15:15, on Zulip):

@nagisa fyi I nominated that bug, @Aaron Turon has been doing some investigation (there is a topic over in #t-compiler/wg-traits I've not caught up on yet)

pnkfelix (Feb 14 2019 at 15:15, on Zulip):

next: "beta 1.33 seems to break tarpaulin on multithreading" #58104

pnkfelix (Feb 14 2019 at 15:15, on Zulip):

eek

pnkfelix (Feb 14 2019 at 15:15, on Zulip):

segfault while running tests

nikomatsakis (Feb 14 2019 at 15:15, on Zulip):

perhaps we can bisect?

pnkfelix (Feb 14 2019 at 15:16, on Zulip):

anyone want to volunteer to investigate further?

pnkfelix (Feb 14 2019 at 15:16, on Zulip):

(I would normally just default to myself but you can see how well that worked out two weeks ago from earlierin the meeting...)

nikomatsakis (Feb 14 2019 at 15:16, on Zulip):

it's going to be tricky to bisect, if the repo rate is only 15%

nikomatsakis (Feb 14 2019 at 15:16, on Zulip):

my sense @pnkfelix is that you have a lot of things on your plate already :)

nagisa (Feb 14 2019 at 15:16, on Zulip):

I can bisect and get a backtrace perhaps

nagisa (Feb 14 2019 at 15:16, on Zulip):

as long as it reproduces on linux

pnkfelix (Feb 14 2019 at 15:17, on Zulip):

Okay I'll assign to @nagisa for now

nikomatsakis (Feb 14 2019 at 15:17, on Zulip):

(but it still feels to me like bisection is the obvious next step)

pnkfelix (Feb 14 2019 at 15:17, on Zulip):

We'll call it P-high for now, but with a 15% reproduction rate ... well, lets just see what comes out of initial investigatin

pnkfelix (Feb 14 2019 at 15:18, on Zulip):

okay that's all the stable-to-beta regressions that were relevant to the compiler (and that we hadn't already hit as part of P-high traversal)

pnkfelix (Feb 14 2019 at 15:19, on Zulip):

next, stable-to-nightly regressions

pnkfelix (Feb 14 2019 at 15:19, on Zulip):

(again, focusing just on ones not already tagged P-high)

pnkfelix (Feb 14 2019 at 15:19, on Zulip):

first: "Cannot create local mono-item" ICE building cortex-m code on nightly #58323

nikomatsakis (Feb 14 2019 at 15:20, on Zulip):

bisect?

mw (Feb 14 2019 at 15:20, on Zulip):

I'll try to look into this

pnkfelix (Feb 14 2019 at 15:20, on Zulip):

maybe. It would be good to find out if this replicates on other targets

mw (Feb 14 2019 at 15:20, on Zulip):

yes

mw (Feb 14 2019 at 15:21, on Zulip):

if it reproduces on other targets, it shouldn't be too hard to bisect

pnkfelix (Feb 14 2019 at 15:21, on Zulip):

next: "Regression from stable to nightly: nested impl trait is not allowed" #57979

mw (Feb 14 2019 at 15:21, on Zulip):

if it's only cortex-m, we'll see if I can do anything then

pnkfelix (Feb 14 2019 at 15:21, on Zulip):

wait, didn't we already discuss this...

pnkfelix (Feb 14 2019 at 15:22, on Zulip):

well in any case, we've got to resolve this. Presumably by implementing a warning cycle or something. Not sure.

nikomatsakis (Feb 14 2019 at 15:23, on Zulip):

I think we talked about it in the lang team meeting

pnkfelix (Feb 14 2019 at 15:23, on Zulip):

@nikomatsakis you have a commit linked saying it adds a regression test for this case

pnkfelix (Feb 14 2019 at 15:23, on Zulip):

are you looking at this with @Zoxc ?

nikomatsakis (Feb 14 2019 at 15:24, on Zulip):

honestly I had forgotten all about this :P

pnkfelix (Feb 14 2019 at 15:24, on Zulip):

ah maybe PR #57981

nikomatsakis (Feb 14 2019 at 15:24, on Zulip):

let me go catch up on what is going on

nikomatsakis (Feb 14 2019 at 15:24, on Zulip):

yes, I pushed a commit to that PR, in part to test the behavior

nikomatsakis (Feb 14 2019 at 15:24, on Zulip):

