Stream: t-compiler

Topic: planning meeting 2019.01.17


nikomatsakis (Jan 17 2020 at 14:55, on Zulip):

Dear @T-compiler/meeting -- planning meeting starts in 5 minutes!

nikomatsakis (Jan 17 2020 at 15:00, on Zulip):

Dear @T-compiler/meeting -- planning meeting!

nikomatsakis (Jan 17 2020 at 15:00, on Zulip):

Our goal is to pick design meetings for the next 3 weeks. However, there are some constraints.

nikomatsakis (Jan 17 2020 at 15:01, on Zulip):
nikomatsakis (Jan 17 2020 at 15:02, on Zulip):

Meeting proposals

nikomatsakis (Jan 17 2020 at 15:02, on Zulip):

/poll What might be good candidates to schedule

simulacrum (Jan 17 2020 at 15:02, on Zulip):

Jan 30th is also a Rust release fwiw

pnkfelix (Jan 17 2020 at 15:03, on Zulip):

its so amazing how we've somehow managed to have releases coincide with Mozilla or Rust all-hands

centril (Jan 17 2020 at 15:03, on Zulip):

I think it's premature to talk parser

centril (Jan 17 2020 at 15:04, on Zulip):

I'm not ready for that discussion, and things like https://github.com/rust-lang/rust/issues/64197 should happen first

nikomatsakis (Jan 17 2020 at 15:04, on Zulip):

I'm not sure about chalk-ty. I'm hoping to put more work into that today. I guess it's not premature necessarily but I would have to do some prep work =)

pnkfelix (Jan 17 2020 at 15:04, on Zulip):

Regarding parser: its premature ... because more experimentation is needed to determine an appropriate interface for it to have (either in terms of exposing to its clients, or the one it uses from rustc) ?

nikomatsakis (Jan 17 2020 at 15:05, on Zulip):

I feel like it's also ok for us to have "early design discussion" sometimes, but I don't know if the design meeting slot is necessarily best place for it? I'm mixed

centril (Jan 17 2020 at 15:05, on Zulip):

@pnkfelix I don't deeply understand the RA parser yet

centril (Jan 17 2020 at 15:05, on Zulip):

and its AST

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

Is there something to discuss for side effect? Maybe we don't need a design meeting there? I feel like we decided we didn't in the past...

nikomatsakis (Jan 17 2020 at 15:06, on Zulip):

One thing I would note: on the merge.llvm-side-effect -- I see no need for a meeting. @mw did measurements on Firefox and found no impact (as expected). I think we can just go for it.

pnkfelix (Jan 17 2020 at 15:06, on Zulip):

@centril and you don't think scheduling the meeting would provide impetus to research more?

centril (Jan 17 2020 at 15:06, on Zulip):

I would be happy if we decide to just r+

nikomatsakis (Jan 17 2020 at 15:06, on Zulip):

We might want to prep an announcement or something

centril (Jan 17 2020 at 15:06, on Zulip):

@pnkfelix at some point yes, but not yet

simulacrum (Jan 17 2020 at 15:06, on Zulip):

I can sign up to write that announcement

centril (Jan 17 2020 at 15:07, on Zulip):

things like https://github.com/rust-lang/rust/issues/64197 are important things that would need to be fixed first at any rate

pnkfelix (Jan 17 2020 at 15:07, on Zulip):

I'm just trying to understand whether you are saying that a meeting can/should be blocked on all participants wanting to first dive in before even scheduling the meeting itself.

pnkfelix (Jan 17 2020 at 15:07, on Zulip):

(a design meeting, note)

pnkfelix (Jan 17 2020 at 15:07, on Zulip):

because, honestly, that kind of road block does not make sense to me.

centril (Jan 17 2020 at 15:07, on Zulip):

@pnkfelix ideally the main maintainers of a particular part of the rustc infra should be ready for it :slight_smile:

nikomatsakis (Jan 17 2020 at 15:08, on Zulip):

well, what is the goal of the meeting? I would think the goal is to try and establish the direction.

