Stream: t-compiler

Topic: steering meeting 2019.05.03 #58850


matklad (May 03 2019 at 14:08, on Zulip):

Aren't we supposed to have a streering meeting right now?

eddyb (May 03 2019 at 14:08, on Zulip):

I thought so too!

Wesley Wiser (May 03 2019 at 14:09, on Zulip):

It's on the compiler team calendar

eddyb (May 03 2019 at 14:09, on Zulip):

@nikomatsakis doesn't seem to be around

eddyb (May 03 2019 at 14:10, on Zulip):

@matklad I realized today that I screwed up by not focusing on choice of allocation, btw, in my writeup. I think allocation scheme, and whether in-place mutation ever happens, is super important. I guess the "mutable visitor" change shook me too much

pnkfelix (May 03 2019 at 14:12, on Zulip):

Maybe I should drive. Let me get to a laptop

eddyb (May 03 2019 at 14:13, on Zulip):

how much do you like minutia of parsing and ASTs?

pnkfelix (May 03 2019 at 14:13, on Zulip):

i didn't think that was topic today?

matklad (May 03 2019 at 14:14, on Zulip):

created https://rust-lang.zulipchat.com/#narrow/stream/131828-t-compiler/topic/mem-management.20of.20syntax.20trees/near/164797396 for this offtopic

mw (May 03 2019 at 14:14, on Zulip):

what _was_ the topic today? :)

pnkfelix (May 03 2019 at 14:14, on Zulip):

in particular, I thought topic was based on: https://rust-lang.zulipchat.com/#narrow/stream/131828-t-compiler/topic/pre-steering.20meeting.202019.2E05.2E03/near/164555548

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

namely, reviewing the design meeting process

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

see also: https://github.com/rust-lang/compiler-team/blob/master/procedures/steering-meeting.md

eddyb (May 03 2019 at 14:15, on Zulip):

@pnkfelix so, @nikomatsakis split my proposal topic out of that topic into the "AST/parsing etc." one

eddyb (May 03 2019 at 14:15, on Zulip):

but this week being meta seems fine

eddyb (May 03 2019 at 14:15, on Zulip):

it's not like my proposal is anything but half-written anyway

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

right. @nikomatsakis also said that the meta stuff may not take up the whole slot

pnkfelix (May 03 2019 at 14:16, on Zulip):

so lets maybe get meta stuff out of the way first

pnkfelix (May 03 2019 at 14:16, on Zulip):

as you can see from the steering-meeting.md, the intention is to keep a four-week cyclic rhythm

pnkfelix (May 03 2019 at 14:17, on Zulip):

where the first week is planning for the next three weeks, and the fourth week is always reserved for a non-technical discussion

pnkfelix (May 03 2019 at 14:18, on Zulip):

one detail here is that, from steering-meeting.md, I infer that we'd prefer more structure than just shooting off ideas in this chat room

pnkfelix (May 03 2019 at 14:19, on Zulip):

namely, there's a hope that actual meeting proposals will have a PR that adds a corresponding file in the proposals/ directory

pnkfelix (May 03 2019 at 14:19, on Zulip):

but I'm not sure if that's meant to be a pre-req before any sort of discussion at the meeting?

pnkfelix (May 03 2019 at 14:19, on Zulip):

I guess it is

mw (May 03 2019 at 14:19, on Zulip):

I think it is

pnkfelix (May 03 2019 at 14:20, on Zulip):

okay

pnkfelix (May 03 2019 at 14:20, on Zulip):

/me looks to see if there are any files/PRs that correspond to such proposals

mw (May 03 2019 at 14:20, on Zulip):

I think this will be interesting

eddyb (May 03 2019 at 14:20, on Zulip):

I talked to @nikomatsakis and he said he'd be fine with issues or other documents (Dropbox paper/hackmd), since I think merging a file into a shared repo seems a bit much to me

mw (May 03 2019 at 14:20, on Zulip):

RFCs have turned out to be too heavyweight for this kind of thing

pnkfelix (May 03 2019 at 14:20, on Zulip):

/me looks to see if there are any files/PRs that correspond to such proposals

(answer appears to be "no")

eddyb (May 03 2019 at 14:20, on Zulip):

but I agree on having a prepared proposal of some sort

pnkfelix (May 03 2019 at 14:20, on Zulip):

