Stream: t-compiler/wg-nll

Topic: weekly meeting 2019.02.20


pnkfelix (Feb 19 2019 at 13:21, on Zulip):

hi @T-compiler/meeting @WG-compiler-nll ; our meeting is tomorrow (not today), but if anyone has anything to report today, it would be great if you could put it on the existing paper doc

oli (Feb 19 2019 at 13:21, on Zulip):

can you link the doc here?

davidtwco (Feb 19 2019 at 13:22, on Zulip):

triage document

lqd (Feb 19 2019 at 13:24, on Zulip):

did you want to notify wg-nll ?

pnkfelix (Feb 19 2019 at 13:25, on Zulip):

omg

pnkfelix (Feb 19 2019 at 13:25, on Zulip):

thanks @lqd

pnkfelix (Feb 19 2019 at 13:25, on Zulip):

Sorry all, the above note was meant for @WG-compiler-nll

pnkfelix (Feb 19 2019 at 21:06, on Zulip):

ordinarily I'd save my pretriage effort for tomorrow, but since I'm about to try to write these notes for the compiler team repo, I'll do it now.

pnkfelix (Feb 19 2019 at 21:07, on Zulip):

(Though I'm also realizing that maybe the right way to handle the NLL notes is simply to use the notes from last Wednesday's meeting as the initial entry for the compiler team notes. Not sure, its sort of goofy since then we'll have the NLL meeting itself in the interim before the actually compiler team meeting.)

pnkfelix (Feb 19 2019 at 21:08, on Zulip):

((tl;dr: we have a race condition.))

pnkfelix (Feb 19 2019 at 21:10, on Zulip):

anyway, pretriage: We currently have zero uncategorized issues. :tada:

pnkfelix (Feb 19 2019 at 21:11, on Zulip):

we have 27 categorized issues that are lacking a P-label.

pnkfelix (Feb 19 2019 at 21:13, on Zulip):

lets nip these in the bud

pnkfelix (Feb 19 2019 at 21:13, on Zulip):

"MIR borrowck: obscure error message for move out of struct with destructor" #46634

pnkfelix (Feb 19 2019 at 21:13, on Zulip):

this is a feature request; the message from NLL is superior to that from AST-borrowck.

pnkfelix (Feb 19 2019 at 21:13, on Zulip):

therefore, marking as P-low

pnkfelix (Feb 19 2019 at 21:14, on Zulip):

Next: "augmented-assignments test output has poor spans in NLL" #51000

pnkfelix (Feb 19 2019 at 21:15, on Zulip):

The issue here is a bit inflated since I think the pattern

LHS
+=
RHS;

is quite rare

pnkfelix (Feb 19 2019 at 21:16, on Zulip):

marking P-low. And also pinging @Matthew Jasper to see if they still want it assigned to them.

pnkfelix (Feb 19 2019 at 21:19, on Zulip):

Next: "Unclear "cannot infer" error message instead of expected "cannot move out of borrowed content"" #51759

pnkfelix (Feb 19 2019 at 21:20, on Zulip):

the issue here, as niko noted back in June 2018, predates NLL: essentially our diagnostics when it comes to auto-injected reborrows+derefs coercions is subpar and confusing.

pnkfelix (Feb 19 2019 at 21:21, on Zulip):

I think this is an aspect of Rust that we need to do a better job of communicating. A lot of our audience is unlikely to be familiar with the notions here.

pnkfelix (Feb 19 2019 at 21:21, on Zulip):

so marking P-medium.

pnkfelix (Feb 19 2019 at 21:23, on Zulip):

next: "[nll] change how MIR represents places" #52708

pnkfelix (Feb 19 2019 at 21:24, on Zulip):

this is a cross-cutting concern, not NLL specific at all, IMO.

lqd (Feb 19 2019 at 21:25, on Zulip):

seems it has moved to "t-compiler place 2.0" MIR2/etc after all hands, but I'm not sure if there's an issue for it there yet ?

pnkfelix (Feb 19 2019 at 21:26, on Zulip):

yep. The Paper doc for t-compiler place 2.0 does link to issue #52708

pnkfelix (Feb 19 2019 at 21:26, on Zulip):

so I'm just going to remove the NLL tag(s) from it

Matthew Jasper (Feb 19 2019 at 21:29, on Zulip):

marking P-low. And also pinging @Matthew Jasper to see if they still want it assigned to them.