nikomatsakis (Jan 17 2020 at 15:08, on Zulip):

so I guess the question is whether doing #64197 has to be done first before we can talk about other questions

centril (Jan 17 2020 at 15:08, on Zulip):

yes, and to assess the direction you need some readiness

nikomatsakis (Jan 17 2020 at 15:09, on Zulip):

I seem to recall that @matklad had done a fair amount of prep there, I'm going to skim for a sec

centril (Jan 17 2020 at 15:09, on Zulip):

re-engineering the parser & AST is a massive change

centril (Jan 17 2020 at 15:09, on Zulip):

since it has implications for lowering and macros as well

nikomatsakis (Jan 17 2020 at 15:11, on Zulip):

I don't disagree with that

centril (Jan 17 2020 at 15:12, on Zulip):

Also, e.g.

For this reason, I propose to focus on extracting only the parser itself, and continuing using different syntax tree data structures in rust-analyzer and in rustc.

I've already extracted the parser

nikomatsakis (Jan 17 2020 at 15:12, on Zulip):

I'm just not sure that it means we can't talk about the things in the issue :)

centril (Jan 17 2020 at 15:12, on Zulip):

it is in the crate rustc_parse

nikomatsakis (Jan 17 2020 at 15:12, on Zulip):

Does it support multiple forms of ASTs?

centril (Jan 17 2020 at 15:13, on Zulip):

no it works on the one AST, and for now it also entangled with conditional compilation (due to the issue aforementioned)

centril (Jan 17 2020 at 15:14, on Zulip):

Note that the parser does inline mutation of the AST in places for better diagnostics, which is presumably hard to square with this new design

nikomatsakis (Jan 17 2020 at 15:14, on Zulip):

This is I think precisely the sort of thing that would be good to discuss in a meeting

centril (Jan 17 2020 at 15:16, on Zulip):

I don't disagree; all I'm saying is that I think it could wait some more time, and to e.g. extract the conditional compilation part so that we know we can make the parser more pure

nikomatsakis (Jan 17 2020 at 15:16, on Zulip):

I guess it seems to me that we can "look a bit ahead", and predict what sorts of problems we might encounter if we tried to do this, and whether we should look further into it or pick another route altogether

nikomatsakis (Jan 17 2020 at 15:17, on Zulip):

I see that the issue describes a few things to discuss:

nikomatsakis (Jan 17 2020 at 15:17, on Zulip):
nikomatsakis (Jan 17 2020 at 15:17, on Zulip):

this might be more than we can reasonably do regardless :)

pnkfelix (Jan 17 2020 at 15:17, on Zulip):

I don't disagree; all I'm saying is that I think it could wait some more time, and to e.g. extract the conditional compilation part so that we know we can make the parser more pure

While I can understand that doing upfront implementation work can inform subsequent design work ...

pnkfelix (Jan 17 2020 at 15:17, on Zulip):

... I still think its odd to block design work on the implementation work.

nikomatsakis (Jan 17 2020 at 15:18, on Zulip):

Yes, it feels to me like it's a cycle. You think about where you want to go, you try to get there, you hit problems, you come back to the design. =)

nikomatsakis (Jan 17 2020 at 15:18, on Zulip):

But also i'm not sure I 100% see the connection

nikomatsakis (Jan 17 2020 at 15:18, on Zulip):

by this I mean that many of the details here seem different (to clarify: there are details, like what AST maybe should look like, that seems pretty separate from how we drive parser and handle cfg)

pnkfelix (Jan 17 2020 at 15:18, on Zulip):

but I am definitely willing to let the people most actively maintaining the code in question be the ones who have heaviest impact on scheduling of design work.

pnkfelix (Jan 17 2020 at 15:19, on Zulip):

I guess I'd like to hear from @matklad about the pro's/con's of trying to do this design meeting sooner vs waiting a bit

nikomatsakis (Jan 17 2020 at 15:19, on Zulip):

yes, that would be helpful

matklad (Jan 17 2020 at 15:19, on Zulip):

I think the only pre-requsitie for the design meeting is that relevant folks should be aware of both rustc and rust-analyzer parser designs

centril (Jan 17 2020 at 15:20, on Zulip):