I believe that @Zoxc's PR fixed an obvious bug

nikomatsakis (Feb 14 2019 at 15:24, on Zulip):

but did not cause the code in this example to work again

nikomatsakis (Feb 14 2019 at 15:25, on Zulip):

which...is maybe good

nikomatsakis (Feb 14 2019 at 15:25, on Zulip):

(unless we want to issue a warning)

nikomatsakis (Feb 14 2019 at 15:25, on Zulip):

I think the bors failure is due to a lack of a stderr file

nikomatsakis (Feb 14 2019 at 15:25, on Zulip):

I guess the question we have to decide is whether we want a warning period here

nikomatsakis (Feb 14 2019 at 15:25, on Zulip):

if so, we will need some follow-up work

pnkfelix (Feb 14 2019 at 15:26, on Zulip):

I'm going to mark this as P-high and assign it to myself, not to implement the follow-up work necessarily, but at least to push some solution forward

nikomatsakis (Feb 14 2019 at 15:26, on Zulip):

but regardless we can land @Zoxc's PR (since it doesn't actually change the behavior here)

nikomatsakis (Feb 14 2019 at 15:26, on Zulip):

ok, I just pushed the missing stderr file to #57981

pnkfelix (Feb 14 2019 at 15:27, on Zulip):

okay that's all the stable-to-nightly regressions relevant to us that we hadn't already hit via P-high traversal

pnkfelix (Feb 14 2019 at 15:27, on Zulip):

PR #57214 is marked waiting on our team

pnkfelix (Feb 14 2019 at 15:28, on Zulip):

and of course I dont think we specifically discussed this issue last week

pnkfelix (Feb 14 2019 at 15:28, on Zulip):

which is okay. We can work this out async

pnkfelix (Feb 14 2019 at 15:28, on Zulip):

(it would have been nice to figure it out face-to-face but we can manage.)

nikomatsakis (Feb 14 2019 at 15:29, on Zulip):

Yeah. So it's an interesting test case for how to approach this sort of decision

pnkfelix (Feb 14 2019 at 15:29, on Zulip):

and actually the decision here may bleed into one of the working-groups that we are planning to set up

nikomatsakis (Feb 14 2019 at 15:29, on Zulip):

there is another similar one nominated below

pnkfelix (Feb 14 2019 at 15:29, on Zulip):

so lets table this for a moment

pnkfelix (Feb 14 2019 at 15:29, on Zulip):

next T-compiler I-nominated

pnkfelix (Feb 14 2019 at 15:30, on Zulip):

looks like some these are bugs that should be triaged

pnkfelix (Feb 14 2019 at 15:31, on Zulip):

namely #58158 and #58435 are both ICE's. Probably both default to P-high

pnkfelix (Feb 14 2019 at 15:31, on Zulip):

I'll just tag them both as P-high and then we'll discuss them next week

pnkfelix (Feb 14 2019 at 15:32, on Zulip):

oh #58435 is actually fixed in nightly; so the only (potential) task is to maybe look for a PR that we might backport.

pnkfelix (Feb 14 2019 at 15:33, on Zulip):

I think we do need to set up a policy where if anyone nominates an issue, they're expected to write a comment saying why they are doing so

nikomatsakis (Feb 14 2019 at 15:33, on Zulip):

So https://github.com/rust-lang/rust/issues/58158 is not a regression

pnkfelix (Feb 14 2019 at 15:33, on Zulip):

at least its not great to be forced to infer it from comment backlog, in the general case.

nikomatsakis (Feb 14 2019 at 15:33, on Zulip):

I don't know that it should default to P-high

nikomatsakis (Feb 14 2019 at 15:33, on Zulip):

@centril nominated it, but didn't leave any references as to what they intended by that -- @centril, you around?

pnkfelix (Feb 14 2019 at 15:34, on Zulip):

okay well I don't want to further delay the next exciting parts to our meeting

pnkfelix (Feb 14 2019 at 15:34, on Zulip):

namely, my intention was to keep triage to <= 30minutes

pnkfelix (Feb 14 2019 at 15:34, on Zulip):

and then we could discuss the meta-plan for how the latter 30minutes will go in general in the future

pnkfelix (Feb 14 2019 at 15:34, on Zulip):

namely, the plan we outlined at the All Hands last week was to dedicate the latter 30 minutes of each weekly meeting to a check in with 1 (or more?) working-groups.