I don't mind there being other formats

pnkfelix (May 03 2019 at 14:21, on Zulip):

but we do need a central way to at least gather them

pnkfelix (May 03 2019 at 14:21, on Zulip):

right?

eddyb (May 03 2019 at 14:21, on Zulip):

I wish you could just link something like hackmd with github better

eddyb (May 03 2019 at 14:21, on Zulip):

other than a link in an issue

eddyb (May 03 2019 at 14:21, on Zulip):

or maybe we discuss proposals in our own forks?

eddyb (May 03 2019 at 14:21, on Zulip):

and merge them when we want to commit to them

pnkfelix (May 03 2019 at 14:22, on Zulip):

maybe this meeting is the central gathering point ?

pnkfelix (May 03 2019 at 14:22, on Zulip):

e.g. once you have something written up, you post it here

eddyb (May 03 2019 at 14:22, on Zulip):

yeah I suppose we can do whatever, and then put it into the compiler-team repo if accepted

eddyb (May 03 2019 at 14:22, on Zulip):

especially so everyone can see the plan

pnkfelix (May 03 2019 at 14:22, on Zulip):

that makes as much sense as anything else, as long as the content at the links remains live long enough for it to be copied over

eddyb (May 03 2019 at 14:23, on Zulip):

this would partly replace the scattering of issues we create for various things medium/long-term plans, in rust-lang/rust

pnkfelix (May 03 2019 at 14:23, on Zulip):

Does anyone see a flaw in that approach? Seem lightweight enough

eddyb (May 03 2019 at 14:23, on Zulip):

(of which some of them cough promptly get forgotten cough)

pnkfelix (May 03 2019 at 14:23, on Zulip):

(of which some of them cough promptly get forgotten cough)

by design, right? ;)

pnkfelix (May 03 2019 at 14:24, on Zulip):

okay, well this sounds fine to me.

pnkfelix (May 03 2019 at 14:24, on Zulip):

I had originally thought that this meeting might represent week 1 of a four week cycle

pnkfelix (May 03 2019 at 14:24, on Zulip):

but I don't think that is fair

pnkfelix (May 03 2019 at 14:24, on Zulip):

in that there hasn't been enough time for anyone to prepare anything reasonable

pnkfelix (May 03 2019 at 14:24, on Zulip):

so maybe we can say that next week will be week 1 of our first four week cycle?

pnkfelix (May 03 2019 at 14:26, on Zulip):

we could spend the remainder of the time slot: 1. debating other aspects of steering-meeting.md, or 2. debating what topics for which people might consider writing up proposals, or 3. diving into whatever @eddyb was mentioning above, or 4. ending the meeting early ... ?

pnkfelix (May 03 2019 at 14:26, on Zulip):

(feel free to suggest other options in that enum)

eddyb (May 03 2019 at 14:27, on Zulip):

WOW, GitHub get on this level

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

did ... did eddyb just see something they liked about zulip ... ?

eddyb (May 03 2019 at 14:28, on Zulip):

Discord also allows it so it's not unique :P

eddyb (May 03 2019 at 14:28, on Zulip):

I just forgot unicode had emojis for numbers

pnkfelix (May 03 2019 at 14:29, on Zulip):

okay, lets maybe spend some time brainstorming about potential proposals that people might want to write up[

pnkfelix (May 03 2019 at 14:29, on Zulip):

and maybe interested parties can then collaborate, if they want, on a shared propoisal

pnkfelix (May 03 2019 at 14:30, on Zulip):

so lets see: there's the future of the rust AST and/or parser ... @eddyb has previously linked this doc https://hackmd.io/zrZhb94HS6KxW3sguwXNqA

eddyb (May 03 2019 at 14:31, on Zulip):

I have two lines of somewhat-related proposals:

eddyb (May 03 2019 at 14:31, on Zulip):

there are MIR things that could be proposed but I don't think I can get into a MIR (re)design any time soon

pnkfelix (May 03 2019 at 14:32, on Zulip):

As a non-tech proposal: i think @nikomatsakis and @pnkfelix would both like for us to revisit our prioritization/labelling scheme

pnkfelix (May 03 2019 at 14:32, on Zulip):

@pnkfelix would also like for us to at some point talk about how effort fixing bugs is actually distributed/assigned/tracked

pnkfelix (May 03 2019 at 14:33, on Zulip):

