Stream: t-compiler

Topic: #54818 weekly meeting 2018-01-03


nagisa (Jan 03 2019 at 13:08, on Zulip):

o/

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

@nagisa I was going to ask, do you want to run meeting today or shall I?

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

but I guess maybe that answers my question

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

I didn't have internet connectivity until a few minutes ago (silly cafe) so I'm not going to be able to get nearly as much done as @pnkfelix, but I'll at least get a head-start on triaging the P-high issues.

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

# matching against an associated constant gives type error #57280 -- pending PR

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

# rustdoc crashing on linux-sparc64 due to unaligned access #56927 -- fixed by @Nikita Popov, but we need a beta backport still

nagisa (Jan 03 2019 at 14:58, on Zulip):

@nikomatsakis I can run, and I do have a prepass in my notepad.

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

# PhantomData fields in repr(C) structs change ABI on aarch64 #56877 -- currently assigned to @pnkfelix, but not progress as far as I know. We might want to find someone else.

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

oh, ok I'll just stop :)

nagisa (Jan 03 2019 at 14:59, on Zulip):

Its almost time for the meeting anyway

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

yeah :)

nagisa (Jan 03 2019 at 14:59, on Zulip):

Hi, @T-compiler!

davidtwco (Jan 03 2019 at 15:00, on Zulip):

(@T-compiler/meeting)

nagisa (Jan 03 2019 at 15:01, on Zulip):

# matching against an associated constant gives type error #57280 -- pending PR

@nikomatsakis has a [zulip log] in their last comment but no corresponding [zulip log]:.

davidtwco (Jan 03 2019 at 15:01, on Zulip):

I just submitted a PR that fixes this.

davidtwco (Jan 03 2019 at 15:01, on Zulip):

Ah yeah, you said that. :face_palm:

nagisa (Jan 03 2019 at 15:01, on Zulip):

Yeah, niko already noted that there is a pending PR.

nagisa (Jan 03 2019 at 15:02, on Zulip):

I’ll leave off all P-highs assigned to pnkfelix for last P-highs, as they are absent till probably mid-January?

nagisa (Jan 03 2019 at 15:02, on Zulip):

So going through what @nikomatsakis already looked at...

nagisa (Jan 03 2019 at 15:03, on Zulip):

Tracking beta-backport of #57053, which is beta-nominated, we’ll get to that later.

nagisa (Jan 03 2019 at 15:03, on Zulip):

There has been a fair amount of discussion, but it does not seem like a proper strategy has been formed yet…

nikomatsakis (Jan 03 2019 at 15:03, on Zulip):

I am wondering if we should reassign

nikomatsakis (Jan 03 2019 at 15:04, on Zulip):

although I would guess @pnkfelix will be back around probably next week

nagisa (Jan 03 2019 at 15:04, on Zulip):

If any of you have any good ideas on how and what to filter and how to consider what is ABI-relevant and what is not, a note on PR would be great.

nikomatsakis (Jan 03 2019 at 15:04, on Zulip):

but they don't seem like an obvious candidate, more like @eddyb

nikomatsakis (Jan 03 2019 at 15:05, on Zulip):

although I would guess @pnkfelix will be back around probably next week

er actually I think that's wrong

nagisa (Jan 03 2019 at 15:05, on Zulip):

mid-January according to their own words.

nikomatsakis (Jan 03 2019 at 15:05, on Zulip):

right

nikomatsakis (Jan 03 2019 at 15:05, on Zulip):

@nagisa maybe reassign to me and I'll try to sketch out some suggestions

nagisa (Jan 03 2019 at 15:06, on Zulip):

Okay, moving on (as I write up a note on strategy to the issue discussed above)

nagisa (Jan 03 2019 at 15:06, on Zulip):

Has been discussed at NLL workgroup meeting, assigned to @nikomatsakis for investigation.

nikomatsakis (Jan 03 2019 at 15:07, on Zulip):

yeah that's fine, I'll try to get to that too :)

nagisa (Jan 03 2019 at 15:07, on Zulip):

Okay. Next is...

nagisa (Jan 03 2019 at 15:08, on Zulip):

Left open to figure if #56850 needs to be backported and if so, how. #56850 is beta-accepted.

nagisa (Jan 03 2019 at 15:08, on Zulip):

@Alexander Regueiro is it fine to assign to you to prepare a backport?

nagisa (Jan 03 2019 at 15:09, on Zulip):

(you’re already an assignee)

nagisa (Jan 03 2019 at 15:09, on Zulip):

I’ll write a note on the PR, I guess.

nagisa (Jan 03 2019 at 15:09, on Zulip):

Haven’t touched it since the first look. It still bothers me that there are two issues mashed into one here. I could not reproduce the originally reported one. The crossbeam one is reproducible and has been bisected. The PR seems relevant, but without more investigation it is not clear whom to blame – crossbeam for assuming too much or the PR.