nikomatsakis (Feb 14 2019 at 15:34, on Zulip):

Do you want to start that discussion now? Should we decide the fate of #58158 -- just stay nominated?

pnkfelix (Feb 14 2019 at 15:35, on Zulip):

lets mark it P-high

pnkfelix (Feb 14 2019 at 15:35, on Zulip):

to ensure we talk about it next week

nikomatsakis (Feb 14 2019 at 15:35, on Zulip):

ok, we can revisit next week

Esteban Küber (Feb 14 2019 at 15:35, on Zulip):

(I just had an idea. We have many issues that get fixed when fixing other tickets. Should we have a separate test directory for bug reproductions? That way when something in that directory changes we can investigate and proactively close the tickets. Should I follow up on this idea with a ticket or internals convo?)

nikomatsakis (Feb 14 2019 at 15:35, on Zulip):

@Esteban Küber I'm not sure exactly what you mean by that -- do you mean adding minimal reproductions even before the bug is fixed to the repo?

nikomatsakis (Feb 14 2019 at 15:36, on Zulip):

If so, I am a big believer in having test cases that document the current behavior of the compiler (but don't necessarily represent the correct behavior)

pnkfelix (Feb 14 2019 at 15:36, on Zulip):

but boy oh boy you better have some good tags in the comments to highlight such tests

nikomatsakis (Feb 14 2019 at 15:37, on Zulip):

re: the checking in with WGs, I had hoped to have a start on laying out that struture -- e.g., a list of WGs etc -- but I didn't get to it. I would like to call attention to #t-compiler/wg-meta -- I had hoped to have an initial meeting whose goal was to get ourselves organized, and one of the topics I was going to raise first was creating the list of working groups and trying to pick a good order for presentations.

nikomatsakis (Feb 14 2019 at 15:37, on Zulip):

but boy oh boy you better have some good tags in the comments to highlight such tests

in a previous company I worked at, we had .ref (for "correct behavior") and .badref files (for documenting current behavior)

pnkfelix (Feb 14 2019 at 15:37, on Zulip):

@nikomatsakis do you think we should aim to have only 1 group per 30 minute slot?

nikomatsakis (Feb 14 2019 at 15:37, on Zulip):

not necessarily suggesting that :)

nikomatsakis (Feb 14 2019 at 15:37, on Zulip):

@nikomatsakis do you think we should aim to have only 1 group per 30 minute slot?

that's a good question. It may be overkill a lot of the time. I think what I would prefer is

pnkfelix (Feb 14 2019 at 15:37, on Zulip):

we seemed to have a large number of potential WG's on the list last week

nikomatsakis (Feb 14 2019 at 15:38, on Zulip):

that we kind of check-in with the "on deck" working groups

nikomatsakis (Feb 14 2019 at 15:38, on Zulip):

perhaps at the end of the triage meeting

nikomatsakis (Feb 14 2019 at 15:38, on Zulip):

to get a sense for how much they have to say

nikomatsakis (Feb 14 2019 at 15:38, on Zulip):

and to remind them :)

nikomatsakis (Feb 14 2019 at 15:38, on Zulip):

perhaps at the end of the triage meeting

perhaps async, as not everyone can attend meeting

nikomatsakis (Feb 14 2019 at 15:38, on Zulip):

and then schedule somewhere from like 1-3 depending?

pnkfelix (Feb 14 2019 at 15:38, on Zulip):

hmm, I see. So we have a time slot for full traversal of on-deck WG's

nikomatsakis (Feb 14 2019 at 15:39, on Zulip):

perhaps 3 is ambitious :) I'm imagining a common thing will be "no significant change"

pnkfelix (Feb 14 2019 at 15:39, on Zulip):

and then 1-3 deep dive?

