Stream: t-compiler

Topic: design meeting 2019-09-27


mw (Sep 27 2019 at 13:03, on Zulip):

Is there a design meeting today?

Wesley Wiser (Sep 27 2019 at 13:04, on Zulip):

There's a design meeting happening over in wg-mir-opt: https://rust-lang.zulipchat.com/#narrow/stream/189540-t-compiler.2Fwg-mir-opt/topic/Place's.20projection.20interning.20design.20meeting.20.3F

Wesley Wiser (Sep 27 2019 at 13:04, on Zulip):

If that's what you're thinking of

pnkfelix (Sep 27 2019 at 13:07, on Zulip):

There is a steering meeting today, which we sometimes call design meetings

pnkfelix (Sep 27 2019 at 13:07, on Zulip):

The meeting today is specifically a planning meeting for the proposals that we have on the table

pnkfelix (Sep 27 2019 at 13:07, on Zulip):

in fact this seems as good a time as any to let everyone know ...

mw (Sep 27 2019 at 13:07, on Zulip):

Ah ok. Thanks!

pnkfelix (Sep 27 2019 at 13:07, on Zulip):

Just a heads up to @T-compiler/meeting : the planning meeting will be starting in 53 minutes in this zulip topic

pnkfelix (Sep 27 2019 at 13:09, on Zulip):

In the meantime, you all can look over the meeting proposals over here

nikomatsakis (Sep 27 2019 at 14:01, on Zulip):

Hi @T-compiler/meeting -- steering meeting is now! This is a planning meeting, meaning that we'll be deciding what to talk about over the next few weeks.

nikomatsakis (Sep 27 2019 at 14:01, on Zulip):

I'm going to start by giving 5 minutes for

Announcements

nikomatsakis (Sep 27 2019 at 14:01, on Zulip):

Also last minute design proposals :)

nikomatsakis (Sep 27 2019 at 14:01, on Zulip):

/me goes to open something he didn't get around to doing yesterday

pnkfelix (Sep 27 2019 at 14:03, on Zulip):

I was going to ask what it meant that compiler-team#159 was "descheduled and replaced" with compiler-team#166; but I think I've inferred that in this context, "descheduled and replaced with" means pre-empted by, right?

pnkfelix (Sep 27 2019 at 14:04, on Zulip):

(my original interpretation was that the phrase meant 159 was subsumed by 166; but I no longer believe that to be a correct intepretation.)

Wesley Wiser (Sep 27 2019 at 14:05, on Zulip):

IIRC the meeting slot was replaced with a different topic, but the topic itself is still valid.

nikomatsakis (Sep 27 2019 at 14:06, on Zulip):

I was going to ask what it meant that compiler-team#159 was "descheduled and replaced" with compiler-team#166; but I think I've inferred that in this context, "descheduled and replaced with" means pre-empted by, right?

confirm

nikomatsakis (Sep 27 2019 at 14:06, on Zulip):

OK, not a lot of announcements, that's cool, we had a very active announcement time yesterday :)

nikomatsakis (Sep 27 2019 at 14:06, on Zulip):

We have a lot of meeting proposals this week

pnkfelix (Sep 27 2019 at 14:06, on Zulip):

Its been a while since I was at one of these things

nikomatsakis (Sep 27 2019 at 14:06, on Zulip):

Well, it's been 4 weeks

nikomatsakis (Sep 27 2019 at 14:07, on Zulip):

You just weren't at the last one :P

nikomatsakis (Sep 27 2019 at 14:07, on Zulip):

Anyway, reminder: every 4 weeks, we have one of these planning meetings to discuss what to do over the next 3 meetings

nikomatsakis (Sep 27 2019 at 14:07, on Zulip):

As a general guideline, we are supposed to discuss at least one "non-technical" topic, meaning not some kind of technical deep dive but rather a discussion about how we run the group, priorities, etc (of course it often will still hit on some technical topics)

pnkfelix (Sep 27 2019 at 14:08, on Zulip):

there are 7 8 proposals. Looking at them now, I am wondering if there is an opportunity to combine topics.

nikomatsakis (Sep 27 2019 at 14:08, on Zulip):

With that in mind, let me start by dumping out a list of the topics, and people can generally vote on what they think has highest priority. We'll do this with approval voting, meaning put a :+1: emoji on as many as you like.