Yes, this is still on my list for when I have time

pnkfelix (Feb 19 2019 at 21:29, on Zulip):

next: "[nll] hash borrows in scope for better performance" #53159

pnkfelix (Feb 19 2019 at 21:29, on Zulip):

Hmm. It sounds like it is not clear whether this idea would still be much of a win. Maybe it would be good, but I'm not in a position to judge at the moment.

pnkfelix (Feb 19 2019 at 21:31, on Zulip):

I'm tempted to mark it P-medium, though I'm a little worried then it will be ignored forever even though it shouldn't be

lqd (Feb 19 2019 at 21:31, on Zulip):

I think it's related to place 2.0 as well ?

pnkfelix (Feb 19 2019 at 21:31, on Zulip):

is it?

lqd (Feb 19 2019 at 21:32, on Zulip):

I feel those issues came up around the same piece of code

lqd (Feb 19 2019 at 21:32, on Zulip):

requiring unroll_place and the likes

pnkfelix (Feb 19 2019 at 21:33, on Zulip):

I'll explicitly leave it unprioritized for now

pnkfelix (Feb 19 2019 at 21:35, on Zulip):

next: ""invert" borrow computation" #53328

pnkfelix (Feb 19 2019 at 21:36, on Zulip):

now this one niko explicitly pointed out last September that we are better off waiting until we've made more progress on Polonius before we invest much effort here.

lqd (Feb 19 2019 at 21:36, on Zulip):

(lol or maybe it was this one related to place 2.0 sorry felix :) (but these are all more or less interconnected)

pnkfelix (Feb 19 2019 at 21:37, on Zulip):

I don't know, in either case it seems like the computation here is tied to how we propagate borrows

pnkfelix (Feb 19 2019 at 21:38, on Zulip):

which I think of as being mostly decoupled from the underlying representation of Places themselves?

lqd (Feb 19 2019 at 21:40, on Zulip):

yeah from the representation but not the uses of the new representation, I'm probably unclear

pnkfelix (Feb 19 2019 at 21:41, on Zulip):

okay I'm going to stop pre-triaging now, I wasn't planning to spend too much time on this

lqd (Feb 19 2019 at 21:41, on Zulip):

do you need any help with it ?

pnkfelix (Feb 19 2019 at 21:41, on Zulip):

no, I just realized its not a good use of my time right now

pnkfelix (Feb 19 2019 at 21:42, on Zulip):

I had hoped it would "naturally" lead to whatever notes I enter into rust-lang/compiler-team WG-nll notes

pnkfelix (Feb 19 2019 at 21:43, on Zulip):

actually ... Place 2.0 is probably a decent topic for us to plug in there, given that different people have thoughts on it.

lqd (Feb 20 2019 at 17:58, on Zulip):

(sorry but I might not be able to make it to tonight's meeting)

pnkfelix (Feb 20 2019 at 20:24, on Zulip):

oh yeah, hey @WG-compiler-nll , we have a meeting in ... about six minutes :fear:

pnkfelix (Feb 20 2019 at 20:25, on Zulip):

I didn't finish the pretriage that I started last night

pnkfelix (Feb 20 2019 at 20:25, on Zulip):

but I guess I can spend five minutes now doing more of it. :working_on_it:

pnkfelix (Feb 20 2019 at 20:28, on Zulip):