(Admittedly I'm skeptical of this new design and fond of the current one, which I think serves rustc & e.g. clippy better than the new one, which is partly why I want this to take more time to fully assess the implications)

nikomatsakis (Jan 17 2020 at 15:21, on Zulip):

I feel like that's all the more reason to talk about it ..

matklad (Jan 17 2020 at 15:21, on Zulip):

It's also important to note that parser library-ification is the next blocker on merging rust-analyzer and rustc, so it also makes sense to schedule the implementaion work earlier, and, by extention, to schedule design work earlier as well

nikomatsakis (Jan 17 2020 at 15:21, on Zulip):

Let me ask this

nikomatsakis (Jan 17 2020 at 15:22, on Zulip):

If I understand, the refactoring @centril that you're referring to is kind of changing how cfg is handled, so that name resolution drives parsing -- and in particular parsing of submodule files

pnkfelix (Jan 17 2020 at 15:22, on Zulip):

honest Q because I am ignorant: of the so-called "relevant people" to the parser work, which ones will be at Mozilla all-hands?

centril (Jan 17 2020 at 15:22, on Zulip):

no one, I think

matklad (Jan 17 2020 at 15:23, on Zulip):

(deleted)

matklad (Jan 17 2020 at 15:23, on Zulip):

oh, mozzila, not rust

nikomatsakis (Jan 17 2020 at 15:23, on Zulip):

@matklad is at least in Berlin :)

nikomatsakis (Jan 17 2020 at 15:23, on Zulip):

but I'm not sure that anyone else is

centril (Jan 17 2020 at 15:23, on Zulip):

"relevant people" is probably matklad, petrochenkov, estebank, and myself

nikomatsakis (Jan 17 2020 at 15:23, on Zulip):

but it seems like most of what @matklad is talking about here

centril (Jan 17 2020 at 15:23, on Zulip):

also oli for lowering?

nikomatsakis (Jan 17 2020 at 15:23, on Zulip):

has to do with parsing of a single file

nikomatsakis (Jan 17 2020 at 15:23, on Zulip):

and representing the AST

nikomatsakis (Jan 17 2020 at 15:24, on Zulip):

so I'm not sure that I entirely see why one blocks the other, and if it does, I feel like that alone is maybe a fruitful focus for discussion

centril (Jan 17 2020 at 15:24, on Zulip):

If I understand, the refactoring @centril that you're referring to is kind of changing how cfg is handled, so that name resolution drives parsing -- and in particular parsing of submodule files

yep, that's right

nikomatsakis (Jan 17 2020 at 15:25, on Zulip):

In any case, I guess I'm pushing back because I want to see us making progress on library-ification goals. I feel like we need more talking early on so that we can be converging. If it's not a design meeting, then, I'd sort of like to know what the mechanism is that is going to help us converge

centril (Jan 17 2020 at 15:25, on Zulip):

currently upon mod foo;, the parser does some inappropriate things: it does conditional compilation, if it is in the configuration, it opens opens the submodule file, and does parsing

nikomatsakis (Jan 17 2020 at 15:25, on Zulip):

(Also, I'm totally open to other mechanisms :)

centril (Jan 17 2020 at 15:25, on Zulip):

@nikomatsakis I'm not arguing for another mechanism

centril (Jan 17 2020 at 15:26, on Zulip):

I'm just arguing for "wait more weeks" :P

nikomatsakis (Jan 17 2020 at 15:26, on Zulip):

@matklad do you feel like there are things to talk about that are independent of how we handle mod foo;?

centril (Jan 17 2020 at 15:26, on Zulip):

(not a lot of more weeks)

matklad (Jan 17 2020 at 15:27, on Zulip):

Yes

nikomatsakis (Jan 17 2020 at 15:27, on Zulip):

so maybe we should focus on those

matklad (Jan 17 2020 at 15:27, on Zulip):

More specifically, I feel everyone agrees that the handling of mod foo; should be changed from the current one, so this is a question on which we have full consensus

centril (Jan 17 2020 at 15:28, on Zulip):

definitely

centril (Jan 17 2020 at 15:28, on Zulip):