varkor (Feb 14 2019 at 15:39, on Zulip):

there was a suggestion of having the WGs prepare some notes ahead of time for the meeting — that way it would be easier to tell how many WGs could fit into each meeting

nikomatsakis (Feb 14 2019 at 15:39, on Zulip):

yes, this is part of my motivation too

nikomatsakis (Feb 14 2019 at 15:39, on Zulip):

i.e., reach out week before to remind them to prep notes

nikomatsakis (Feb 14 2019 at 15:39, on Zulip):

however

nikomatsakis (Feb 14 2019 at 15:39, on Zulip):

there was a thought I've had in other contexts that may be relevant

nikomatsakis (Feb 14 2019 at 15:40, on Zulip):

basically that, without notes that are published sufficiently in advance, we should not proceed

nikomatsakis (Feb 14 2019 at 15:40, on Zulip):

to try and encourage said notes

nikomatsakis (Feb 14 2019 at 15:40, on Zulip):

I don't know if that's something we would want to apply here

pnkfelix (Feb 14 2019 at 15:40, on Zulip):

so maybe we could have a deadline of Tuesday for the notes

pnkfelix (Feb 14 2019 at 15:40, on Zulip):

and then whomever is running the meeting (usually me presumably) will figure out whose going to present based on those notes?

mw (Feb 14 2019 at 15:41, on Zulip):

sgtm

nikomatsakis (Feb 14 2019 at 15:41, on Zulip):

well, who prepares notes?

nikomatsakis (Feb 14 2019 at 15:41, on Zulip):

everybody?

pnkfelix (Feb 14 2019 at 15:41, on Zulip):

probably not everybody

nikomatsakis (Feb 14 2019 at 15:41, on Zulip):

only the 'next few'?

mw (Feb 14 2019 at 15:41, on Zulip):

the next 3 wgs in the queue

pnkfelix (Feb 14 2019 at 15:41, on Zulip):

we could decide at tail of each weekly meting

mw (Feb 14 2019 at 15:41, on Zulip):

or something like that

nikomatsakis (Feb 14 2019 at 15:41, on Zulip):

yeah ok

pnkfelix (Feb 14 2019 at 15:41, on Zulip):

and default to the next 3 wg's in the queue

nikomatsakis (Feb 14 2019 at 15:41, on Zulip):

I think we can kind of decide that bit on the fly

pnkfelix (Feb 14 2019 at 15:41, on Zulip):

in terms of providing guidance

nikomatsakis (Feb 14 2019 at 15:41, on Zulip):

Ideally, I think preparing notes should be a no-op

nikomatsakis (Feb 14 2019 at 15:42, on Zulip):

because I think that it'd be good to have WGs have a structure where

nikomatsakis (Feb 14 2019 at 15:42, on Zulip):

they collect notes as they come up

nikomatsakis (Feb 14 2019 at 15:42, on Zulip):

in a separate spot

nikomatsakis (Feb 14 2019 at 15:42, on Zulip):

e.g., at the end of a weekly meeting, when major things get done, or an interesting question comes up, etc

davidtwco (Feb 14 2019 at 15:42, on Zulip):

I could imagine a scenario, for example, where an NLL wg's notes would just be the status update document that is produced every week by the members of the wg for their meeting anyway.

nikomatsakis (Feb 14 2019 at 15:43, on Zulip):

maybe that's the thing to encourage? e.g., we have a central repository (or repo per WG, as the WG prefers), and it has a file like "notes.md" that you can prepend things to, and when a meeting occurs, we kind of skim those files for the WGs in question to see what's going on?

mw (Feb 14 2019 at 15:43, on Zulip):

yes, I think wgs should take every opportunity to optimize this kind of buerocratic overhead

nikomatsakis (Feb 14 2019 at 15:44, on Zulip):

I'm into this idea

nikomatsakis (Feb 14 2019 at 15:44, on Zulip):

it's worth trying out, anyway

pnkfelix (Feb 14 2019 at 15:44, on Zulip):

as long as each WG has a stable link for its notes (that get imperatively updated in place)

pnkfelix (Feb 14 2019 at 15:44, on Zulip):

we can add it to #54818

nikomatsakis (Feb 14 2019 at 15:44, on Zulip):