(the latter may well tie into the former; perhaps a unified proposal would be best there)

eddyb (May 03 2019 at 14:34, on Zulip):

we should maybe make it easier for people who are familiar with relevant parts of the implementation to give useful suggestions. "mentoring" doesn't feel ideal for this

eddyb (May 03 2019 at 14:35, on Zulip):

suggestions for either diagnosing or fixing a certain bug, that is

pnkfelix (May 03 2019 at 14:35, on Zulip):

ah you mean in terms of assigning work? yes.

pnkfelix (May 03 2019 at 14:36, on Zulip):

(in particular, writing up mentoring for fixing a bug has often appeared to me like it would require as much or more work as just fixing it)

eddyb (May 03 2019 at 14:36, on Zulip):

there is a sliding scale from being able to utter a couple nouns and them knowing how to go from there, all the way to linking to parts of the code with detailed explanations

eddyb (May 03 2019 at 14:37, on Zulip):

this is tricky even knowing who will actually work on it, let alone if you're doing it ahead of time

pnkfelix (May 03 2019 at 14:37, on Zulip):

actually, the interesting thing there about that sliding scale

pnkfelix (May 03 2019 at 14:37, on Zulip):

is that the people with the most knowledge may, via uttering those words, give enough hints to people in the middle that the person in a middle could then assist a beginner, pair-programming style.

pnkfelix (May 03 2019 at 14:38, on Zulip):

but that's just a hypothetical, and probably assumes both more people and more person-hours than we actually have available.

eddyb (May 03 2019 at 14:38, on Zulip):

oh noes are we creating NetBurst for mentoring?!

pnkfelix (May 03 2019 at 14:39, on Zulip):

Regarding other potential tech proposals: does anyone have any genius ideas for reducing memory pressure in rustc? I think at some point the idea was floated of simply restricting the number of codegen-units ...

eddyb (May 03 2019 at 14:40, on Zulip):

yes but you won't like it :P

eddyb (May 03 2019 at 14:40, on Zulip):

configurable limit for in-memory persistence of GC'd objects

eddyb (May 03 2019 at 14:41, on Zulip):

a more down-to-earth proposal would be file-backed arenas, that we are careful to avoid disk IO on if there is enough RAM

eddyb (May 03 2019 at 14:41, on Zulip):

this is how I see single-session whole-crate-DAG compilation playing out long-term

eddyb (May 03 2019 at 14:42, on Zulip):

and a similar threshold for disk usage, discarding the "least important to persist queries" first

eddyb (May 03 2019 at 14:43, on Zulip):

Cargo really needs to do similar GC to git, I think right now you can just end up with year-old versions of libraries just sitting around, with different hashes

pnkfelix (May 03 2019 at 14:43, on Zulip):

okay yes disk usage pressure is also an issue

pnkfelix (May 03 2019 at 14:43, on Zulip):

but its arguably a separate one

eddyb (May 03 2019 at 14:43, on Zulip):

so I don't have any idea better than "configurable age/size limits"

pnkfelix (May 03 2019 at 14:43, on Zulip):

since I think people are somewhat more willing/able to blow away build artifacts

pnkfelix (May 03 2019 at 14:44, on Zulip):

in response to disk usage pressure

eddyb (May 03 2019 at 14:44, on Zulip):

btw we can do "less relevant queries" relatively cheaply

pnkfelix (May 03 2019 at 14:44, on Zulip):

versus being basically unable to do anything, as users, in response to memory pressure.

eddyb (May 03 2019 at 14:44, on Zulip):

and maybe even track "usefulness" of queries without a huge perf hit

eddyb (May 03 2019 at 14:44, on Zulip):