nikomatsakis (Sep 27 2019 at 14:09, on Zulip):
nikomatsakis (Sep 27 2019 at 14:09, on Zulip):
nikomatsakis (Sep 27 2019 at 14:09, on Zulip):
nikomatsakis (Sep 27 2019 at 14:09, on Zulip):
nikomatsakis (Sep 27 2019 at 14:09, on Zulip):
nikomatsakis (Sep 27 2019 at 14:09, on Zulip):
nikomatsakis (Sep 27 2019 at 14:09, on Zulip):
nikomatsakis (Sep 27 2019 at 14:10, on Zulip):
nikomatsakis (Sep 27 2019 at 14:10, on Zulip):
nikomatsakis (Sep 27 2019 at 14:10, on Zulip):

there are 7 8 proposals. Looking at them now, I am wondering if there is an opportunity to combine topics.

indeed. I think one of the other things we can do in this meeting is to give feedback on a proposal, or maybe close, or combine, or just generally tweak.

nikomatsakis (Sep 27 2019 at 14:12, on Zulip):

(In particular, I think we may should consider what amount of prep we want before a given meeting, and assign someone to do it, since most of these proposals are really just paragraphs -- and that's ok, I think that should be our new process.)

nikomatsakis (Sep 27 2019 at 14:12, on Zulip):

(Is anyone still considering what to vote for who wants to vote? I'm going to wait another minute or so)

nikomatsakis (Sep 27 2019 at 14:12, on Zulip):

@pnkfelix did you have specific "merges" in mind?

pnkfelix (Sep 27 2019 at 14:13, on Zulip):

I was considering merging ICE-breaker into Maintenance and Triage

nikomatsakis (Sep 27 2019 at 14:13, on Zulip):

I was considering just closing that

pnkfelix (Sep 27 2019 at 14:13, on Zulip):

the working-group review may or may not fit in with them as well. It probably deserves its own slot given how many working groups there are

nikomatsakis (Sep 27 2019 at 14:13, on Zulip):

Maybe I'll just do that -- we're moving forward now, we can schedule a meeting if it seems like there are open questions worth discussing.

pnkfelix (Sep 27 2019 at 14:13, on Zulip):

Just closing ICE-breakers, that is?

nikomatsakis (Sep 27 2019 at 14:14, on Zulip):

yes

nikomatsakis (Sep 27 2019 at 14:14, on Zulip):

I'm going to do it! :decisive:

nikomatsakis (Sep 27 2019 at 14:14, on Zulip):

OK, enough voting.

nikomatsakis (Sep 27 2019 at 14:15, on Zulip):

Top candidates, non-technical:

Top candidates, technical:

pnkfelix (Sep 27 2019 at 14:16, on Zulip):

I think the roadmap is higher priority than constitutional convention, and I think the voting tallies support that thinking.

pnkfelix (Sep 27 2019 at 14:16, on Zulip):

we've been muddling along so far, we can keep going for a while yet

nikomatsakis (Sep 27 2019 at 14:17, on Zulip):

Yes, I agree -- though I think it's probably a good idea to start trying to actually draft a more specific proposal on the side. I'd rather not discuss that further in this meeting though.

nikomatsakis (Sep 27 2019 at 14:17, on Zulip):

(I have thoughts I'll post in a separate topic)

pnkfelix (Sep 27 2019 at 14:17, on Zulip):

yes. Maybe in fact a road map item for 2020 would be to have such a convention

nikomatsakis (Sep 27 2019 at 14:17, on Zulip):

One question I have about the merge side-effect change:

nikomatsakis (Sep 27 2019 at 14:17, on Zulip):

So I think we need to reach a decision here

nikomatsakis (Sep 27 2019 at 14:18, on Zulip):

The current status is "we have some data that shows some slowdown, and we have (I think) no better ideas on how to do this"

nikomatsakis (Sep 27 2019 at 14:18, on Zulip):

I'm not sure if this is really a good use of meeting time. Maybe we want to delegate some folks to decide? This is kind of entangled in the "constitution" topic :P but yeah

mw (Sep 27 2019 at 14:19, on Zulip):

yes, I would expect people to show up more or less uninformed

pnkfelix (Sep 27 2019 at 14:19, on Zulip):

well, is our concern that we might have underestimated the true cost for some important customers?

nikomatsakis (Sep 27 2019 at 14:19, on Zulip):

I guess the point of the meeting is maybe more "how do we decide". (Caveat: I'm kind of inclined to think we should just bite the bullet and do this)

pnkfelix (Sep 27 2019 at 14:19, on Zulip):

or are we assuming that our estimate is accurate, and the question is just whether the cost is acceptable?

nikomatsakis (Sep 27 2019 at 14:19, on Zulip):

well, is our concern that we might have underestimated the true cost for some important customers?