"NLL reports "borrow used here in later iteration of loop" in cases outside of loop" #53773 => P-medium. (Already assigned to @Santiago Pastorino and has a PR)

pnkfelix (Feb 20 2019 at 20:28, on Zulip):

" fix "bivariant wf" bug in the NLL subtyping code" #54105 => leaving unprioritized for now but I'm inclined to tag as P-medium given that we still don't know whether this can be exploited.

pnkfelix (Feb 20 2019 at 20:29, on Zulip):

"High memory usage compiling keccak benchmark" #54208 => WG-compiler-performance, P-medium

pnkfelix (Feb 20 2019 at 20:31, on Zulip):

"Poor borrow checker error message" #54256 => needs a better title. P-medium

pnkfelix (Feb 20 2019 at 20:32, on Zulip):

okay @WG-compiler-nll lets get started.

pnkfelix (Feb 20 2019 at 20:33, on Zulip):

So our agenda has effectively three items on it: Status updates, the Meta-Question of "What is this meeting for", and a new item I added "last minute" last night: Place 2.0

pnkfelix (Feb 20 2019 at 20:34, on Zulip):

Everyone feel free to add whatever you like about status updates to the NLL triage paper

pnkfelix (Feb 20 2019 at 20:35, on Zulip):

Regarding status, I had the rough impression that bors/homu had a really time this past week. Did anyone else feel that way? It seemed like pulling teeth to get anything merged.

davidtwco (Feb 20 2019 at 20:36, on Zulip):

I believe there were a variety of issues with the CI providers throughout the week.

pnkfelix (Feb 20 2019 at 20:37, on Zulip):

Anyway: Lets maybe try to open the floor now to talk about this meeting itself

pnkfelix (Feb 20 2019 at 20:37, on Zulip):

(and I plan to put a cap of about 10-12 minutes on such discussion before we move on to Place 2.0)

Santiago Pastorino (Feb 20 2019 at 20:38, on Zulip):

I really liked the idea of converting this meeting into a space where we can "hack" together or ask questions

Santiago Pastorino (Feb 20 2019 at 20:38, on Zulip):

but unsure if triaging and all that is helping you @pnkfelix in any way

pnkfelix (Feb 20 2019 at 20:38, on Zulip):

We can try that. I would like to see something more vibrant.

pnkfelix (Feb 20 2019 at 20:38, on Zulip):

But of course issues will still need to get triaged.

Santiago Pastorino (Feb 20 2019 at 20:39, on Zulip):

I recognize I haven't done any issue triaging and neither help there, I can start doing so and tackling random tickets too :)

pnkfelix (Feb 20 2019 at 20:39, on Zulip):

so far the meeting has been a kind of driver to force me to do triage

Santiago Pastorino (Feb 20 2019 at 20:39, on Zulip):

hehe that's important :)

pnkfelix (Feb 20 2019 at 20:39, on Zulip):

I don't mind doing triage, at least for now, but I need to keep myself honest about making sure it gets done.

pnkfelix (Feb 20 2019 at 20:39, on Zulip):

more importantly, though: Part of triage is to actually assign work

pnkfelix (Feb 20 2019 at 20:40, on Zulip):

namely the P-high issues

Santiago Pastorino (Feb 20 2019 at 20:40, on Zulip):

we can trick you that this meeting is for triaging and changing to something different everytime :P

Santiago Pastorino (Feb 20 2019 at 20:40, on Zulip):

well yeah, that's important

pnkfelix (Feb 20 2019 at 20:40, on Zulip):

my current thinking is that I can have a weekly topic dedicated to triage, e.g. "triage 2019.02.27" would be next week's.

davidtwco (Feb 20 2019 at 20:40, on Zulip):

I’m happy with that. I’ve mostly not been doing a lot of NLL tasks so the meeting for me is just checking if there’s any high priority items that nobody has earmarked that I might have time for. It seems like most everything has assignee perpetually now?