I doubt we can implement a full adaptive caching algorithm, but we can at least approximate one

pnkfelix (May 03 2019 at 14:45, on Zulip):

okay. there might be interesting fruit here

eddyb (May 03 2019 at 14:45, on Zulip):

and queries that are not "properties of a definition", especially those that are cheap to recompute, we could throw away relatively easily, in response to memory pressure

eddyb (May 03 2019 at 14:45, on Zulip):

well, arena-allocating everything makes it hard to actually reduce the memory usage of doing so

pnkfelix (May 03 2019 at 14:45, on Zulip):

i'm a little worried that right now this meeting is playing out mostly like a conversation between @eddyb and @pnkfelix

pnkfelix (May 03 2019 at 14:46, on Zulip):

i mean, that's not the worst thing in the world

eddyb (May 03 2019 at 14:46, on Zulip):

but file-backing means you could tell the OS to unload those pages from memory

eddyb (May 03 2019 at 14:46, on Zulip):

actually, @Zoxc should be in here, yeah

pnkfelix (May 03 2019 at 14:46, on Zulip):

but I want to take a moment and see if anyone else has other topics they've had in the back of their mind that they might like to champion for a future steering meeting

pnkfelix (May 03 2019 at 14:46, on Zulip):

oh man

pnkfelix (May 03 2019 at 14:46, on Zulip):

did we not ping @T-compiler/meeting at the start ... ?

eddyb (May 03 2019 at 14:47, on Zulip):

ahahahaha

pnkfelix (May 03 2019 at 14:47, on Zulip):

/me scrolls

eddyb (May 03 2019 at 14:47, on Zulip):

yeah since we weren't sure what was happening, without @nikomatsakis here

pnkfelix (May 03 2019 at 14:47, on Zulip):

what a fail on my part

eddyb (May 03 2019 at 14:47, on Zulip):

@pnkfelix btw, if nothing else happens, and you have the time, I could walk you through the AST stuff (if you're curious)

eddyb (May 03 2019 at 14:47, on Zulip):

maybe write a better proposal based on that, etc.

pnkfelix (May 03 2019 at 14:48, on Zulip):

(I don't actually have the time; I have a hard limit to leave in 12 minutes)

qmx (May 03 2019 at 14:48, on Zulip):

it's fine, at least we can read the backlog

qmx (May 03 2019 at 14:48, on Zulip):

:)

pnkfelix (May 03 2019 at 14:48, on Zulip):

the good news is, since we aren't starting the four week cycle this week, but rather next week

pnkfelix (May 03 2019 at 14:48, on Zulip):

its not a disaster if a bunch of people missed this meeting

pnkfelix (May 03 2019 at 14:49, on Zulip):

just as long as we asynchronously do get some number of proposals up by next week

eddyb (May 03 2019 at 14:49, on Zulip):

fwiw I have a headache right now, so not having to explain my incomplete proposal would be fine :P

pnkfelix (May 03 2019 at 14:49, on Zulip):

and the other detail there is that writing up the proposals is itself a task

eddyb (May 03 2019 at 14:49, on Zulip):

I should maybe wrap up other things, so I can focus on it more soon

pnkfelix (May 03 2019 at 14:50, on Zulip):

(as well as even thinking of whatever is being proposed, of course)

pnkfelix (May 03 2019 at 14:50, on Zulip):

since I have a personal stake in the priorization/labelling/work-assignment non-tech issue

eddyb (May 03 2019 at 14:50, on Zulip):

wait I'm not sure I'll be able to attend next week, there's an overlap I forgot about

pnkfelix (May 03 2019 at 14:51, on Zulip):

its probably sensible that I champion (and potentially author) a proposal there.

eddyb (May 03 2019 at 14:51, on Zulip):

maybe we should let everyone else have a go first, and I'll try to be present from my phone

pnkfelix (May 03 2019 at 14:51, on Zulip):

maybe we should let everyone else have a go first, and I'll try to be present from my phone

by this, @eddyb , do you mean that you would not plan to have any proposals for next week's meeting?

nikomatsakis (May 03 2019 at 14:51, on Zulip):

Argh argh my apologies -- I was dragging away by something here that kept me AFK for far longer than I expected, and I faile to make it back in time for meeting.

nikomatsakis (May 03 2019 at 14:52, on Zulip):

I will however re-read the minutes

pnkfelix (May 03 2019 at 14:52, on Zulip):

and so if no one takes up the torch to author one (or champion something you authored), you're willing to wait four weeks for the next planning meeting?

