Stream: t-compiler/wg-nll

Topic: weekly meeting October 30


pnkfelix (Oct 30 2018 at 19:23, on Zulip):

hi @WG-compiler-nll we'll be having our weekly meeting in about 7 minutes. (There was a time change in EU, but not in USA yet I think, and thus the time may come an hour earlier than some of you may have expected...)

davidtwco (Oct 30 2018 at 19:25, on Zulip):

:astonished:

pnkfelix (Oct 30 2018 at 19:31, on Zulip):

hi all

Andrew Poelstra (Oct 30 2018 at 19:31, on Zulip):

I don't want to start a DST discussion, but FYI Icelandic time is UTC+0 and does not have DST, and you can use this with google calendar (which inexplicably does not have "UTC" as an option)

Andrew Poelstra (Oct 30 2018 at 19:31, on Zulip):

so it's a convenient timezone to use for scheduling international meetings

pnkfelix (Oct 30 2018 at 19:32, on Zulip):

there's a whole topic area under T-compiler on this

pnkfelix (Oct 30 2018 at 19:33, on Zulip):

in terms of deciding whether we should pin our weekly meetings to 1. UTC, or 2. USA time (the current). (I don't think anyone has actually suggested options like 3. EU time, though it certainly could be an option ...)

nikomatsakis (Oct 30 2018 at 19:33, on Zulip):

seemed like the current consensus was to stick with USA time

nikomatsakis (Oct 30 2018 at 19:34, on Zulip):

for better and worse

nikomatsakis (Oct 30 2018 at 19:35, on Zulip):

shall we get started?

nikomatsakis (Oct 30 2018 at 19:35, on Zulip):

/me remembers the paper doc

pnkfelix (Oct 30 2018 at 19:36, on Zulip):

so lets see

pnkfelix (Oct 30 2018 at 19:36, on Zulip):

I put out a metaphorical fire last night. :smile: :fire: :fire_engine:

pnkfelix (Oct 30 2018 at 19:37, on Zulip):

the nice thing is that I think I discovered that fire via the NLL diagnostic review

pnkfelix (Oct 30 2018 at 19:38, on Zulip):

so: thank you to @davidtwco for taking care of porting that to use migrate

nikomatsakis (Oct 30 2018 at 19:38, on Zulip):

:my hero:

pnkfelix (Oct 30 2018 at 19:38, on Zulip):

because that was crucial to discovering the fire

nikomatsakis (Oct 30 2018 at 19:38, on Zulip):

I'm annoyed that there is no "first place ribbon" emoji

nikomatsakis (Oct 30 2018 at 19:38, on Zulip):

anyway

nikomatsakis (Oct 30 2018 at 19:39, on Zulip):

what the. I searched for "medal" specifically, I thought =)

nikomatsakis (Oct 30 2018 at 19:39, on Zulip):

did you complete the review, @pnkfelix ?

nikomatsakis (Oct 30 2018 at 19:39, on Zulip):

that was a monumental task indeed

pnkfelix (Oct 30 2018 at 19:39, on Zulip):

no the review is not done; toddler bathtime took precedence this evening :bath:

pnkfelix (Oct 30 2018 at 19:39, on Zulip):

but I've got 35 tests left

nikomatsakis (Oct 30 2018 at 19:39, on Zulip):

so nearly

pnkfelix (Oct 30 2018 at 19:40, on Zulip):

so it should get done tonight

pnkfelix (Oct 30 2018 at 19:40, on Zulip):

I guess the big question(s) are: What next? and What next?

nikomatsakis (Oct 30 2018 at 19:40, on Zulip):

yeah so

nikomatsakis (Oct 30 2018 at 19:41, on Zulip):

I was thinking about this :)

pnkfelix (Oct 30 2018 at 19:41, on Zulip):

i.e. that's what I take the "life after milestones" topic to mean

nikomatsakis (Oct 30 2018 at 19:41, on Zulip):

yes to some extent :)

pnkfelix (Oct 30 2018 at 19:41, on Zulip):

to be clear: When you say the release is "out"

nikomatsakis (Oct 30 2018 at 19:41, on Zulip):

although a big part of that was also just "how should we organize things"

pnkfelix (Oct 30 2018 at 19:41, on Zulip):

you mean it is in the beta channel, right?

nikomatsakis (Oct 30 2018 at 19:41, on Zulip):

yes

nikomatsakis (Oct 30 2018 at 19:41, on Zulip):