Santiago Pastorino (Feb 20 2019 at 20:41, on Zulip):

what about redifining the meeting as work assign + working together/questions?

pnkfelix (Feb 20 2019 at 20:41, on Zulip):

as for assigning work, what would people think if I just assigned the tasks during the triage process, but people would obviously feel free to object/unassign/delegate?

davidtwco (Feb 20 2019 at 20:41, on Zulip):

When I started typing that message, “that” referred to @Santiago Pastorino’s suggestion about this meetings purpose.

Santiago Pastorino (Feb 20 2019 at 20:41, on Zulip):

as for assigning work, what would people think if I just assigned the tasks during the triage process, but people would obviously feel free to object/unassign/delegate?

yeah, maybe that's even better

pnkfelix (Feb 20 2019 at 20:42, on Zulip):

@davidtwco yeah I don't think we've had too many instances of unassigned P-high issues.

Santiago Pastorino (Feb 20 2019 at 20:42, on Zulip):

or you could even chat with us async

davidtwco (Feb 20 2019 at 20:42, on Zulip):

Also fine with that.

Santiago Pastorino (Feb 20 2019 at 20:42, on Zulip):

but I'm fine with you assigning stuff to me

pnkfelix (Feb 20 2019 at 20:42, on Zulip):

okay. So lets plan to give the open discussion/office hours time a shot

Santiago Pastorino (Feb 20 2019 at 20:42, on Zulip):

the only issue I see with that is that we may be working on something different

pnkfelix (Feb 20 2019 at 20:42, on Zulip):

normally I'd say something like "starting next week"

Santiago Pastorino (Feb 20 2019 at 20:42, on Zulip):

like Place 2.0, Polonius or something like that

pnkfelix (Feb 20 2019 at 20:42, on Zulip):

but in fact the next agenda item is Place 2.0

pnkfelix (Feb 20 2019 at 20:43, on Zulip):

which is arguably a good "office hour" topic anyway

pnkfelix (Feb 20 2019 at 20:43, on Zulip):

@Santiago Pastorino I saw some back-and-forth between you and @oli about it

Santiago Pastorino (Feb 20 2019 at 20:43, on Zulip):

the only issue I see with that is that we may be working on something different

I was referring to your idea of you assigning stuff to us

Santiago Pastorino (Feb 20 2019 at 20:43, on Zulip):

but ok, we moved to the next topic

pnkfelix (Feb 20 2019 at 20:43, on Zulip):

Well

Santiago Pastorino (Feb 20 2019 at 20:43, on Zulip):

@Santiago Pastorino I saw some back-and-forth between you and @oli about it

yes

pnkfelix (Feb 20 2019 at 20:43, on Zulip):

we don't need to move on quite yet

davidtwco (Feb 20 2019 at 20:43, on Zulip):

like Place 2.0, Polonius or something like that

I assume Polonius is purview of #t-compiler/wg-polonius now.

pnkfelix (Feb 20 2019 at 20:44, on Zulip):

I don't want to move on if there are questions/ideas to continue with

pnkfelix (Feb 20 2019 at 20:44, on Zulip):

regarding the meeting

Santiago Pastorino (Feb 20 2019 at 20:44, on Zulip):

ok

davidtwco (Feb 20 2019 at 20:44, on Zulip):

I’m happy to move on.

pnkfelix (Feb 20 2019 at 20:44, on Zulip):

so lets recap: For now, I'll continue doing weekly triage. I'll asynchronously assign work to people.

pnkfelix (Feb 20 2019 at 20:44, on Zulip):

I'll try to be semi-balanced

Santiago Pastorino (Feb 20 2019 at 20:44, on Zulip):

so ... I think Felix's idea of assigning stuff to us is great because it removes some process of the table

Santiago Pastorino (Feb 20 2019 at 20:45, on Zulip):

but on the other hand, what if we are working on something heavy like that

Santiago Pastorino (Feb 20 2019 at 20:45, on Zulip):

regardless of the working group it belongs

pnkfelix (Feb 20 2019 at 20:45, on Zulip):

in terms of trying to recognize which people seem to already have a lot on their plate and/or not enough time