Zoxc (May 03 2019 at 14:52, on Zulip):

I'd bring up my chain of PRs which makes progress on end to end queries - https://github.com/rust-lang/rust/pull/59904

Wesley Wiser (May 03 2019 at 14:53, on Zulip):

Maybe it would be a good idea to try to have any proposals available by Thursday? That way people have a chance to read the proposal and prepare questions rather than spend meeting time reading the proposal.

pnkfelix (May 03 2019 at 14:53, on Zulip):

@Zoxc great, .that ... definitely is a linked list implemented via github PR's.

pnkfelix (May 03 2019 at 14:53, on Zulip):

I do think it would be good for proposals to be ready by thursday

eddyb (May 03 2019 at 14:53, on Zulip):

@Zoxc oh right! that's great and I could build on that for my own proposals :D

pnkfelix (May 03 2019 at 14:53, on Zulip):

I don't know whether it will be reasonable to make it a hard requirement that they be posted by thursday

pnkfelix (May 03 2019 at 14:53, on Zulip):

especially since, as noted above

pnkfelix (May 03 2019 at 14:53, on Zulip):

we don't have a common indexing method

eddyb (May 03 2019 at 14:54, on Zulip):

since some of them are mostly "what deep changes do we need to make to push this even futher"

pnkfelix (May 03 2019 at 14:54, on Zulip):

apart from the meeting itself

nikomatsakis (May 03 2019 at 14:54, on Zulip):

(I think we should not have the meeting be the way)

pnkfelix (May 03 2019 at 14:54, on Zulip):

@nikomatsakis but you also are not opposed to other proposal formats

nikomatsakis (May 03 2019 at 14:54, on Zulip):

I'd rather have opening issues and tagging with a label, or adding to a central doc, or something

nikomatsakis (May 03 2019 at 14:54, on Zulip):

Yeah, I agree that a file in repo is maybe too heavy weight

pnkfelix (May 03 2019 at 14:54, on Zulip):

/me tried to say this at the top

nikomatsakis (May 03 2019 at 14:54, on Zulip):

I just think there's value in having a place people can go beforehand to read up

pnkfelix (May 03 2019 at 14:55, on Zulip):

(that we needed a central index of some form, even if it is just the issue/PR list)

nikomatsakis (May 03 2019 at 14:55, on Zulip):

and also persisting from week to week

nikomatsakis (May 03 2019 at 14:55, on Zulip):

My hope is that these proposals can "accumulate" during the 4 weeks

nikomatsakis (May 03 2019 at 14:55, on Zulip):

(sorry for jumping in the middle!!)

nikomatsakis (May 03 2019 at 14:55, on Zulip):

Why not just have issues on compiler-team repo tagged with a label and a link to a hackmd/gist/whatever document?

nikomatsakis (May 03 2019 at 14:55, on Zulip):

Seems pretty light-weight

pnkfelix (May 03 2019 at 14:56, on Zulip):

its no problem as long as everyone's aware of it and follows through

nikomatsakis (May 03 2019 at 14:56, on Zulip):

(I wasn't wild about issues because I didn't want a bunch of commentary to accumulate there but we could try it)

pnkfelix (May 03 2019 at 14:56, on Zulip):

maybe we can as a rule, lock the issues?

pnkfelix (May 03 2019 at 14:56, on Zulip):

(or is that too heavy handed?)

nikomatsakis (May 03 2019 at 14:56, on Zulip):

maybe we can as a rule, lock the issues?

I was going to suggest it :)

eddyb (May 03 2019 at 14:56, on Zulip):

heh that is kinda neat?

nikomatsakis (May 03 2019 at 14:57, on Zulip):

I like to lock + link to a zulip thread

pnkfelix (May 03 2019 at 14:57, on Zulip):

basically we don't want a repeat of the overwhelming RFC thread issues

nikomatsakis (May 03 2019 at 14:57, on Zulip):

but it doesn't matter really ;)

eddyb (May 03 2019 at 14:57, on Zulip):

or maybe have a single issue or "GitHub Project" to track them

nikomatsakis (May 03 2019 at 14:57, on Zulip):

Let's try issues to start, I'll make a label like "Meeting Proposal"

eddyb (May 03 2019 at 14:57, on Zulip):

actually, what does that stuff look like now? the "kanban" or w/e style thing

pnkfelix (May 03 2019 at 14:57, on Zulip):

/me goes to google yet more terminology

eddyb (May 03 2019 at 14:58, on Zulip):

https://github.com/rust-lang/compiler-team/projects

nikomatsakis (May 03 2019 at 14:58, on Zulip):

we can add a github project on top of that? I'm not quite sure what the columns etc would be, but that could be useful

eddyb (May 03 2019 at 14:58, on Zulip):