(in fact I'm sorta experimenting with fixing it)

centril (Jan 17 2020 at 15:29, on Zulip):

(though it's a bit difficult)

nikomatsakis (Jan 17 2020 at 15:30, on Zulip):

(Is @Esteban Küber here by any chance, or @Vadim Petrochenkov?)

pnkfelix (Jan 17 2020 at 15:30, on Zulip):

so I'm still trying to understand: @centril doesn't want to schedule a parser meeting in the current cycle. I.e. even Feb 7 would be too soon. Right? So what happens if we get to Feb 14 and @centril is in same position w.r.t. their (dis)comfort with knowledge of RA etc ?

nikomatsakis (Jan 17 2020 at 15:30, on Zulip):

anyway maybe we can pick a more narrow topic to discuss, such as digging deeper into how we are going to bridge the rust-analyzer ASTs (presuming we do want to do that...)

nikomatsakis (Jan 17 2020 at 15:30, on Zulip):

and/or I think we could just talk about how rust-analyzer works

nikomatsakis (Jan 17 2020 at 15:30, on Zulip):

and just try to learn something :)

pnkfelix (Jan 17 2020 at 15:30, on Zulip):

to me, Feb 7 seems a ways off. But of course its not, not really.

nikomatsakis (Jan 17 2020 at 15:31, on Zulip):

it'll be here soon enough :)

pnkfelix (Jan 17 2020 at 15:32, on Zulip):

would a meeting to discuss how the RA parser works, would even that be premature, @centril ?

matklad (Jan 17 2020 at 15:32, on Zulip):

Note that we have a recent writeup about this: https://rust-lang.zulipchat.com/#narrow/stream/131828-t-compiler/topic/planning.20meeting.202019.2E01.2E17/near/185924217

centril (Jan 17 2020 at 15:32, on Zulip):

I'd be up for a "lecture meeting"

nikomatsakis (Jan 17 2020 at 15:33, on Zulip):

I think the goal should be that active participants read the write-up first

nikomatsakis (Jan 17 2020 at 15:33, on Zulip):

so it's not just like copying and pasting text into the chat ..

nikomatsakis (Jan 17 2020 at 15:34, on Zulip):

the other two topics were

nikomatsakis (Jan 17 2020 at 15:34, on Zulip):

I definitely have thoughts about chalk-ty that I wouldn't mind floating, I've been rethinking in particular some of my plans and wouldn't mind getting feedback, though I've not had time to write them down (on my list of "maybe today" items)

nikomatsakis (Jan 17 2020 at 15:34, on Zulip):

but I also can get that down and put it out there async

nikomatsakis (Jan 17 2020 at 15:34, on Zulip):

I'm asking because if we want to schedule to give time for folks to read, we might consider Feb 7

nikomatsakis (Jan 17 2020 at 15:35, on Zulip):

though really how much time does it take, I don't know. it's mostly a problem of making the time

centril (Jan 17 2020 at 15:35, on Zulip):

Feb 7 seems good; it's an important change and giving folks time to read a lot is a good idea

nikomatsakis (Jan 17 2020 at 15:35, on Zulip):

this also gives as much time as possible to continue refactoring work :)

nikomatsakis (Jan 17 2020 at 15:36, on Zulip):

@pnkfelix do we feel we have something to talk about for maintenance and triage? I still feel like we need "work" there, but I'm also not full of ideas, beyond the "cleanup crew" idea ;)

nikomatsakis (Jan 17 2020 at 15:36, on Zulip):

but I think part of the goal of a meeting can be to "retro and brainstorm"

pnkfelix (Jan 17 2020 at 15:37, on Zulip):

Yeah, I don't know.

pnkfelix (Jan 17 2020 at 15:37, on Zulip):