nagisa (Jan 03 2019 at 15:10, on Zulip):

I guess I’ll ask to fill a separate issue for crossbeam

nagisa (Jan 03 2019 at 15:11, on Zulip):

As one would be expected, holidays masterfully prevented any(?) activity here.

nikomatsakis (Jan 03 2019 at 15:11, on Zulip):

( I'm going to assign myself to that and try to push that ball forward in @pnkfelix's absence )

nagisa (Jan 03 2019 at 15:11, on Zulip):

The PR that was a root cause of this to happen has been reverted. This is now a regression from stable-to-stable, so an opportunity for a timely backport seems to have been missed. Long term fix #55617 based on stacker has not been merged. Current strategy in the PR only works on Tier1 platforms only, but I have made a crate over the holidays that supports most of the architectures and have made a PR against stacker to use it.

nagisa (Jan 03 2019 at 15:12, on Zulip):

So the progress is smooth here.

nagisa (Jan 03 2019 at 15:12, on Zulip):

Assuming @Alex Crichton agrees to merge my PR against stacker.

nagisa (Jan 03 2019 at 15:13, on Zulip):

Two more NLL issues are:

but I struggled to write up any summary for either of them, so there isn’t any

nikomatsakis (Jan 03 2019 at 15:13, on Zulip):

I think that #48225 is likely "not a bug"

nikomatsakis (Jan 03 2019 at 15:13, on Zulip):

not sure about #53114, I'll revisit (looks like tomorrow will be an NLL catch up day )

nikomatsakis (Jan 03 2019 at 15:14, on Zulip):

and/or sync up with @Matthew Jasper

nagisa (Jan 03 2019 at 15:14, on Zulip):

@nikomatsakis do you want to reassign both to yourself?

nagisa (Jan 03 2019 at 15:15, on Zulip):

Also it would be nice to remove the I-nominated from #48225

nikomatsakis (Jan 03 2019 at 15:15, on Zulip):

do you want to reassign both to yourself?

will do

nagisa (Jan 03 2019 at 15:16, on Zulip):

Then after that there also are two more issues assigned to @pnkfelix:

nagisa (Jan 03 2019 at 15:16, on Zulip):

These have likely seen no activity/progress.

nagisa (Jan 03 2019 at 15:17, on Zulip):

The specification issue seems a broad documentation one, which does not seem to block any particular release

nikomatsakis (Jan 03 2019 at 15:17, on Zulip):

those are both kind of the same as the other 2PB issues

nagisa (Jan 03 2019 at 15:17, on Zulip):

but will probably impede work on NLL.

nikomatsakis (Jan 03 2019 at 15:17, on Zulip):

that is, the "rewrite one"

nikomatsakis (Jan 03 2019 at 15:17, on Zulip):

well, at least the first one

nikomatsakis (Jan 03 2019 at 15:17, on Zulip):

so I'll assign that to myself

nikomatsakis (Jan 03 2019 at 15:17, on Zulip):

I guess it already is

nikomatsakis (Jan 03 2019 at 15:18, on Zulip):

the #46901 .. yeah, let's leave that, I don't foresee having time for writing specs in the next week, but I think it's a good question how we will coordinate to make it happen

nagisa (Jan 03 2019 at 15:18, on Zulip):

(rust all-hands seems like a great time to brainstorm the coordination :slight_smile:)

nagisa (Jan 03 2019 at 15:19, on Zulip):

Right, so that’s it for the P-high issues.

nagisa (Jan 03 2019 at 15:19, on Zulip):

# beta-nominated

5 PRs to look at.

nagisa (Jan 03 2019 at 15:19, on Zulip):
nagisa (Jan 03 2019 at 15:20, on Zulip):

(I made emojis you can click on, :back: is for backport, :not_allowed: is for not backport)

nikomatsakis (Jan 03 2019 at 15:21, on Zulip):

the main question is: do we not backport because this already made it to stable? I feel like we might as well close the hole as soon as possible though.

nagisa (Jan 03 2019 at 15:21, on Zulip):

this seems like a fairly serious issue to me

Wesley Wiser (Jan 03 2019 at 15:22, on Zulip):

It's a very small patch and a high-priority issue. Makes sense to backport to me even if it's already hit stable.

nagisa (Jan 03 2019 at 15:22, on Zulip):

I would perhaps even lean towards backporting to stable, but a new stable is coming out fairly soon already.

nagisa (Jan 03 2019 at 15:22, on Zulip):

so there is little value in doing that.

nikomatsakis (Jan 03 2019 at 15:23, on Zulip):

yeah, I don't think it's worth it.

nikomatsakis (Jan 03 2019 at 15:23, on Zulip):

sounds like we should backport

nagisa (Jan 03 2019 at 15:23, on Zulip):

Next in the queue is:

nagisa (Jan 03 2019 at 15:24, on Zulip):

The PR is mostly tests.