I mean, if we're just doing links

eddyb (May 03 2019 at 14:58, on Zulip):

this could be a neat place to store them

nikomatsakis (May 03 2019 at 14:58, on Zulip):

can anybody add cards?

pnkfelix (May 03 2019 at 14:58, on Zulip):

I object to having the links be the permanent record

eddyb (May 03 2019 at 14:58, on Zulip):

I think it's similar to creating an issue

pnkfelix (May 03 2019 at 14:58, on Zulip):

its fine for initial draft

eddyb (May 03 2019 at 14:58, on Zulip):

@pnkfelix this is assuming that once something is accepted we merge it into the repo

eddyb (May 03 2019 at 14:58, on Zulip):

like, full text

nikomatsakis (May 03 2019 at 14:59, on Zulip):

I was imagining that the issue comments would have a purpose, just noit general disussion

pnkfelix (May 03 2019 at 14:59, on Zulip):

but at the latest, post discussion, it shoudl be transcribed. Yes, okay, that.

nikomatsakis (May 03 2019 at 14:59, on Zulip):

more like a blog: minutes/updates

eddyb (May 03 2019 at 14:59, on Zulip):

oh that is fair

nikomatsakis (May 03 2019 at 14:59, on Zulip):

one thing I can imagine for example is

nikomatsakis (May 03 2019 at 14:59, on Zulip):

that in a planning meeting

eddyb (May 03 2019 at 14:59, on Zulip):

so then we can just use locked issues

nikomatsakis (May 03 2019 at 14:59, on Zulip):

we might say something like "We like this ide but can you add X, Y , and Z to the proposal"

nikomatsakis (May 03 2019 at 14:59, on Zulip):

(and that could be a useful thing to use comments for)

pnkfelix (May 03 2019 at 14:59, on Zulip):

/me is going to hand off driving the meeting to @nikomatsakis

nikomatsakis (May 03 2019 at 14:59, on Zulip):

sorry, I'm feeling terrible about being late :embarrassed:

pnkfelix (May 03 2019 at 14:59, on Zulip):

(in case people decide to continue after the end time)

eddyb (May 03 2019 at 14:59, on Zulip):

we can use hackmd/gist comments for transient discussion

nikomatsakis (May 03 2019 at 15:00, on Zulip):

seems fine

nikomatsakis (May 03 2019 at 15:00, on Zulip):

ok, so how about this:

nikomatsakis (May 03 2019 at 15:00, on Zulip):

I'll edit the proposal to make a simple issue template

nikomatsakis (May 03 2019 at 15:00, on Zulip):

(and add it to the repo)

nikomatsakis (May 03 2019 at 15:00, on Zulip):

and we'll lock issues that are meeting proposals (and tag them)

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

(actually I have to go to another meeting too)

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

but @eddyb and others, open issues for next time and we;l figure out the details later?

eddyb (May 03 2019 at 15:21, on Zulip):

@nikomatsakis oops I forgot to make sure you saw this: I likely won't be able to present anything next week due to a scheduling overlap I missed because I expected more will happen this week

nikomatsakis (May 03 2019 at 15:24, on Zulip):

OK, well let's discuss.

nikomatsakis (May 03 2019 at 15:26, on Zulip):

You don't necessarily have to be present, and maybe we'll want to do the first meeting later, or whatever

nikomatsakis (May 03 2019 at 15:26, on Zulip):

I'll try to write-up a quick gh template in the meantime, with an aim to be lightweight

eddyb (May 03 2019 at 15:40, on Zulip):

if I propose anything for next week, it will be reverting https://github.com/rust-lang/rust/pull/58061

eddyb (May 03 2019 at 15:41, on Zulip):

since I think I can argue for that without an extensive look into what we might want to do in the future

eddyb (May 03 2019 at 15:41, on Zulip):

it's mostly "mutation is a dead-end and it ties our hands in the name of performance"

nikomatsakis (May 03 2019 at 17:54, on Zulip):

@T-compiler/meeting It occurs to me that we should make a summary of today's meeting -- anybody up to give that a try?

nikomatsakis (May 03 2019 at 17:54, on Zulip):

Also, I've created https://github.com/rust-lang/compiler-team/pull/79 which I think matches what we said.

Last update: Nov 22 2019 at 05:45UTC