I think that's probably a factor -- I think we don't have very well articulated concerns

nikomatsakis (Sep 27 2019 at 14:20, on Zulip):

I guess I could see e.g. picking two or three "real world" applications, let's say firefox stylo, maybe something from linkerd/tokio, and measuring the impact

nikomatsakis (Sep 27 2019 at 14:20, on Zulip):

i.e., people who have pre-defined benchmarks of their own

nagisa (Sep 27 2019 at 14:20, on Zulip):

I voted for that change to figure out how we want to proceed, there is more than just "lets merge" or "lets not merge and wait" to it

nikomatsakis (Sep 27 2019 at 14:20, on Zulip):

(cc @Carl Lerche -- could you maybe help us in measuring perf impact on linkerd etc?)

nikomatsakis (Sep 27 2019 at 14:20, on Zulip):

I voted for that change to figure out how we want to proceed, there is more than just "lets merge" or "lets not merge and wait" to it

ok!

nikomatsakis (Sep 27 2019 at 14:21, on Zulip):

@nagisa do you think it's worth having a meeting, or do you feel like there is more steps to do first?

pnkfelix (Sep 27 2019 at 14:21, on Zulip):

@nagisa so you think its worth a Friday steering meeting of the team?

nagisa (Sep 27 2019 at 14:21, on Zulip):

but the "more" is also "what the actual heck is happening with LLVM’s direction wrt forward-progress" and I myself have no idea.

pnkfelix (Sep 27 2019 at 14:21, on Zulip):

@nagisa or do you think we could grab some volunteers from those present now to investigate/deep-dive?

nagisa (Sep 27 2019 at 14:21, on Zulip):

so I don’t think meeting is a correct medium to discuss this per se.

mw (Sep 27 2019 at 14:21, on Zulip):

is it possible to merge the change behind a -Z flag? that would make it easier to test for Firefox

nikomatsakis (Sep 27 2019 at 14:21, on Zulip):

Ah, nice

nikomatsakis (Sep 27 2019 at 14:21, on Zulip):

Yes, that is possible

nikomatsakis (Sep 27 2019 at 14:22, on Zulip):

Maybe we should just agree to do that and delegate to someone to get "real world" measurements on firefox and 2 or 3 other projects?

nikomatsakis (Sep 27 2019 at 14:22, on Zulip):

We can leave the meeting proposal open to revisit the topic once we have that data?

mw (Sep 27 2019 at 14:22, on Zulip):

I can do Firefox if this is available in a regular nightly

nagisa (Sep 27 2019 at 14:22, on Zulip):

I don’t mind merging it behind a flag. If anything this will prevent this code from bitrotting endlessly

nikomatsakis (Sep 27 2019 at 14:23, on Zulip):

I'd volunteer to do something else

nikomatsakis (Sep 27 2019 at 14:23, on Zulip):

or at least pester someone to do so :P

nikomatsakis (Sep 27 2019 at 14:23, on Zulip):

OK, let's do that

pnkfelix (Sep 27 2019 at 14:23, on Zulip):

I don’t mind merging it behind a flag. If anything this will prevent this code from bitrotting endlessly

To be clear: merging it on-by-default with flag to disable? Or other way around?

pnkfelix (Sep 27 2019 at 14:23, on Zulip):

(must have meant other way around)

nikomatsakis (Sep 27 2019 at 14:23, on Zulip):

I think the idea was a flag to enable

mw (Sep 27 2019 at 14:23, on Zulip):

the other way round

mw (Sep 27 2019 at 14:23, on Zulip):

yes

nikomatsakis (Sep 27 2019 at 14:23, on Zulip):

default users have no change

pnkfelix (Sep 27 2019 at 14:23, on Zulip):