Santiago Pastorino (Feb 20 2019 at 20:45, on Zulip):

anyway I guess we can just chat about tasks with you directly?

pnkfelix (Feb 20 2019 at 20:45, on Zulip):

I think the most important thing is that people need to be honest/realistic about their time availability

Santiago Pastorino (Feb 20 2019 at 20:45, on Zulip):

yeah

pnkfelix (Feb 20 2019 at 20:46, on Zulip):

namely, if you have a week where you won't be around to even look at what work is assigned to you

pnkfelix (Feb 20 2019 at 20:46, on Zulip):

I really need to know that ahead of time

Santiago Pastorino (Feb 20 2019 at 20:46, on Zulip):

seems good

pnkfelix (Feb 20 2019 at 20:46, on Zulip):

unless we're going to request explicit acknowledgement of work assignment?

davidtwco (Feb 20 2019 at 20:46, on Zulip):

We could perhaps signal that in the triage document for the upcoming week.

Santiago Pastorino (Feb 20 2019 at 20:46, on Zulip):

I'm fine with either, we can always talk about stuff :)

pnkfelix (Feb 20 2019 at 20:47, on Zulip):

that is: I can either 1. assign work by using the "Assignee" list in Github issue, or 2. Leave "Assignee" blank and just ask people to assign themselves.

pnkfelix (Feb 20 2019 at 20:47, on Zulip):

The more I think about it, the more I think (2.) might be better...

davidtwco (Feb 20 2019 at 20:47, on Zulip):

Isn’t 2) what currently happens?

pnkfelix (Feb 20 2019 at 20:47, on Zulip):

since I'm very nervous about giving other deveopers the wrong impression about what is being actively worked on.

Santiago Pastorino (Feb 20 2019 at 20:48, on Zulip):

I think 2 is better

pnkfelix (Feb 20 2019 at 20:48, on Zulip):

Yes, 2. is what currently happens, but we often do it synchronously (right?)

pnkfelix (Feb 20 2019 at 20:48, on Zulip):

e.g. during the meeting

Santiago Pastorino (Feb 20 2019 at 20:48, on Zulip):

Isn’t 2) what currently happens?

I guess, Felix is suggesting to do it async rather than in this meeting

Santiago Pastorino (Feb 20 2019 at 20:48, on Zulip):

if so, agreed :)

pnkfelix (Feb 20 2019 at 20:49, on Zulip):

while the suggestion would be, I guess, to instead make the request asynchronously. My plan would be to post both in a Github comment and in the Zulip triage topic

davidtwco (Feb 20 2019 at 20:49, on Zulip):

Maybe I’ve just not been paying enough attention. I thought most issues already had assignees by the time we got to the meeting in recent weeks. I’m happy with either approach, not too bothered.

pnkfelix (Feb 20 2019 at 20:50, on Zulip):

My plan would be to post both in a Github comment and in the Zulip triage topic

(the reason I would opt to do both, and not just Github comment, is that I know for myself that my Github notifications are totally out of control. So I would want, for others doing this to me, for them to use additional channels rather than github alone.)

pnkfelix (Feb 20 2019 at 20:50, on Zulip):

okay then, I think i think know how to handle things going forward then.

pnkfelix (Feb 20 2019 at 20:51, on Zulip):

Longer-term, we may want to rotate the actual triage duty itself.

pnkfelix (Feb 20 2019 at 20:51, on Zulip):

but I am willing to wait before trying to do that.

pnkfelix (Feb 20 2019 at 20:51, on Zulip):

okay, any other thoughts before Place 2.0 ?