I was planning on creating subdirectories of https://github.com/rust-lang/compiler-team

nikomatsakis (Feb 14 2019 at 15:44, on Zulip):

or perhaps on someone else creating them :)

nikomatsakis (Feb 14 2019 at 15:44, on Zulip):

seems like that is a good starting point, and said subdirectory can include a link to a separate repo if desired

nikomatsakis (Feb 14 2019 at 15:45, on Zulip):

(for more detailed materials)

nikomatsakis (Feb 14 2019 at 15:45, on Zulip):

(I wonder if we should move towards a jekyll-based setup, with gh-pages or something, rather than just relying on GH to render the .md files etc)

nikomatsakis (Feb 14 2019 at 15:45, on Zulip):

/me shrugs

nikomatsakis (Feb 14 2019 at 15:45, on Zulip):

(I wonder if we should move towards a jekyll-based setup, with gh-pages or something, rather than just relying on GH to render the .md files etc)

that sounds like work :) I'm aiming for simplest possible overhead for now...

pnkfelix (Feb 14 2019 at 15:45, on Zulip):

doesn't GH at least have a browser-based editor at this point?

nikomatsakis (Feb 14 2019 at 15:45, on Zulip):

yes, it does

nikomatsakis (Feb 14 2019 at 15:46, on Zulip):

and it's actually kind of tolerable to use ime :)

davidtwco (Feb 14 2019 at 15:46, on Zulip):

Do we know which working groups there will be? Only those mentioned at the All Hands? Less than that? More than that?

nikomatsakis (Feb 14 2019 at 15:46, on Zulip):

I think "roughly" that, but it'll change

pnkfelix (Feb 14 2019 at 15:46, on Zulip):

Do we know which working groups there will be? Only those mentioned at the All Hands? Less than that? More than that?

I think the actual set remains to be established

pnkfelix (Feb 14 2019 at 15:47, on Zulip):

let me find those notes and link them here ...

pnkfelix (Feb 14 2019 at 15:47, on Zulip):

rustc org meeting notes

nikomatsakis (Feb 14 2019 at 15:47, on Zulip):

I was just looking for those =)

pnkfelix (Feb 14 2019 at 15:48, on Zulip):

can we maybe figure out right now how soon we can establish the set of WG's?

pnkfelix (Feb 14 2019 at 15:48, on Zulip):

like, would it be reasonable to do it in ... the next two weeks?

pnkfelix (Feb 14 2019 at 15:49, on Zulip):

or do we want something more organic?

nikomatsakis (Feb 14 2019 at 15:49, on Zulip):

I think that's a good goal

nikomatsakis (Feb 14 2019 at 15:49, on Zulip):

or rather

pnkfelix (Feb 14 2019 at 15:49, on Zulip):

where we just send out feelers and see what kind of response we get, and adjust the set accordingly?

nikomatsakis (Feb 14 2019 at 15:49, on Zulip):

I think that we should have a goal by next week of having some wgs "functional"

nikomatsakis (Feb 14 2019 at 15:49, on Zulip):

by which I mean that they are starting to use the structure and experiment

nikomatsakis (Feb 14 2019 at 15:49, on Zulip):

and others can be "in discussion", where we are deciding who/how to lead?

nikomatsakis (Feb 14 2019 at 15:49, on Zulip):

e.g., WG-nll is bsasically extant

nikomatsakis (Feb 14 2019 at 15:49, on Zulip):

similarly WG-traits "close enough"

pnkfelix (Feb 14 2019 at 15:50, on Zulip):

right, those seem like good initial candidates.

nikomatsakis (Feb 14 2019 at 15:50, on Zulip):

that is, I feel like Ic ould give some updates there :)

nikomatsakis (Feb 14 2019 at 15:50, on Zulip):

not sure about the others, how people feel, but I suspect that e.g. the Wg-parallel-compilation is in a place where we could do initial notes

nikomatsakis (Feb 14 2019 at 15:50, on Zulip):

i.e., "catching up to current state" :)

pnkfelix (Feb 14 2019 at 15:50, on Zulip):

and we can probably spin up a WG-rls2.0 too

nikomatsakis (Feb 14 2019 at 15:50, on Zulip):

confirm

pnkfelix (Feb 14 2019 at 15:50, on Zulip):

maybe not in a week...