Glad we're all in sync

pnkfelix (Sep 27 2019 at 14:24, on Zulip):

so okay, that's one tech topic taken care of

nikomatsakis (Sep 27 2019 at 14:25, on Zulip):

posted comment

nikomatsakis (Sep 27 2019 at 14:25, on Zulip):

I'll post on the PR too

nikomatsakis (Sep 27 2019 at 14:26, on Zulip):

OK, done

nikomatsakis (Sep 27 2019 at 14:26, on Zulip):

current status:

Top candidates, non-technical:

Top candidates, technical:

nikomatsakis (Sep 27 2019 at 14:26, on Zulip):

Let's talk about compiler-team#175?

nikomatsakis (Sep 27 2019 at 14:26, on Zulip):

I think we need to start making some decisions here

pnkfelix (Sep 27 2019 at 14:26, on Zulip):

First off: is @Zoxc present here?

nikomatsakis (Sep 27 2019 at 14:27, on Zulip):

Unknown, but we would definitely want them to attend any meeting we schedule!

pnkfelix (Sep 27 2019 at 14:28, on Zulip):

can we schedule it with a fallback topic if we learn @Zoxc cannot attend the slot we choose?

nikomatsakis (Sep 27 2019 at 14:28, on Zulip):

I also think that @**Zoxc**'s writeup on internals is pretty decent in terms of laying out the various PRs they've opened. I guess one question for me is whether these PRs represent an end-state or just an intermediate step to something else (and what that is)

nikomatsakis (Sep 27 2019 at 14:29, on Zulip):

can we schedule it with a fallback topic if we learn Zoxc cannot attend the slot we choose?

yeah I think we can adjust schedules a bit as needed

pnkfelix (Sep 27 2019 at 14:29, on Zulip):

(where we would try to establish well ahead of time, say Tuesday, whether @Zoxc is available)

pnkfelix (Sep 27 2019 at 14:29, on Zulip):

Okay

nikomatsakis (Sep 27 2019 at 14:29, on Zulip):

I think it'd be useful to sharpen a bit what we aim to get out of the meeting exactly

nikomatsakis (Sep 27 2019 at 14:29, on Zulip):

Though I could also imagine doing that async

nikomatsakis (Sep 27 2019 at 14:30, on Zulip):

no, I can't, I take that back

nikomatsakis (Sep 27 2019 at 14:30, on Zulip):

(what I can imagine is that we discuss here in general terms and that we async do work to try and prep for that goal)

nikomatsakis (Sep 27 2019 at 14:33, on Zulip):

OK. So, one thing I am imagining:

nikomatsakis (Sep 27 2019 at 14:33, on Zulip):
nikomatsakis (Sep 27 2019 at 14:34, on Zulip):
nikomatsakis (Sep 27 2019 at 14:35, on Zulip):

I sort of incline to the first, I think it'd be interesting and helpful to the ultimate reviewers. Depending how it goes, we could rinse-and-repeat for other PRs.

nikomatsakis (Sep 27 2019 at 14:35, on Zulip):

I think part of that discussion would be evaluating the flaws in the setup and how we might try to fix them, which is kind of what I mean by end-state

nikomatsakis (Sep 27 2019 at 14:35, on Zulip):

That said, one concern I have overall is that I think we need to do follow-through

nikomatsakis (Sep 27 2019 at 14:36, on Zulip):

e.g., with the parallel work, we've landed a bunch of PRs, but I think we're now struggling to organize the follow-up work to do documentation, cleanup, refactoring, measurement

nikomatsakis (Sep 27 2019 at 14:36, on Zulip):

(and hence maybe it's valid to say "we don't want to land these PRs right now")

nikomatsakis (Sep 27 2019 at 14:36, on Zulip):

but this might also be topics to discuss in the context of some specific PR?

pnkfelix (Sep 27 2019 at 14:36, on Zulip):

I am interested in the loading the dep graph in the background

nikomatsakis (Sep 27 2019 at 14:37, on Zulip):

If we did this plan I am proposing, I think it woudl be best to pick someone (say you, @pnkfelix) to work with zoxc and the PR ahead of time, to try and prep a bit more before the meeting (i.e., try to understand/document the design more deeply)

pnkfelix (Sep 27 2019 at 14:37, on Zulip):