Santiago Pastorino (Feb 20 2019 at 20:51, on Zulip):

no

Santiago Pastorino (Feb 20 2019 at 20:51, on Zulip):

let's jump into Place 2.0

Santiago Pastorino (Feb 20 2019 at 20:51, on Zulip):

so, yeah, I have been talking to @oli about it

Santiago Pastorino (Feb 20 2019 at 20:51, on Zulip):

there are a couple of documents

pnkfelix (Feb 20 2019 at 20:52, on Zulip):

lets link that topic maybe

pnkfelix (Feb 20 2019 at 20:52, on Zulip):

zulip t-compiler place2.0

Santiago Pastorino (Feb 20 2019 at 20:52, on Zulip):

exactly, went for the link you beat me :)

Santiago Pastorino (Feb 20 2019 at 20:52, on Zulip):

we have https://paper.dropbox.com/doc/Place-2.0--AX5Ih7aoGsihZcG5EiMuTZ4DAg-9NjhX4N9I3dEt6YCJM8Ln

Santiago Pastorino (Feb 20 2019 at 20:53, on Zulip):

and https://paper.dropbox.com/doc/Place-2.0-current-PR-status--AX6hBHPBw5yIf3FWVl2Ouqd8Ag-vmbnFv8VkCEuL57QfWWMH

Santiago Pastorino (Feb 20 2019 at 20:53, on Zulip):

Oli wanted to take a different approach

Santiago Pastorino (Feb 20 2019 at 20:53, on Zulip):

so we can make tiny steps and be able to merge work faster

Santiago Pastorino (Feb 20 2019 at 20:53, on Zulip):

for me would be the same but Oli stated that in the way they proposed is going to be easier to review

Santiago Pastorino (Feb 20 2019 at 20:53, on Zulip):

that's a fair point

pnkfelix (Feb 20 2019 at 20:53, on Zulip):

so the main question I have had is

pnkfelix (Feb 20 2019 at 20:54, on Zulip):

Are both/either of you confident that the API will work with NeoPlace ?

Santiago Pastorino (Feb 20 2019 at 20:54, on Zulip):

what API?

Santiago Pastorino (Feb 20 2019 at 20:54, on Zulip):

the iterate thing?

pnkfelix (Feb 20 2019 at 20:54, on Zulip):

I guess oli's musings

pnkfelix (Feb 20 2019 at 20:54, on Zulip):

but yeah

pnkfelix (Feb 20 2019 at 20:55, on Zulip):

when I was playing with this locally, I got hung up on first trying to redo NeoPlace from scratch, so that I could then make the iteration API on top it

pnkfelix (Feb 20 2019 at 20:55, on Zulip):

because I wanted to make sure that whatever the iteration API was, it would actually work with either Place or NeoPlace

Santiago Pastorino (Feb 20 2019 at 20:56, on Zulip):

well it should be compatible

pnkfelix (Feb 20 2019 at 20:56, on Zulip):

but I think part of the goal in my head was to have something that would cover all the use-cases, some how

Santiago Pastorino (Feb 20 2019 at 20:56, on Zulip):

since we are going to return Projections one by one

Santiago Pastorino (Feb 20 2019 at 20:56, on Zulip):

I guess

pnkfelix (Feb 20 2019 at 20:56, on Zulip):

while Oli's musings expliclty say implement some of the place processing algorithms (emphasis on "some")

Santiago Pastorino (Feb 20 2019 at 20:57, on Zulip):

yeah, Oli wasn't 100%, I told him ok I can start with the first steps and we can see how it ends

Santiago Pastorino (Feb 20 2019 at 20:57, on Zulip):

would be fairly easy to do the first part of the work

Santiago Pastorino (Feb 20 2019 at 20:57, on Zulip):

I ended with a lot of meetings today but I guess I can finish the thing tomorrow, maybe

pnkfelix (Feb 20 2019 at 20:58, on Zulip):

okay

pnkfelix (Feb 20 2019 at 20:58, on Zulip):

one thing I will say

Santiago Pastorino (Feb 20 2019 at 20:58, on Zulip):

even following the previous approach the change is doable and merge conflicts are not that bad or weren't that bad really

Santiago Pastorino (Feb 20 2019 at 20:58, on Zulip):

the thing I understood is that it could be a pain to review

pnkfelix (Feb 20 2019 at 20:59, on Zulip):

By "previous approach" do you mean "have both Place and NeoPlace side by side in the code", converting between them as needed?

pnkfelix (Feb 20 2019 at 21:00, on Zulip):

or are you referring to something else there?