I might need to go back and try to write up another dfraft of the propsoal

centril (Jan 17 2020 at 15:37, on Zulip):

maybe it would be good to discuss the progress on #65031 soon btw

pnkfelix (Jan 17 2020 at 15:37, on Zulip):

maybe I will just close that Proposal PR

centril (Jan 17 2020 at 15:37, on Zulip):

we've come a long way

nikomatsakis (Jan 17 2020 at 15:38, on Zulip):

maybe I will just close that Proposal PR

I feel ok about that

nikomatsakis (Jan 17 2020 at 15:38, on Zulip):

maybe it would be good to discuss the progress on #65031 soon btw

interesting

nikomatsakis (Jan 17 2020 at 15:38, on Zulip):

what would you want to talk about there exactly

centril (Jan 17 2020 at 15:39, on Zulip):

we could talk about the progress (also as announcement), and e.g. what role we envisage for the rustc crate, if there is going to be one

centril (Jan 17 2020 at 15:40, on Zulip):

but maybe that doesn't need a design meeting, not sure

nikomatsakis (Jan 17 2020 at 15:40, on Zulip):

PS, seems like we have agreement to schedule parser meeting for Feb 7, with a primary goal of talking through some of the rust-analyzer AST design notes?

nikomatsakis (Jan 17 2020 at 15:40, on Zulip):

@matklad :point_up: ?

nikomatsakis (Jan 17 2020 at 15:40, on Zulip):

(That work for you?)

centril (Jan 17 2020 at 15:40, on Zulip):

I thought we wanted to talk type system on 7th?

simulacrum (Jan 17 2020 at 15:40, on Zulip):

I don't think a design meeting is warranted on rustc crate quite yet (seems like an odd topic)

nikomatsakis (Jan 17 2020 at 15:40, on Zulip):

/me confused :)

nikomatsakis (Jan 17 2020 at 15:41, on Zulip):

I'm asking because if we want to schedule to give time for folks to read, we might consider Feb 7

ah I guess this was ambiguous

centril (Jan 17 2020 at 15:41, on Zulip):

yes :slight_smile:

nikomatsakis (Jan 17 2020 at 15:41, on Zulip):

I meant consider the parser for Feb 7

nikomatsakis (Jan 17 2020 at 15:42, on Zulip):

let me ask this. We could do

as the two meetings

nikomatsakis (Jan 17 2020 at 15:42, on Zulip):

I don't really care which order :)

centril (Jan 17 2020 at 15:42, on Zulip):

aha; -- an information meeting about the parser on the 7th seems ok

centril (Jan 17 2020 at 15:42, on Zulip):

though we could do maintenance & triage + chalk/ty in the two slots?

nikomatsakis (Jan 17 2020 at 15:43, on Zulip):

I think we decided against maintenance and triage

nikomatsakis (Jan 17 2020 at 15:43, on Zulip):

since we don't know how to drive it

centril (Jan 17 2020 at 15:43, on Zulip):

oh

centril (Jan 17 2020 at 15:43, on Zulip):

ok

matklad (Jan 17 2020 at 15:43, on Zulip):

Parser info on 7th sounds good to me

pnkfelix (Jan 17 2020 at 15:43, on Zulip):

I don't think a design meeting is warranted on rustc crate quite yet (seems like an odd topic)

(we can float some initial discussion of rustc crate at triage meeting. and then see, based on that, whether to try to have follow-up design meeting)

nikomatsakis (Jan 17 2020 at 15:43, on Zulip):

shall we do that? I can up the priority on preparing a chalk-ty write-up then

matklad (Jan 17 2020 at 15:43, on Zulip):

In general, I am willing to teach anybody about parser stuff at any time, just reach out to me :D

pnkfelix (Jan 17 2020 at 15:44, on Zulip):

I think we decided against maintenance and triage

I filed compiler-team#239 for this

nikomatsakis (Jan 17 2020 at 15:46, on Zulip):

I am feeling a nagging doubt. I feel like we've covered rust-analyzer's parser architecture a few times -- at all hands we had a great video, etc. I guess I think talking about the realities of creating a streaming architecture, and whether we could try to take steps toward it, might be a better use of sync discussion time

nikomatsakis (Jan 17 2020 at 15:47, on Zulip):

And I still don't really see what's blocking that.

centril (Jan 17 2020 at 15:47, on Zulip):