I guess a first question is: do you agree we should use the Release milestone for "things that need backports"?

nikomatsakis (Oct 30 2018 at 19:41, on Zulip):

and, if so, what criteria would we use for that?

pnkfelix (Oct 30 2018 at 19:41, on Zulip):

I just want everyone to be clear about this. As in, there is still both the power and responsibility to make fixes on the beta channel before it actually leaks to the world on stable

nikomatsakis (Oct 30 2018 at 19:42, on Zulip):

I'm not convinced that e.g. the remaining "user annotation" things merit backports, although they are strictly speaking NLL-sound

pnkfelix (Oct 30 2018 at 19:42, on Zulip):

I think the Release milestone is as good a tag as any for identifying the "things that need backports"

nikomatsakis (Oct 30 2018 at 19:43, on Zulip):

so my working hypothesis is "we don't have to backport anything just now"

nikomatsakis (Oct 30 2018 at 19:43, on Zulip):

unless we find NLL-sound (or, potentially, NLL-complete) bugs we think are likely to affect a "lot" of folks

nikomatsakis (Oct 30 2018 at 19:43, on Zulip):

(well or diagnostics if they are truly awful I guess)

pnkfelix (Oct 30 2018 at 19:44, on Zulip):

so that probably means that a lot of stuff that is currently tagged for Release milestone

pnkfelix (Oct 30 2018 at 19:44, on Zulip):

should be untagged

nikomatsakis (Oct 30 2018 at 19:44, on Zulip):

yes, that's what I was thinking

pnkfelix (Oct 30 2018 at 19:44, on Zulip):

right?

davidtwco (Oct 30 2018 at 19:45, on Zulip):

@pnkfelix roughly how many of the minor diagnostic hiccups are distinct issues?

pnkfelix (Oct 30 2018 at 19:45, on Zulip):

not sure. I started trying to collapse them at some point in the process, but they aren't all collapsed.

pnkfelix (Oct 30 2018 at 19:45, on Zulip):

but I don't think any of them are worth a backport

pnkfelix (Oct 30 2018 at 19:46, on Zulip):

I've been trying to use different columns for things that are more serious

pnkfelix (Oct 30 2018 at 19:46, on Zulip):

but the only one of those (besides the aforementioned fire) is a problem with unions that is not a stable feature.

pnkfelix (Oct 30 2018 at 19:46, on Zulip):

(and I think its an NLL-complete thing, not an NLL-sound one, so migration saves our bacon there. I think.)

nikomatsakis (Oct 30 2018 at 19:46, on Zulip):

for reference, @pnkfelix you are referring to the cards from this GH project, right?

pnkfelix (Oct 30 2018 at 19:47, on Zulip):

oh, yes, sorry

nikomatsakis (Oct 30 2018 at 19:47, on Zulip):

so it seems like if we use "Release" for backport

nikomatsakis (Oct 30 2018 at 19:47, on Zulip):

we probably want some other way to tag "targeted" issues

nikomatsakis (Oct 30 2018 at 19:47, on Zulip):

e.g., it seems like this union thing merits attention

pnkfelix (Oct 30 2018 at 19:47, on Zulip):

P-high ?

pnkfelix (Oct 30 2018 at 19:47, on Zulip):

like, my inclination

pnkfelix (Oct 30 2018 at 19:47, on Zulip):

is to start treating NLL issues

nikomatsakis (Oct 30 2018 at 19:47, on Zulip):

perhaps; that has the side-effect of triggering T-compiler review, which is maybe ok

pnkfelix (Oct 30 2018 at 19:47, on Zulip):

as if they are any other T-compiler issue

nikomatsakis (Oct 30 2018 at 19:47, on Zulip):

yes, ok

nikomatsakis (Oct 30 2018 at 19:48, on Zulip):

that makes a measure of sense, yes

nikomatsakis (Oct 30 2018 at 19:49, on Zulip):

what do we think we ought to do regarding migration mode? i.e, there is some amount of work "strengthening" tests, should we harden migration mode too? (as you were contemplating)

nikomatsakis (Oct 30 2018 at 19:49, on Zulip):

e.g., to require //~ ERROR annotations

nikomatsakis (Oct 30 2018 at 19:49, on Zulip):