Santiago Pastorino (Feb 20 2019 at 21:00, on Zulip):

no, to that

pnkfelix (Feb 20 2019 at 21:00, on Zulip):

hmm okay. I actually would have expected lots of merge conflicts.

pnkfelix (Feb 20 2019 at 21:01, on Zulip):

but ... well okay I'll just be surprised for now.

Santiago Pastorino (Feb 20 2019 at 21:01, on Zulip):

the main problem with that approach was that we converted from Place to NeoPlace and viceversa freely

Santiago Pastorino (Feb 20 2019 at 21:01, on Zulip):

and that's not mergeable until we finish all the work

Santiago Pastorino (Feb 20 2019 at 21:01, on Zulip):

otherwise would be a huge performance regression

pnkfelix (Feb 20 2019 at 21:01, on Zulip):

yes okay

pnkfelix (Feb 20 2019 at 21:02, on Zulip):

I believe there could exist a way to structure a single PR broken up into several commits where each commit was relatively easy to review

pnkfelix (Feb 20 2019 at 21:02, on Zulip):

but I also prefer the "iterator API" approach, assuming it works.

Santiago Pastorino (Feb 20 2019 at 21:02, on Zulip):

yeah

davidtwco (Feb 20 2019 at 21:03, on Zulip):

Are eddy’s concerns from the current status document still relevant or have those been resolved?

pnkfelix (Feb 20 2019 at 21:03, on Zulip):

the main advice I have there is that I had worried (and I still worry) about someone getting bogged down with trying to make the iterator API "look nice"

pnkfelix (Feb 20 2019 at 21:03, on Zulip):

by this, I mean, I don't know, something like trying to figure out how to make a method on Place that returns an Iterator over its projections

pnkfelix (Feb 20 2019 at 21:04, on Zulip):

but reading @oli 's notes now, I'm happy to see that they did not impose that in their sketch.

Santiago Pastorino (Feb 20 2019 at 21:04, on Zulip):

Are eddy’s concerns from the current status document still relevant or have those been resolved?

and reinterning a lot instead of slicing/iterating projection lists (which is what the recursive code would’ve done had it been written for a flat structure), unsure about this, the rest I guess it's solved

Santiago Pastorino (Feb 20 2019 at 21:04, on Zulip):

but anyway, we are kind of starting from scratch

pnkfelix (Feb 20 2019 at 21:04, on Zulip):

okay

Santiago Pastorino (Feb 20 2019 at 21:05, on Zulip):

a lot of the code was not written by me, so I'm not 100% sure

Santiago Pastorino (Feb 20 2019 at 21:05, on Zulip):

now I can pay more attention to this kind of details

pnkfelix (Feb 20 2019 at 21:05, on Zulip):

is @csmoe jumping back in on this? Or are you doing it on your own? Or is something else assisting?

Santiago Pastorino (Feb 20 2019 at 21:06, on Zulip):

unsure about @csmoe but I can do it on my own

pnkfelix (Feb 20 2019 at 21:06, on Zulip):

okay

pnkfelix (Feb 20 2019 at 21:06, on Zulip):

I need to walk my dog. But this all sounds fine with me.

Santiago Pastorino (Feb 20 2019 at 21:06, on Zulip):

I could pair or something with @csmoe unsure their schedule and stuff

pnkfelix (Feb 20 2019 at 21:07, on Zulip):

So if you are starting from scratch, are you going to create an entirely new branch? Or are you going to force-push to your old place2 branch?