nikomatsakis (Feb 14 2019 at 15:50, on Zulip):

we only need one or two

nikomatsakis (Feb 14 2019 at 15:50, on Zulip):

to do the first meeting :)

nikomatsakis (Feb 14 2019 at 15:50, on Zulip):

then we can add as we go

nikomatsakis (Feb 14 2019 at 15:51, on Zulip):

lay the track as the train passes over

pnkfelix (Feb 14 2019 at 15:52, on Zulip):

okay. So I take it that what's currently on the table

mw (Feb 14 2019 at 15:52, on Zulip):

is there a meta emoji?

nikomatsakis (Feb 14 2019 at 15:52, on Zulip):

man that question is breaking my mind

nikomatsakis (Feb 14 2019 at 15:52, on Zulip):

I am trying to imagine what it would even look like

pnkfelix (Feb 14 2019 at 15:52, on Zulip):

is that we're going to actually adopt this structure in time for next weeks' meeeting

nagisa (Feb 14 2019 at 15:52, on Zulip):

we can add custom ones!

pnkfelix (Feb 14 2019 at 15:53, on Zulip):

i.e. WG-traits and WG-nll should have ... "structure" ...

nikomatsakis (Feb 14 2019 at 15:53, on Zulip):

@pnkfelix I'll create two directories on compiler-team right now, one for each team, and let's (for now) add files like:

nikomatsakis (Feb 14 2019 at 15:53, on Zulip):

something like that?

davidtwco (Feb 14 2019 at 15:54, on Zulip):

Are there other things from the organization meeting that need to be discussed and implemented or is that the task of #t-compiler/wg-meta?

pnkfelix (Feb 14 2019 at 15:54, on Zulip):

@nikomatsakis the only Q I had there is whether we would let WG-nll continue using Paper or not

nikomatsakis (Feb 14 2019 at 15:54, on Zulip):

I imagined that being more the wg-meta role

nikomatsakis (Feb 14 2019 at 15:54, on Zulip):

but I think we can do it now :)

nikomatsakis (Feb 14 2019 at 15:54, on Zulip):

I just didn't want to clog up the meeting

pnkfelix (Feb 14 2019 at 15:54, on Zulip):

I personally am not wedded to continued use of Paper

nikomatsakis (Feb 14 2019 at 15:54, on Zulip):

I think it's better to use the repository, but it could be moved from paper

pnkfelix (Feb 14 2019 at 15:54, on Zulip):

its fun but it does not always load quickly.

nikomatsakis (Feb 14 2019 at 15:54, on Zulip):

i.e., paper to write the "draft"

pnkfelix (Feb 14 2019 at 15:54, on Zulip):

so yes, lets plan to migrate to the repository

davidtwco (Feb 14 2019 at 15:54, on Zulip):

I think that paper could be used and then the final notes could be added to markdown afterwards - it's just easier for everyone to check in and add to rather than making a PR with their week's notes.

nikomatsakis (Feb 14 2019 at 15:55, on Zulip):

main problem with repo is that it will be hard for people to interactively add notes etc, hence paper

pnkfelix (Feb 14 2019 at 15:55, on Zulip):

right, that sounds reasonable to me

varkor (Feb 14 2019 at 15:55, on Zulip):

are the reports literally taken from notes.md?

nikomatsakis (Feb 14 2019 at 15:55, on Zulip):

@varkor I think they can be but don't have to be

varkor (Feb 14 2019 at 15:55, on Zulip):

if so, can notes.md just have date headings?

nikomatsakis (Feb 14 2019 at 15:55, on Zulip):

we could also do that

pnkfelix (Feb 14 2019 at 15:55, on Zulip):

As long as everyone uses markdown, then any service (Paper, hackmd.io, gist) should be usable

nagisa (Feb 14 2019 at 15:55, on Zulip):

technically git preserves history itself...