I guess it depends how long "migration compare-mode" will remain in use

nikomatsakis (Oct 30 2018 at 19:49, on Zulip):

but my feeling is.. we have some time yet

nikomatsakis (Oct 30 2018 at 19:49, on Zulip):

and new tests will be added in the meantime etc

pnkfelix (Oct 30 2018 at 19:49, on Zulip):

maybe that's a good topic

pnkfelix (Oct 30 2018 at 19:50, on Zulip):

i.e. what is our timeline

pnkfelix (Oct 30 2018 at 19:50, on Zulip):

for when we might attempt to migrate rust2015

nikomatsakis (Oct 30 2018 at 19:50, on Zulip):

maybe a reasonable question is

pnkfelix (Oct 30 2018 at 19:50, on Zulip):

I guess it can't come sooner than the actual release of the 2018 edition

nikomatsakis (Oct 30 2018 at 19:50, on Zulip):

what is the "trigger criteria"?

pnkfelix (Oct 30 2018 at 19:50, on Zulip):

which means we have at least six more weeks of migration

nikomatsakis (Oct 30 2018 at 19:50, on Zulip):

e.g., if we go a whole release without any NLL-specific reported issues

nikomatsakis (Oct 30 2018 at 19:50, on Zulip):

or something

pnkfelix (Oct 30 2018 at 19:51, on Zulip):

I think if the hordes of developers for 2015

pnkfelix (Oct 30 2018 at 19:51, on Zulip):

say they demand it

pnkfelix (Oct 30 2018 at 19:51, on Zulip):

after seeing the joy experienced by the edition=2018 folk

nikomatsakis (Oct 30 2018 at 19:51, on Zulip):

lol

pnkfelix (Oct 30 2018 at 19:51, on Zulip):

You think I'm kidding

pnkfelix (Oct 30 2018 at 19:51, on Zulip):

but I've been doing this diagnostic review

pnkfelix (Oct 30 2018 at 19:51, on Zulip):

NLL is so much nicer in so many cases

nikomatsakis (Oct 30 2018 at 19:51, on Zulip):

indeed

davidtwco (Oct 30 2018 at 19:52, on Zulip):

...and that's without the region errors, right? which IMO are particularly improved from AST.

pnkfelix (Oct 30 2018 at 19:52, on Zulip):

right, even without the NLL region errors

pnkfelix (Oct 30 2018 at 19:53, on Zulip):

so @nikomatsakis alluded above to my desire to "harden" the migration mode

pnkfelix (Oct 30 2018 at 19:53, on Zulip):

basically one of the takeaways I've had from the .stderr review

pnkfelix (Oct 30 2018 at 19:53, on Zulip):

is that its too hard to keep doing this

pnkfelix (Oct 30 2018 at 19:53, on Zulip):

and that we shoudl instead have some policy (that is enforced via compiletest)

pnkfelix (Oct 30 2018 at 19:54, on Zulip):

that the //~ERROR annotations "work" for both default and migration modes

nikomatsakis (Oct 30 2018 at 19:54, on Zulip):

do you have a sense for how often that will be the case?

pnkfelix (Oct 30 2018 at 19:54, on Zulip):

for any case where you can't get that effect, you have to switch to using // revisions: ast nll

pnkfelix (Oct 30 2018 at 19:54, on Zulip):

or some such

pnkfelix (Oct 30 2018 at 19:54, on Zulip):

mmmm

pnkfelix (Oct 30 2018 at 19:55, on Zulip):

that's a good question, I sort of wish I had been keeping that data as I did this pass

nikomatsakis (Oct 30 2018 at 19:55, on Zulip):

I suppose

nikomatsakis (Oct 30 2018 at 19:55, on Zulip):

that if we migrate 2015

nikomatsakis (Oct 30 2018 at 19:55, on Zulip):

that is, enable migration mode globally

nikomatsakis (Oct 30 2018 at 19:55, on Zulip):

this becomes a non-issue

pnkfelix (Oct 30 2018 at 19:55, on Zulip):

yes

nikomatsakis (Oct 30 2018 at 19:55, on Zulip):

hmm

nikomatsakis (Oct 30 2018 at 19:55, on Zulip):

that's a good incentive to hurry up :)

pnkfelix (Oct 30 2018 at 19:55, on Zulip):

that is one motivation for trying to migrate sooner. But its sort of a bad motivation

nikomatsakis (Oct 30 2018 at 19:55, on Zulip):

not the best :)

pnkfelix (Oct 30 2018 at 19:55, on Zulip):

i.e. "our tools are bad so lets just take this risky path"

pnkfelix (Oct 30 2018 at 19:56, on Zulip):

anyway

nikomatsakis (Oct 30 2018 at 19:56, on Zulip):

my feeling is

davidtwco (Oct 30 2018 at 19:56, on Zulip):