Santiago Pastorino (Feb 20 2019 at 21:08, on Zulip):

probably will make a copy of the current thing as place2_old and force push to place2 the new code

pnkfelix (Feb 20 2019 at 21:11, on Zulip):

Okay. I'm going to go walk my dog now. If anyone has any questions, feel free to throw them in here.

Matthew Jasper (Feb 20 2019 at 21:13, on Zulip):

Did you agree whether you or NIko would review #57609 and #57202 ?

pnkfelix (Feb 20 2019 at 21:14, on Zulip):

that hasn't been discussed but should be.

pnkfelix (Feb 20 2019 at 21:14, on Zulip):

I think I should review the match changes just to get it off niko's plate.

pnkfelix (Feb 20 2019 at 21:15, on Zulip):

@Matthew Jasper I hate to ask again regarding #57609, but: The diagram in the description reflects the current status, right?

pnkfelix (Feb 20 2019 at 21:15, on Zulip):

(I'm sure you've answered that Q before, that's why I hate to ask)

Matthew Jasper (Feb 20 2019 at 21:17, on Zulip):

It shows both the old and the new status and should be up-to date. I'll just double check.

pnkfelix (Feb 20 2019 at 21:17, on Zulip):

and for #57202 -- you say there's another branch that you think is more principled and a smaller diff? .... what's the drawback? Why not just switch to that?

pnkfelix (Feb 20 2019 at 21:18, on Zulip):

or you already pushed that version to the branch that's currently reflected on PR #57202 ?

Matthew Jasper (Feb 20 2019 at 21:19, on Zulip):

There are 3 versions. The first version is gone now. The other branch has the second version. And the current PR is the third version.

Matthew Jasper (Feb 20 2019 at 21:19, on Zulip):

It's probably best to just review the third.

Matthew Jasper (Feb 20 2019 at 21:20, on Zulip):

It shows both the old and the new status and should be up-to date. I'll just double check.

Yes, the diagram is correct.

pnkfelix (Feb 20 2019 at 21:20, on Zulip):

It's probably best to just review the third.

okay. I'll assign review of that one to me as well.

Matthew Jasper (Feb 20 2019 at 21:24, on Zulip):

Thanks!

Jake Goulding (Feb 20 2019 at 21:37, on Zulip):

Okay. I'm going to go walk my dog now. If anyone has any questions, feel free to throw them in here.

What is the dog's name? :dog:

pnkfelix (Feb 20 2019 at 21:40, on Zulip):

"Elmo" ; it is short for "El Monstre" (he is originally from Spain and is a monster). :alien:

csmoe (Feb 20 2019 at 23:32, on Zulip):

@Santiago Pastorino I can still help if you wanna.

csmoe (Feb 20 2019 at 23:37, on Zulip):

@pnkfelix And for the "assignment", please consider me a chance if suitable. I'm eager to dive deeper into rustc.

Santiago Pastorino (Feb 21 2019 at 14:22, on Zulip):

@csmoe yeah, we can go something if you want

csmoe (Feb 21 2019 at 15:26, on Zulip):

@Santiago Pastorino okay so I'm gonna follow you path

Santiago Pastorino (Feb 21 2019 at 15:44, on Zulip):

@csmoe I mean, I don't know exactly how to share the work still but we could pair if you want

Santiago Pastorino (Feb 21 2019 at 15:45, on Zulip):

I can also just do the work if you want to do something different

Santiago Pastorino (Feb 21 2019 at 15:45, on Zulip):

whatever you prefer ;)

Santiago Pastorino (Feb 21 2019 at 15:45, on Zulip):

or you could just follow the branch and jump when you want to add something :)

csmoe (Feb 21 2019 at 15:52, on Zulip):

or you could just follow the branch and jump when you want to add something :slight_smile:

will track your branch

Last update: Nov 21 2019 at 13:55UTC