I could add some inline notes to @matklads design doc re. the RA parser also

nikomatsakis (Jan 17 2020 at 15:47, on Zulip):

that would be good for sure

nikomatsakis (Jan 17 2020 at 15:47, on Zulip):

I think as much engagement in leadup as possible is a win

nikomatsakis (Jan 17 2020 at 15:49, on Zulip):

so I propose this:

The only bit there is that I have to do more prep around chalk-ty, so maybe it makes more sense to swap the order, since parser prep is done.

matklad (Jan 17 2020 at 15:50, on Zulip):

I guess I think talking about the realities of creating a streaming architecture

@nikomatsakis could you unpack this? I am not sure I undestand what you are saying

nikomatsakis (Jan 17 2020 at 15:50, on Zulip):

Sorry, what I'm saying is

centril (Jan 17 2020 at 15:50, on Zulip):

as long as we're not making decisions re. the parser I'd be fine with that

nikomatsakis (Jan 17 2020 at 15:50, on Zulip):

of the two things you talked about

nikomatsakis (Jan 17 2020 at 15:50, on Zulip):
nikomatsakis (Jan 17 2020 at 15:50, on Zulip):

I might rather focus more on the second

nikomatsakis (Jan 17 2020 at 15:50, on Zulip):

but really I'd defer to you on this point

nikomatsakis (Jan 17 2020 at 15:51, on Zulip):

which ever you prefer feedback on

matklad (Jan 17 2020 at 15:51, on Zulip):

I definitelly would love to focus on the second one, provided that folks do the homework for the first one :)

nikomatsakis (Jan 17 2020 at 15:52, on Zulip):

So that's what I think we should do.

centril (Jan 17 2020 at 15:52, on Zulip):

I don't agree; one week is not enough to do "the homework"

nikomatsakis (Jan 17 2020 at 15:53, on Zulip):

OK, then let's do chalk-ty next week, Feb 7 parser.

nikomatsakis (Jan 17 2020 at 15:53, on Zulip):

I wonder if @Jack Huey or @detrumi can attend this slot

matklad (Jan 17 2020 at 15:53, on Zulip):

I don't believe we are ready to make any decisions about end-game parser architecture, but we already should have at least a possible map of that state, and some possible roads towards it.

nikomatsakis (Jan 17 2020 at 15:54, on Zulip):

Yes. I'm not so worried about using meetings to make decisions

nikomatsakis (Jan 17 2020 at 15:54, on Zulip):

That can happen but I think it's best if we're using them more to develop the design and uncover problems and constraints

nikomatsakis (Jan 17 2020 at 15:54, on Zulip):

as well as to explain

centril (Jan 17 2020 at 15:54, on Zulip):

implementation strategy & roads isn't too hard to see right now

pnkfelix (Jan 17 2020 at 15:55, on Zulip):

I personally am happy with either date. But I am not happy with the prospect that @nikomatsakis is going to have to do overly quick writeup on chalk-ty in order to accommodate other people not being able to review design docs that have already been written.

pnkfelix (Jan 17 2020 at 15:56, on Zulip):

So that's my official feedback on @centril 's influence on the scheduling here.

centril (Jan 17 2020 at 15:56, on Zulip):

@pnkfelix it's not just reviewing a design document; I can do that fairly quickly; reading the RA parser code is also necessary, imo

pnkfelix (Jan 17 2020 at 15:57, on Zulip):

@matklad has said they are happy to engage. Seems like maybe you two could have private session to answer your Q's?

pnkfelix (Jan 17 2020 at 15:57, on Zulip):

rather than you having to go through code independently?

centril (Jan 17 2020 at 15:57, on Zulip):

sure; well, I'd also like to read though parts of the code

nikomatsakis (Jan 17 2020 at 15:58, on Zulip):

/me suggests fixing a small bug :)

nikomatsakis (Jan 17 2020 at 15:58, on Zulip):

not sure if any of those still exist

pnkfelix (Jan 17 2020 at 15:59, on Zulip):

"Bugs? Those don't happen in Rust"

Jack Huey (Jan 17 2020 at 15:59, on Zulip):

I'll be around Feb. 7!