I wrote up some ideas in a notebook recently for a "diagnostic review tool" that would allow reviews to be incrementally done. You'd run all the tests in two modes, it would provide you with a UI where you can go through each comparison and approve or disapprove with a message. The tool could then post those as GitHub issues or some such. In the future, months later, you could re-run the tool and it'd load the previous results. It'd run all the tests again and you'd only need to review results that have changed. There were some other details but I forget.

nikomatsakis (Oct 30 2018 at 19:56, on Zulip):

we should migrate 2015 if the Rust 2018 launch goes "smoothly"

pnkfelix (Oct 30 2018 at 19:56, on Zulip):

let's make a guess that the so-far 300 cases of "acceptable/desirable deviation" all fall into that category

davidtwco (Oct 30 2018 at 19:56, on Zulip):

Would have been more useful months ago.

nikomatsakis (Oct 30 2018 at 19:57, on Zulip):

(that is, 2 or 3 weeks after Rust 2018 is released)

nikomatsakis (Oct 30 2018 at 19:57, on Zulip):

Would have been more useful months ago.

we had plans to make such a tool as I recall but compare-mode was the "worse is better" alternative we landed on

pnkfelix (Oct 30 2018 at 19:57, on Zulip):

let's make a guess that the so-far 300 cases of "acceptable/desirable deviation" all fall into that category

(finishing thought: that would leave 100 to 135 or so other tests that would need to use // revisions instead)

nikomatsakis (Oct 30 2018 at 19:57, on Zulip):

a fair number

pnkfelix (Oct 30 2018 at 19:58, on Zulip):

true

pnkfelix (Oct 30 2018 at 19:58, on Zulip):

its possible that the other bugs we hope to fix

pnkfelix (Oct 30 2018 at 19:58, on Zulip):

could cause that number to drop

nikomatsakis (Oct 30 2018 at 19:58, on Zulip):

(that is, 2 or 3 weeks after Rust 2018 is released)

what do you all think of this? basically, we switch nightly at this time if we feel comfortable

pnkfelix (Oct 30 2018 at 19:58, on Zulip):

but its possble that this is just fact of life

nikomatsakis (Oct 30 2018 at 19:58, on Zulip):

but that's still a good 8 weeks or so from now

davidtwco (Oct 30 2018 at 19:59, on Zulip):

what do you all think of this? basically, we switch nightly at this time if we feel comfortable

I like this idea.

pnkfelix (Oct 30 2018 at 19:59, on Zulip):

hmm

pnkfelix (Oct 30 2018 at 19:59, on Zulip):

I guess the assumption is that the ... "marketing push" associated with the 2018 release

pnkfelix (Oct 30 2018 at 19:59, on Zulip):

should imply that the bulk of reports should come very early in the process?

pnkfelix (Oct 30 2018 at 19:59, on Zulip):

its not a bad theory

nikomatsakis (Oct 30 2018 at 19:59, on Zulip):

part of my motivation also is

nikomatsakis (Oct 30 2018 at 19:59, on Zulip):

to give us some "lead time"

nikomatsakis (Oct 30 2018 at 20:00, on Zulip):

i.e., that gives us a few weeks to roll back :)

pnkfelix (Oct 30 2018 at 20:00, on Zulip):

I dont follow

nikomatsakis (Oct 30 2018 at 20:00, on Zulip):

maybe 2 weeks in is a good metric, leaving 4 weeks for people using nightly to complain

pnkfelix (Oct 30 2018 at 20:00, on Zulip):

rollback which?

nikomatsakis (Oct 30 2018 at 20:00, on Zulip):

rollback 2015

pnkfelix (Oct 30 2018 at 20:00, on Zulip):

ah

pnkfelix (Oct 30 2018 at 20:00, on Zulip):

certainly whenever we do it

pnkfelix (Oct 30 2018 at 20:00, on Zulip):

we definitely want that lead time

nikomatsakis (Oct 30 2018 at 20:00, on Zulip):