(I'm not clear on where that change falls in the series oF PR's; its just something that stood out to me reading over @Zoxc 's post on internals)

nikomatsakis (Sep 27 2019 at 14:37, on Zulip):

in my ideal world, this would take the form of a hackmd doc that becomes a PR to the rustc-guide

nikomatsakis (Sep 27 2019 at 14:38, on Zulip):

and that would be the outcome of the meeting :)

mw (Sep 27 2019 at 14:38, on Zulip):

we are already loading the dep-graph in the background, IIRC

pnkfelix (Sep 27 2019 at 14:38, on Zulip):

yes I was wondering about trying to work more directly with Zoxc to make, at least, a more goal-oriented introduction to these changes

pnkfelix (Sep 27 2019 at 14:38, on Zulip):

we are already loading the dep-graph in the background, IIRC

Is that something that was introduced "recently", e.g. after @Zoxc 's post which was put up on Jun 20th?

mw (Sep 27 2019 at 14:39, on Zulip):

it's been in there for quite a while. maybe we are talking about different things though

pnkfelix (Sep 27 2019 at 14:39, on Zulip):

hmm. okay.

nikomatsakis (Sep 27 2019 at 14:39, on Zulip):

@pnkfelix there is a PR that changes various aspects of dep-graph loading, though, in connection with the parallel stuff

nikomatsakis (Sep 27 2019 at 14:39, on Zulip):

and the internals post talks about changing some things there to help extend queries to parsing

nikomatsakis (Sep 27 2019 at 14:40, on Zulip):

OK, it's 10:40, I'd like to reach a decision :)

pnkfelix (Sep 27 2019 at 14:40, on Zulip):

anyway, I think I would be willing to try to work with @Zoxc ; the one doubt I have about it is that I do not know off-hand what time zone they are in nor what hours they tend to be active online

nikomatsakis (Sep 27 2019 at 14:40, on Zulip):

I believe they are in Europe

nikomatsakis (Sep 27 2019 at 14:40, on Zulip):

Though whether that makes their hours align with yours, I can't say :P

pnkfelix (Sep 27 2019 at 14:40, on Zulip):

right. Okay.

pnkfelix (Sep 27 2019 at 14:40, on Zulip):

well lets tentatively go with this as the plan.

nikomatsakis (Sep 27 2019 at 14:40, on Zulip):

Why don't we tentatively schedule a meeting 2 weeks out

nikomatsakis (Sep 27 2019 at 14:40, on Zulip):

Here are the dates:

nikomatsakis (Sep 27 2019 at 14:41, on Zulip):
nikomatsakis (Sep 27 2019 at 14:41, on Zulip):

I realize now I will not be here on Oct 18

nikomatsakis (Sep 27 2019 at 14:41, on Zulip):

because of RustBeltRust

mw (Sep 27 2019 at 14:41, on Zulip):

/me is not available Oct 4

nikomatsakis (Sep 27 2019 at 14:42, on Zulip):

:)

nikomatsakis (Sep 27 2019 at 14:42, on Zulip):

I feel like I would definitely want to be here for roadmap discussion

pnkfelix (Sep 27 2019 at 14:42, on Zulip):

is Oct 4th too soon for a roadmap discussion?

nikomatsakis (Sep 27 2019 at 14:42, on Zulip):

I would like to be here for all discussions of course, but could probably "catch up" a bit on some of the more technical ones, at least...in theory

pnkfelix (Sep 27 2019 at 14:43, on Zulip):

(Oct 4th definitely too soon for me to coordinate with Zoxc)

nikomatsakis (Sep 27 2019 at 14:43, on Zulip):

hmm, maybe we should talk about roadmap in a bit more detail?

nikomatsakis (Sep 27 2019 at 14:43, on Zulip):

what kind of prep would we want to do beforehand, and who would do it?

nikomatsakis (Sep 27 2019 at 14:43, on Zulip):

I don't think our goal should be to have a decision or anything

nikomatsakis (Sep 27 2019 at 14:43, on Zulip):

I could imagine trying to evaluate what we had planned from 2019

nikomatsakis (Sep 27 2019 at 14:43, on Zulip):

and get a sense for what progress we made

eddyb (Sep 27 2019 at 14:44, on Zulip):

(random aside: is 2020 going to be the main design/impl year for Rust 2021?)

nikomatsakis (Sep 27 2019 at 14:44, on Zulip):

