Stream: wg-traits

Topic: weekly meeting 2019-03-25


nikomatsakis (Mar 25 2019 at 18:02, on Zulip):

Hey @WG-traits -- meeting :)

Alexander Regueiro (Mar 25 2019 at 18:02, on Zulip):

hi hi

nikomatsakis (Mar 25 2019 at 18:03, on Zulip):

Current sprint document

Sunjay Varma (Mar 25 2019 at 18:04, on Zulip):

I seem to have finished my task :tada:

Sunjay Varma (Mar 25 2019 at 18:04, on Zulip):

(we decided not to do it lol)

nikomatsakis (Mar 25 2019 at 18:04, on Zulip):

best way to finish a task :)

nikomatsakis (Mar 25 2019 at 18:05, on Zulip):

I don't know if @Aaron Turon will show up today or not (haven't heard from him today)

nikomatsakis (Mar 25 2019 at 18:05, on Zulip):

Let's check in first with @Alexander Regueiro re: associated type bounds -- last we spoke, you were prep'ing a PR, did you open it?

Alexander Regueiro (Mar 25 2019 at 18:05, on Zulip):

thanks

Alexander Regueiro (Mar 25 2019 at 18:05, on Zulip):

yeah

nikomatsakis (Mar 25 2019 at 18:06, on Zulip):

link?

Alexander Regueiro (Mar 25 2019 at 18:06, on Zulip):

I tagged and r?'ed @Aaron Turon yesterday to review. :-)

Alexander Regueiro (Mar 25 2019 at 18:06, on Zulip):

https://github.com/rust-lang/rust/pull/57428

nikomatsakis (Mar 25 2019 at 18:06, on Zulip):

great

Alexander Regueiro (Mar 25 2019 at 18:07, on Zulip):

there are only two notable test failures:

nikomatsakis (Mar 25 2019 at 18:07, on Zulip):

And I know that you have been interested in the "RPIT Existential Lifetimes"

nikomatsakis (Mar 25 2019 at 18:07, on Zulip):

ah, carry on

Alexander Regueiro (Mar 25 2019 at 18:07, on Zulip):

yeah. did you see the PR I opened about that? (we'll get to it in a minunte though if that's okay)

Alexander Regueiro (Mar 25 2019 at 18:08, on Zulip):

1) nested-lifetime-bounds -- it seems that with ATB in place, that can't be detected so easily now. it's a diagnostics issue

Alexander Regueiro (Mar 25 2019 at 18:08, on Zulip):

I could do with advice on that

Alexander Regueiro (Mar 25 2019 at 18:08, on Zulip):

2) I might have accidentally broke scoping rules around defining uses for existential type -- see the ui/existential_types/declared_but_not_defined_in_scope.rs test

Alexander Regueiro (Mar 25 2019 at 18:08, on Zulip):

the other 2 failures just require blessing to fix :-)

Alexander Regueiro (Mar 25 2019 at 18:09, on Zulip):

@nikomatsakis ^

nikomatsakis (Mar 25 2019 at 18:10, on Zulip):

OK. I guess we'll handle that stuff on the PR, this meeting is prob not time to dive in

Alexander Regueiro (Mar 25 2019 at 18:10, on Zulip):

no prob

Alexander Regueiro (Mar 25 2019 at 18:10, on Zulip):

I thought we want to stick to higher-level stuff here yes

nikomatsakis (Mar 25 2019 at 18:10, on Zulip):

So the other major topic was rustc-chalk integration

Alexander Regueiro (Mar 25 2019 at 18:10, on Zulip):

maybe we can discuss RPIT existential lifetimes briefly though?

Alexander Regueiro (Mar 25 2019 at 18:10, on Zulip):

if we have a sec

Alexander Regueiro (Mar 25 2019 at 18:10, on Zulip):

sorry

nikomatsakis (Mar 25 2019 at 18:10, on Zulip):

Not sure @tmandry how y'all fared with the builtin bounds stuff?