I'm assuming though that yes the PR push around 2018 helps

pnkfelix (Oct 30 2018 at 20:00, on Zulip):

i.e. it should be timed to happen soon after some release

nikomatsakis (Oct 30 2018 at 20:01, on Zulip):

and that we've done enough crater runs etc

nikomatsakis (Oct 30 2018 at 20:01, on Zulip):

and we have migration mode

davidtwco (Oct 30 2018 at 20:01, on Zulip):

Do either of you anticipate a few NLL issues being raised when 2018 is released or do you think it'll go smoothly?

nikomatsakis (Oct 30 2018 at 20:01, on Zulip):

I'm not really anticipating any problems

nikomatsakis (Oct 30 2018 at 20:01, on Zulip):

but I wouldn't be surprised

pnkfelix (Oct 30 2018 at 20:01, on Zulip):

my dream: we get a bunch of people complaining that tutorial examples (of borrow-check errors) stopped working (... and that's the bulk of the complaints)

pnkfelix (Oct 30 2018 at 20:01, on Zulip):

(as mentioned in "general discussion" I think)

pnkfelix (Oct 30 2018 at 20:02, on Zulip):

I do see #55344 as an example of the kind of thing that we might not have expected

pnkfelix (Oct 30 2018 at 20:03, on Zulip):

jeez louise its been 30min

nikomatsakis (Oct 30 2018 at 20:03, on Zulip):

yes

nikomatsakis (Oct 30 2018 at 20:03, on Zulip):

I was just thinking the same thing :)

pnkfelix (Oct 30 2018 at 20:03, on Zulip):

and we didn't touch on polonius

pnkfelix (Oct 30 2018 at 20:03, on Zulip):

(which I was thinking of the other "what's next?" Q)

nikomatsakis (Oct 30 2018 at 20:03, on Zulip):

I keep meaning to put my head down and spend a day or two thinking about polonius, haven't done it yet -- I think what I'd really like to do is to write out some plans to make it "take over" all the region inferencing jobs

nikomatsakis (Oct 30 2018 at 20:04, on Zulip):

I'm going to put that on my list for this week though:)

nikomatsakis (Oct 30 2018 at 20:04, on Zulip):

I guess the question is: is there other NLL work that I (or others) should be doing?

nikomatsakis (Oct 30 2018 at 20:04, on Zulip):

(beyond polonius)

nikomatsakis (Oct 30 2018 at 20:04, on Zulip):

I basically don't know what should be P-high

nikomatsakis (Oct 30 2018 at 20:04, on Zulip):

unions? user type annotations?

nikomatsakis (Oct 30 2018 at 20:05, on Zulip):

I feel kind of "ok" about diagnostics, although I was talking about scheduling a "diagnostic review" meeting over in #t-compiler/diagnostics

nikomatsakis (Oct 30 2018 at 20:05, on Zulip):

short of that though, I feel like we are well into diminishing returns territory?

pnkfelix (Oct 30 2018 at 20:05, on Zulip):

right

pnkfelix (Oct 30 2018 at 20:05, on Zulip):

I think we are definitely moving into terrain where we'd be better off letting a dedicating WG solve the problem of how to present these errors

pnkfelix (Oct 30 2018 at 20:06, on Zulip):

I do think I might want help deciding which diagnostic changes are actual "regressions"

pnkfelix (Oct 30 2018 at 20:06, on Zulip):

e.g. this card https://github.com/rust-lang/rust/projects/10#card-14186759

pnkfelix (Oct 30 2018 at 20:06, on Zulip):

summary: We have a bunch of cases where AST-borrowck will say "returned data needs to live for lifetime 'a

nikomatsakis (Oct 30 2018 at 20:07, on Zulip):

I did not find that link to work very well

pnkfelix (Oct 30 2018 at 20:07, on Zulip):

and NLL just says "you're returning a borrow of locally allocated state", and doesn't mention 'a

pnkfelix (Oct 30 2018 at 20:07, on Zulip):

I did not find that link to work very well

yes this is a serious problem with the project system

pnkfelix (Oct 30 2018 at 20:07, on Zulip):

but I think my summary will work

nikomatsakis (Oct 30 2018 at 20:07, on Zulip):

I feel like I'd have to see the example to decide if that's a problem or not

pnkfelix (Oct 30 2018 at 20:07, on Zulip):

the point is that: Some people might say "its good NLL doesn't mention 'a.

nikomatsakis (Oct 30 2018 at 20:08, on Zulip):