A good question! I think the answer is yes, that's how we should think about it.

pnkfelix (Sep 27 2019 at 14:44, on Zulip):

is the next edition indeed going to be 2021 ?

nikomatsakis (Sep 27 2019 at 14:44, on Zulip):

It's not something that's been firmly or broadly discussed

mw (Sep 27 2019 at 14:44, on Zulip):

my roadmap proposal for 2020, just one goal: remove Rust dylibs :drop: :mic:

nikomatsakis (Sep 27 2019 at 14:44, on Zulip):

heh =)

mw (Sep 27 2019 at 14:45, on Zulip):

that's a good point, @eddyb

nikomatsakis (Sep 27 2019 at 14:45, on Zulip):

It's not something that's been firmly or broadly discussed

I think that's something that should be discussed as part of the overall roadmap planning effort, tbh

pnkfelix (Sep 27 2019 at 14:45, on Zulip):

across the project as a whole you mean, yeah?

nikomatsakis (Sep 27 2019 at 14:45, on Zulip):

In partcular, I would want to identify any potential 'edition migrations' (i.e., "breaking" changes) early

nikomatsakis (Sep 27 2019 at 14:45, on Zulip):

across the project as a whole you mean, yeah?

yeah

eddyb (Sep 27 2019 at 14:46, on Zulip):

Rust 2021 dylibs will be just cdylibs :P

nikomatsakis (Sep 27 2019 at 14:46, on Zulip):

My opinion though is that I take our Edition messaging seriously

nikomatsakis (Sep 27 2019 at 14:46, on Zulip):

And Rust 2021 is not just about "migations"

nikomatsakis (Sep 27 2019 at 14:46, on Zulip):

i.e., it should be about trying to fix some big problems people feel

nikomatsakis (Sep 27 2019 at 14:46, on Zulip):

I would include IDE/compilation time in that list...perhaps top of it.

nikomatsakis (Sep 27 2019 at 14:47, on Zulip):

So I think yeah we should be talking about Rust 2021 and thinking on that scale

nikomatsakis (Sep 27 2019 at 14:47, on Zulip):

put another way:

nikomatsakis (Sep 27 2019 at 14:47, on Zulip):

whatever we thought we could get done in 2019

nikomatsakis (Sep 27 2019 at 14:47, on Zulip):

is probably what we can realistically get done by 2021

nikomatsakis (Sep 27 2019 at 14:47, on Zulip):

:P

mw (Sep 27 2019 at 14:47, on Zulip):

that seems about right :)

nikomatsakis (Sep 27 2019 at 14:47, on Zulip):

anyway, @mw, how would you feel about missing a roadmap meeting?

eddyb (Sep 27 2019 at 14:48, on Zulip):

every year is actually 3 years when you account for all the slowdowns

nikomatsakis (Sep 27 2019 at 14:48, on Zulip):

I say "a" beacuse I expect further iteration

mw (Sep 27 2019 at 14:48, on Zulip):

I would be OK with that

nikomatsakis (Sep 27 2019 at 14:48, on Zulip):

every year is actually 3 years when you account for all the slowdowns

a colleague of mine used to say... to get a realistic time estimate, take your real estimate, double it, and then up the unit. So 4 days == 8 weeks, 2 months == 4 years, etc

nikomatsakis (Sep 27 2019 at 14:48, on Zulip):

I'm sorry to say that's proven more true to me than I would ever have known.

eddyb (Sep 27 2019 at 14:49, on Zulip):

Zulip redraws every half a minute for me r ight now so this adds up tbh

eddyb (Sep 27 2019 at 14:49, on Zulip):

I'm sorry to say that's proven more true to me than I would ever have known.

ditto

nikomatsakis (Sep 27 2019 at 14:49, on Zulip):
nikomatsakis (Sep 27 2019 at 14:49, on Zulip):

what are some of the "less voted" topics...

pnkfelix (Sep 27 2019 at 14:50, on Zulip):

debuginfo ?

nikomatsakis (Sep 27 2019 at 14:50, on Zulip):
pnkfelix (Sep 27 2019 at 14:50, on Zulip):

I imagine I could run that without niko present

eddyb (Sep 27 2019 at 14:50, on Zulip):