Alexander Regueiro (Mar 25 2019 at 18:10, on Zulip):

do that if it's priority :-)

tmandry (Mar 25 2019 at 18:11, on Zulip):

we decided how to split up the work, but I didn't get any farther than that on this yet

nikomatsakis (Mar 25 2019 at 18:11, on Zulip):

ok :)

nikomatsakis (Mar 25 2019 at 18:12, on Zulip):

so I started cataloging what the next steps are around rustc-chalk integration and dropping notes in this paper doc

nikomatsakis (Mar 25 2019 at 18:12, on Zulip):

it's not really complete yet though

nikomatsakis (Mar 25 2019 at 18:12, on Zulip):

I think the meta goal is to really dive head on towards the unknowns, probably

tmandry (Mar 25 2019 at 18:12, on Zulip):

Aaron Turon: i don't think these are trivial since each of these "built ins" involves some custom logic
Aaron Turon: which we'll have to crib from the old trait system

tmandry (Mar 25 2019 at 18:13, on Zulip):

I haven't been able to look into it yet, though

nikomatsakis (Mar 25 2019 at 18:13, on Zulip):

and I think the biggest unknowns are two-fold:

nikomatsakis (Mar 25 2019 at 18:13, on Zulip):

the first one is tied up, to some extent, in builtin-bounds

nikomatsakis (Mar 25 2019 at 18:13, on Zulip):

well at leaset in that things like ?T: Sized are good examples of challenging problems (as we discussed on the call)

nikomatsakis (Mar 25 2019 at 18:14, on Zulip):

I haven't been able to look into it yet, though

ok

nikomatsakis (Mar 25 2019 at 18:14, on Zulip):

I still think that making some progress there is a good first step before anything. I'm feeling a bit uncertain what I thnk is the next step

nikomatsakis (Mar 25 2019 at 18:14, on Zulip):

I'm going to be away (on vacation) next week

nikomatsakis (Mar 25 2019 at 18:14, on Zulip):

and this week I'm traveling a bit and busy

nikomatsakis (Mar 25 2019 at 18:14, on Zulip):

but ideally i'd like to have enough work lined up =)

nikomatsakis (Mar 25 2019 at 18:14, on Zulip):

one thing that might be a good thing to shoot for this week I thnk

nikomatsakis (Mar 25 2019 at 18:15, on Zulip):

(for me)

Alexander Regueiro (Mar 25 2019 at 18:15, on Zulip):

yeah... Rust Latin American conference, judging by GitHub notifications? :-)

nikomatsakis (Mar 25 2019 at 18:15, on Zulip):

is to try and schedule some time to dig into "the chalk-engine crate and how it works" -- i.e., the actual solver algorithm

Alexander Regueiro (Mar 25 2019 at 18:15, on Zulip):

have fun at that!

Alexander Regueiro (Mar 25 2019 at 18:15, on Zulip):

and spread the Gospel of Rust.

nikomatsakis (Mar 25 2019 at 18:16, on Zulip):

basically to try and talk out what is implemented today and (if we would have time, which I guess is unlikely) to then talk a bit about possible sol'ns

Alexander Regueiro (Mar 25 2019 at 18:16, on Zulip):

@nikomatsakis the region bounds integration in Chalk is tricky it sounds like... probably not something I should look into as a newbie with Chalk?

nikomatsakis (Mar 25 2019 at 18:16, on Zulip):

anyway I guess for me I will try to identify 1 or 2 actionable things after builtin bounds

nikomatsakis (Mar 25 2019 at 18:17, on Zulip):

the region bounds integration in Chalk is tricky it sounds like... probably not something I should look into as a newbie with Chalk?

I think the full sol'n probably not but there might be room to introduce some simpler stuff to start

nikomatsakis (Mar 25 2019 at 18:18, on Zulip):

OK, well, that's probably enough for this week. Main goals here then are to (a) pursue the builtin bounds a bit more and (b) carve out steps for next week :)

Alexander Regueiro (Mar 25 2019 at 18:19, on Zulip):

hmm okay

Alexander Regueiro (Mar 25 2019 at 18:19, on Zulip):

about RPIT existential lifetimes... here's my PR https://github.com/rust-lang/rust/pull/59402

Alexander Regueiro (Mar 25 2019 at 18:19, on Zulip):

discussed this briefly with Matthew Jasper. there are evidently some problems. my last comment is perhaps most relevant

nikomatsakis (Mar 25 2019 at 18:21, on Zulip):

ok, well, let's dive into that in a separate topic

nikomatsakis (Mar 25 2019 at 18:22, on Zulip):

and close up this meeting for today?

nikomatsakis (Mar 25 2019 at 18:22, on Zulip):

(I was just skimming PR, will answer in the RPIT Lifetime issues topic)

nikomatsakis (Mar 25 2019 at 18:22, on Zulip):

Thanks all :tada:

Last update: Nov 18 2019 at 00:45UTC