I think that may have been something we did intentionally

pnkfelix (Oct 30 2018 at 20:08, on Zulip):

oh let me just link one of them then

pnkfelix (Oct 30 2018 at 20:09, on Zulip):

compare this: ast to this nll

pnkfelix (Oct 30 2018 at 20:09, on Zulip):

the NLL is better in almost every way

pnkfelix (Oct 30 2018 at 20:09, on Zulip):

but someone might look at the AST one and say "well it was nice that it was telling me what the min lifetime was that I have to make my return value live for"

pnkfelix (Oct 30 2018 at 20:10, on Zulip):

anyway

pnkfelix (Oct 30 2018 at 20:11, on Zulip):

we don't have to make this decision now

pnkfelix (Oct 30 2018 at 20:11, on Zulip):

its more an example

pnkfelix (Oct 30 2018 at 20:11, on Zulip):

of the kind of thing where I think I'll be creating low priority diagnostic issues

nikomatsakis (Oct 30 2018 at 20:11, on Zulip):

yes I see

nikomatsakis (Oct 30 2018 at 20:11, on Zulip):

hard to say

pnkfelix (Oct 30 2018 at 20:11, on Zulip):

and then we'll just need to make decisions about whether to "solve" them or not

nikomatsakis (Oct 30 2018 at 20:11, on Zulip):

I often get really annoyed when our diganostics tell me "extraneous" information that is not particularly relevant

nikomatsakis (Oct 30 2018 at 20:11, on Zulip):

this is borderline

pnkfelix (Oct 30 2018 at 20:11, on Zulip):

right

nikomatsakis (Oct 30 2018 at 20:11, on Zulip):

borderline because it kind of suggests like "some other lifetime might work"

nikomatsakis (Oct 30 2018 at 20:11, on Zulip):

but of course it won't

pnkfelix (Oct 30 2018 at 20:11, on Zulip):

it'd be one thing if I were returning a &'a _ and I needed a &'b _

nikomatsakis (Oct 30 2018 at 20:11, on Zulip):

but it at least draws attention to the return type -- but then it's not highlighting the return type

nikomatsakis (Oct 30 2018 at 20:12, on Zulip):

I'd rather see the return type underlined and say "--- return type requires a reference" or something

nikomatsakis (Oct 30 2018 at 20:12, on Zulip):

anyway I agree low priority :)

nikomatsakis (Oct 30 2018 at 20:12, on Zulip):

and I'm not sure I'd even file an issue

pnkfelix (Oct 30 2018 at 20:12, on Zulip):

okay

pnkfelix (Oct 30 2018 at 20:12, on Zulip):

well I'll file an issue just beause its how I plan to close out these cards I'm creating

pnkfelix (Oct 30 2018 at 20:12, on Zulip):

(and then we'll have a paper trail for how they were dealt with)

nikomatsakis (Oct 30 2018 at 20:13, on Zulip):

ok

pnkfelix (Oct 30 2018 at 20:13, on Zulip):

(when they are eventually triaged and closed)

nikomatsakis (Oct 30 2018 at 20:13, on Zulip):

I'm gonna have to run soon

pnkfelix (Oct 30 2018 at 20:13, on Zulip):

So... is there any really pressing topic since we're already over our alloted time?

nikomatsakis (Oct 30 2018 at 20:13, on Zulip):

I still don't know what to do with myself :)

nikomatsakis (Oct 30 2018 at 20:14, on Zulip):

I think I will look at polonius

nikomatsakis (Oct 30 2018 at 20:14, on Zulip):

and try to "chart a course" for what it would mean to "enable" it

nikomatsakis (Oct 30 2018 at 20:14, on Zulip):

but I'm amenable to being told there are better things to do

pnkfelix (Oct 30 2018 at 20:15, on Zulip):

I don't know if there's anything better for you to do with respect to NLL

pnkfelix (Oct 30 2018 at 20:15, on Zulip):

okay then I'm going to say this meeting's over then!

pnkfelix (Oct 30 2018 at 20:15, on Zulip):

bye all!

blitzerr (Oct 30 2018 at 22:28, on Zulip):

Can I request to be added to the list that gets pinged when this meeting starts ? I miss it otherwise.

pnkfelix (Oct 30 2018 at 23:38, on Zulip):

@blitzerr okay I think I've added you to @WG-compiler-nll now

blitzerr (Oct 30 2018 at 23:45, on Zulip):

Thanks a lot @pnkfelix

Last update: Nov 21 2019 at 23:25UTC