"?? (no niko)" sounds like my time to shine but I'm not going to make the mistake of promising anything ever again

nikomatsakis (Sep 27 2019 at 14:50, on Zulip):

lol

nikomatsakis (Sep 27 2019 at 14:51, on Zulip):

(man, Zulip won't let me add emoji to any comments right now -- the little emoji cursor shows up, but as soon as my mouse moves over it, it disappears! :angry: I feel like I'm being played with)

nikomatsakis (Sep 27 2019 at 14:51, on Zulip):

debuginfo ?

that works for me

nikomatsakis (Sep 27 2019 at 14:51, on Zulip):

(man, Zulip won't let me add emoji to any comments right now -- the little emoji cursor shows up, but as soon as my mouse moves over it, it disappears! :angry: I feel like I'm being played with)

(I suspect this bug is induced by some FF plugin I have or something, to be fair)

nikomatsakis (Sep 27 2019 at 14:52, on Zulip):

I imagine I could run that without niko present

we can also take a break if we want

nikomatsakis (Sep 27 2019 at 14:52, on Zulip):

but I agree I don't feel the need to be present for debuginfo

pnkfelix (Sep 27 2019 at 14:52, on Zulip):

lets schedule it for now

mw (Sep 27 2019 at 14:52, on Zulip):

well "debuginfo" is not really about debuginfo, right?

nikomatsakis (Sep 27 2019 at 14:52, on Zulip):
nikomatsakis (Sep 27 2019 at 14:53, on Zulip):

well "debuginfo" is not really about debuginfo, right?

yeah it's really about "how can we get a sustainable story"

mw (Sep 27 2019 at 14:53, on Zulip):

it's about making a decision what to do with tier 1 features that get tier 3 treatment

nikomatsakis (Sep 27 2019 at 14:53, on Zulip):

(well put)

pnkfelix (Sep 27 2019 at 14:53, on Zulip):

it's about making a decision what to do with tier 1 features that get tier 3 treatment

that's a more general question though, right?

nikomatsakis (Sep 27 2019 at 14:53, on Zulip):

I don't thikn we have to decide this "for every case right now" or something though

mw (Sep 27 2019 at 14:53, on Zulip):

it's the question we'll end up discussing :)

nikomatsakis (Sep 27 2019 at 14:53, on Zulip):

maybe good to talk about what we plan to talk about in the last 6-7 minutes

pnkfelix (Sep 27 2019 at 14:54, on Zulip):

I guess it will end up there unless a miracle occurs

nikomatsakis (Sep 27 2019 at 14:54, on Zulip):

I would imagine trying to figure out (a) what steps can we do to recruit folks

nikomatsakis (Sep 27 2019 at 14:54, on Zulip):

(b) what are some thresholds and timelines where upon we consider more drastic action

nikomatsakis (Sep 27 2019 at 14:54, on Zulip):

(c) and what might that drastic action be :)

nikomatsakis (Sep 27 2019 at 14:54, on Zulip):

maybe we already know the answers to those questions

nikomatsakis (Sep 27 2019 at 14:55, on Zulip):

well not the exact timeline, but it seems like the obvious first step is to try and do some kind of "call for help" sort of thing

mw (Sep 27 2019 at 14:56, on Zulip):

it sounds like something where it would make sense for Niko to be present

pnkfelix (Sep 27 2019 at 14:56, on Zulip):

well, the other candidate I could imagine would be "compile time analysis and assistance "

nikomatsakis (Sep 27 2019 at 14:57, on Zulip):

that was my proposal :P but yeah

pnkfelix (Sep 27 2019 at 14:57, on Zulip):

but to be honest I think we'd want Niko there for that too

mw (Sep 27 2019 at 14:57, on Zulip):

we can just skip that date then

nikomatsakis (Sep 27 2019 at 14:57, on Zulip):

it seems ok to me to leave it unscheduled

nikomatsakis (Sep 27 2019 at 14:57, on Zulip):

I like my fridays :)

pnkfelix (Sep 27 2019 at 14:57, on Zulip):

yeah lets leave it unscheduled.

pnkfelix (Sep 27 2019 at 14:57, on Zulip):

but

nikomatsakis (Sep 27 2019 at 14:57, on Zulip):

though my dang friday will be busy anyway :shakes-fist-in-impotent-rage:

pnkfelix (Sep 27 2019 at 14:57, on Zulip):

well

pnkfelix (Sep 27 2019 at 14:57, on Zulip):

we often get beta-backport requests ...