nikomatsakis (Feb 14 2019 at 15:56, on Zulip):

my main thought @varkor was that I suspect it'll be useful to have "notes" on what happened, but then to try and structure them into a "narrative" for the sync up

nikomatsakis (Feb 14 2019 at 15:56, on Zulip):

but that could be done "in place" in the notes.md file

pnkfelix (Feb 14 2019 at 15:56, on Zulip):

it may depend on the team, too

varkor (Feb 14 2019 at 15:56, on Zulip):

yes, I could imagine that being convenient

pnkfelix (Feb 14 2019 at 15:56, on Zulip):

or on the particular meeting

pnkfelix (Feb 14 2019 at 15:57, on Zulip):

okay so the main thing for me is that I want to make sure by Tuesday that I know what the stable link is for the notes for each WG

pnkfelix (Feb 14 2019 at 15:57, on Zulip):

and whom to ping to ensure that the notes exist in time for me to review

pnkfelix (Feb 14 2019 at 15:57, on Zulip):

(in principle there won't be much "need" to review this time around, since there's only two teams planned...)

pnkfelix (Feb 14 2019 at 15:59, on Zulip):

so, this all sounds great

pnkfelix (Feb 14 2019 at 15:59, on Zulip):

and we're just about at the hour mark.

nikomatsakis (Feb 14 2019 at 15:59, on Zulip):

it may depend on the team, too

I imagine that each working group will have a directory whose structure includes some standard files (README.md, notes.md) plus whatever other things they want

nikomatsakis (Feb 14 2019 at 15:59, on Zulip):

I'm pushing some commits to compiler-team repo right now btw

nikomatsakis (Feb 14 2019 at 15:59, on Zulip):

and I think all compiler team members should have admin access

nikomatsakis (Feb 14 2019 at 15:59, on Zulip):

we can add more as needed

pnkfelix (Feb 14 2019 at 15:59, on Zulip):

Are there any questions or pressing concerns that anyone wants to raise while we are all here?

Cem Karan (Feb 14 2019 at 16:01, on Zulip):

Newbie question: Is there a calendar that shows when meetings are taking place?

pnkfelix (Feb 14 2019 at 16:02, on Zulip):

@Cem Karan not to my knowledge

pnkfelix (Feb 14 2019 at 16:02, on Zulip):

that would probably be a decent thing to try to arrange across all the teams/WG's...

pnkfelix (Feb 14 2019 at 16:02, on Zulip):

(At least the ones whose meetings occur in public spaces)

davidtwco (Feb 14 2019 at 16:02, on Zulip):

Are the journeyperson/apprentice idea/plans responsibility of #t-compiler/wg-meta too?

Cem Karan (Feb 14 2019 at 16:02, on Zulip):

Yes. I have no idea when meetings are going to occur, so I just kind of lurk on zulipchat...

pnkfelix (Feb 14 2019 at 16:04, on Zulip):

Are the journeyperson/apprentice idea/plans responsibility of #t-compiler/wg-meta too?

Seems like it might fall under "document how the compiler team functions" part of the charter of WG-meta

Cem Karan (Feb 14 2019 at 16:07, on Zulip):

As long as everyone uses markdown, then any service (Paper, hackmd.io, gist) should be usable

FYI, some places block things like paper and google docs. Not sure why (they do when I'm at work, I'm teleworking right now, which is the only reason I can reach it).

nikomatsakis (Feb 14 2019 at 16:07, on Zulip):

Newbie question: Is there a calendar that shows when meetings are taking place?

@Cem Karan great question! On my "short list" of things to figure out is how to make a public google calendar people can subscribe to

nikomatsakis (Feb 14 2019 at 16:07, on Zulip):

Are the journeyperson/apprentice idea/plans responsibility of #t-compiler/wg-meta too?

@davidtwco I think yes=)

nikomatsakis (Feb 14 2019 at 16:08, on Zulip):

ps, @pnkfelix, I just pushed some files to the compiler-team repository

nikomatsakis (Feb 14 2019 at 16:08, on Zulip):

feel free to change whatever or give feedback, but let's take that to #t-compiler/wg-meta

pnkfelix (Feb 14 2019 at 16:08, on Zulip):

sounds great

pnkfelix (Feb 14 2019 at 16:08, on Zulip):

okay, bye everyone @T-compiler/meeting !

centril (Feb 14 2019 at 16:28, on Zulip):

@nikomatsakis I am here now :slight_smile:

lqd (Feb 14 2019 at 17:03, on Zulip):

Perhaps @lqd, if they have time

@nikomatsakis (no need to reply) sure I'll take a look at the #58451 ICE

Last update: Nov 22 2019 at 04:30UTC