Jack Huey (Jan 17 2020 at 15:59, on Zulip):

Oh, also next week

nikomatsakis (Jan 17 2020 at 15:59, on Zulip):

ok, good to know

pnkfelix (Jan 17 2020 at 16:00, on Zulip):

okay, well, lets be clear

matklad (Jan 17 2020 at 16:00, on Zulip):

Might be a good idea to ask about @Vadim Petrochenkov and @Esteban Küber availability as well, if they can't next week, but can on 7th, that's an argument for a swap

nikomatsakis (Jan 17 2020 at 16:00, on Zulip):

I agree with that

nikomatsakis (Jan 17 2020 at 16:00, on Zulip):

though it looks to me like we are tentatively doing

but we'll potentially swap based on @Vadim Petrochenkov and @Esteban Küber's availability

pnkfelix (Jan 17 2020 at 16:02, on Zulip):

Okay that's the clarity I wanted.

pnkfelix (Jan 17 2020 at 16:02, on Zulip):

we'll maybe plan to finalize the date by end of day Monday?

centril (Jan 17 2020 at 16:02, on Zulip):

and "parser" still involves discussing how RA does things?

pnkfelix (Jan 17 2020 at 16:04, on Zulip):

I think the intent was to try to encourage people to do more upfront work of the first bullet so that the meeting could focus on the second bullet

pnkfelix (Jan 17 2020 at 16:04, on Zulip):

but @matklad is going to make the decision in the end about what is best to focus on, I think ?

nikomatsakis (Jan 17 2020 at 16:05, on Zulip):

right, the real hope is that we can start to discuss ways to share more code between r-a and rustc. honestly I'm not sure how much the concrete details of r-a's AST matter for that

nikomatsakis (Jan 17 2020 at 16:05, on Zulip):

though I think it's good to take a look!

centril (Jan 17 2020 at 16:05, on Zulip):

I feel like you're switching back & forth from the 7th to the other date and it's confusing

nikomatsakis (Jan 17 2020 at 16:06, on Zulip):

We did switch back and forth a few times, yes. @pnkfelix raised the point that it really makes sense to try and move faster on rust-analyzer parser discussions since the write-ups are more prepared.

centril (Jan 17 2020 at 16:06, on Zulip):

Yes, and I responded to said point

pnkfelix (Jan 17 2020 at 16:07, on Zulip):

I have to go, but my silence in response to your response was not agreement with it.

pnkfelix (Jan 17 2020 at 16:07, on Zulip):

(and actually, I wasn't silent)

pnkfelix (Jan 17 2020 at 16:08, on Zulip):

Would we be better off not having a meeting next week???

nikomatsakis (Jan 17 2020 at 16:09, on Zulip):

Maybe we should just do that. There's not really time to prep the chalk-ty stuff, centril would like time to review, let's run with Feb 7 for parser.

nikomatsakis (Jan 17 2020 at 16:09, on Zulip):

I'll work on chalk-ty stuff regardless today

nikomatsakis (Jan 17 2020 at 16:09, on Zulip):

this meeting has to end :)

pnkfelix (Jan 17 2020 at 16:09, on Zulip):

well, that, or check with availability of @Vadim Petrochenkov and @Esteban Küber

pnkfelix (Jan 17 2020 at 16:09, on Zulip):

and let that dictate

nikomatsakis (Jan 17 2020 at 16:09, on Zulip):

right

pnkfelix (Jan 17 2020 at 16:10, on Zulip):

for all we know, they might be available next week and not available on Feb 7th

nikomatsakis (Jan 17 2020 at 16:12, on Zulip):

OK, I'm sending an e-mail since they may find all the pings confusing :)

pnkfelix (Jan 17 2020 at 16:13, on Zulip):

bye everyone in @T-compiler/meeting . Thanks for attending.

nagisa (Jan 17 2020 at 16:14, on Zulip):

How come I always get notification sound for "thanks for attending" only T_T

nikomatsakis (Jan 17 2020 at 16:14, on Zulip):

huh, that...is weird

Last update: Jun 07 2020 at 10:45UTC