pnkfelix (Sep 27 2019 at 14:58, on Zulip):

that end up spilling into there

pnkfelix (Sep 27 2019 at 14:58, on Zulip):

(maybe "often" is too strong a word)

nikomatsakis (Sep 27 2019 at 14:58, on Zulip):

well

nikomatsakis (Sep 27 2019 at 14:58, on Zulip):

another option would be to move the meeting earlier in the week

nikomatsakis (Sep 27 2019 at 14:58, on Zulip):

I'm also not sure how much I'm really needed

nikomatsakis (Sep 27 2019 at 14:58, on Zulip):

I could do some pre-discussion maybe with felix

pnkfelix (Sep 27 2019 at 14:58, on Zulip):

hmm. That might be a good idea in general... I like my fridays too.

nikomatsakis (Sep 27 2019 at 14:58, on Zulip):

put in my general feeling

pnkfelix (Sep 27 2019 at 14:58, on Zulip):

but I don't know whether that will help or hurt our effort to get engagement here.

nikomatsakis (Sep 27 2019 at 14:59, on Zulip):

OK -- meeting ends in 1 minute!

nikomatsakis (Sep 27 2019 at 14:59, on Zulip):

question is basically:

nikomatsakis (Sep 27 2019 at 15:00, on Zulip):

I'm not inclined to change the day

nikomatsakis (Sep 27 2019 at 15:00, on Zulip):

because I have a phobia of calendars and doodle

nikomatsakis (Sep 27 2019 at 15:00, on Zulip):

(but I think it might be worth consdering at some point)

nikomatsakis (Sep 27 2019 at 15:00, on Zulip):

I'm good with either of the other two, but I think if we do "nothing" we should still try to take some steps

pnkfelix (Sep 27 2019 at 15:00, on Zulip):

yes

pnkfelix (Sep 27 2019 at 15:00, on Zulip):

this problem is causing issues for T-release

nikomatsakis (Sep 27 2019 at 15:01, on Zulip):

I say schedule it

pnkfelix (Sep 27 2019 at 15:01, on Zulip):

lets schedule it and still try to resolve the specific problem T-release is facing

nikomatsakis (Sep 27 2019 at 15:01, on Zulip):

yes, that

nikomatsakis (Sep 27 2019 at 15:01, on Zulip):

I was just typing that :)

pnkfelix (Sep 27 2019 at 15:01, on Zulip):

in a nearer term

nikomatsakis (Sep 27 2019 at 15:02, on Zulip):

OK, thanks all for a great meeting! y'all are the best! :hug:

pnkfelix (Sep 27 2019 at 15:02, on Zulip):

and that way, if we end up cancelling Oct 18th meeting, it will not hurt as much

pnkfelix (Sep 27 2019 at 15:02, on Zulip):

okay bye @T-compiler/meeting ; thanks for attending!

nikomatsakis (Sep 27 2019 at 15:02, on Zulip):
nikomatsakis (Sep 27 2019 at 15:03, on Zulip):

I'll go update a few things

nikomatsakis (Sep 27 2019 at 15:24, on Zulip):

OK, calendar is updated. One thing I wanted to ask -- @pnkfelix -- do you think that it might make sense in the debuginfo meeting to talk a bit about some specific PRs, too? (Also, where should we discuss the more immediate follow-up that's needed? Maybe break out a separate topic for that...)

Carl Lerche (Sep 27 2019 at 15:53, on Zulip):

@nikomatsakis I was pinged... but i don't know how to use zulip really... i would be happy to help w/ linkerd stuff... for linkerd, my guess is there is a lot that we can do on our end to speed up compilation times, but it's a black box

centril (Sep 27 2019 at 16:37, on Zulip):

Missed the meeting but... re. #59546 it is good that there's some movement here. But... I would like to see a deadline (e.g. 6 months) after which the -Z flag becomes the default. I think it's not OK to call Rust a safe language and then let a soundness hole with known miscompilations to sit around for so long.

pnkfelix (Sep 30 2019 at 07:37, on Zulip):

@nikomatsakis talking about specific PR's makes sense, if for no other reason than to have a reference point for the overall conversation.

Last update: Nov 20 2019 at 01:10UTC