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

looks pretty harmless

davidtwco (Jan 03 2019 at 15:24, on Zulip):

How do we know which way you vote @nagisa since you need to click on both options?

nagisa (Jan 03 2019 at 15:25, on Zulip):

@davidtwco I end up removing the other vote once I get to the next ticket.

nagisa (Jan 03 2019 at 15:26, on Zulip):

So it appears to not have been confirmed that the issue is a regression at all

nagisa (Jan 03 2019 at 15:26, on Zulip):

Ah the issue says this is a stable-to-stable regression

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

seems like a similar situation to the one before

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

i.e., problem discovered from Rust 2018

nagisa (Jan 03 2019 at 15:27, on Zulip):

Right, except the bug this fixes is not as critical.

Alexander Regueiro (Jan 03 2019 at 15:27, on Zulip):

@nagisa I’m pretty sure that was already backported no?

Alexander Regueiro (Jan 03 2019 at 15:27, on Zulip):

To beta at least

nagisa (Jan 03 2019 at 15:28, on Zulip):

@nagisa I’m pretty sure that was already backported no?

Are you talking about the #56850? Could you find the ticket with the backport?

nagisa (Jan 03 2019 at 15:28, on Zulip):

Seems like the consensus is to backport the PR at hand.

Wesley Wiser (Jan 03 2019 at 15:29, on Zulip):

I think this is the backport https://github.com/rust-lang/rust/pull/57236

Alexander Regueiro (Jan 03 2019 at 15:29, on Zulip):

I don’t know it sorry. The release team closed my backport PR ages ago, for some reason.

nagisa (Jan 03 2019 at 15:29, on Zulip):

Next is another one from @Vadim Petrochenkov (spoiler, there is more!)

Alexander Regueiro (Jan 03 2019 at 15:29, on Zulip):

Could test on playground though. I’m not on a computer however.

nagisa (Jan 03 2019 at 15:30, on Zulip):

@Wesley Wiser okay, thanks, I’ll note it on the issue and close it.

Alexander Regueiro (Jan 03 2019 at 15:30, on Zulip):

:+1:🏻

nagisa (Jan 03 2019 at 15:31, on Zulip):

So back to #57181: a small issue, mostly tests.

nagisa (Jan 03 2019 at 15:31, on Zulip):

A 2018 regression.

nagisa (Jan 03 2019 at 15:31, on Zulip):

just like the previous PR.

nagisa (Jan 03 2019 at 15:32, on Zulip):
nagisa (Jan 03 2019 at 15:33, on Zulip):

More of the same :slight_smile:

nikomatsakis (Jan 03 2019 at 15:34, on Zulip):

seems like this is to be expected, given that we landed such a major rewrite

nagisa (Jan 03 2019 at 15:34, on Zulip):

Well, I admit that I didn’t participate in testing 2018 edition at all, and I suspect that this is true for most.

nagisa (Jan 03 2019 at 15:35, on Zulip):

Final backport nomination is

nikomatsakis (Jan 03 2019 at 15:35, on Zulip):

Well, I admit that I didn’t participate in testing 2018 edition at all, and I suspect that this is true for most.

well this particular rewrite landed shortly before the stable release anyway :P

nagisa (Jan 03 2019 at 15:36, on Zulip):

now this is an interesting one as it technically solves a soundness issue

nagisa (Jan 03 2019 at 15:36, on Zulip):

(codegen issues pertaining to alignment are classified as such, usually)

nagisa (Jan 03 2019 at 15:36, on Zulip):

Although the worst that happens to our knowledge is that SPARC is crashing

nagisa (Jan 03 2019 at 15:37, on Zulip):

(added emojis, click away!)

nagisa (Jan 03 2019 at 15:39, on Zulip):

Okay, that’s it for the backports.

Pietro Albini (Jan 03 2019 at 15:39, on Zulip):

preparing the rollup right now

nagisa (Jan 03 2019 at 15:39, on Zulip):

# stable-to beta regressions

5 of them. 3 T-rustdoc. 1 is a P-high that has been looked at before. Which leaves...

nagisa (Jan 03 2019 at 15:40, on Zulip):

This is a fairly recently filled one, which explains lack of investiation :slight_smile:

nagisa (Jan 03 2019 at 15:40, on Zulip):

We need a P-flag.

nagisa (Jan 03 2019 at 15:40, on Zulip):

and possibly a volunteer to look at it.

nikomatsakis (Jan 03 2019 at 15:40, on Zulip):

Hmm. I can investigate that.

nikomatsakis (Jan 03 2019 at 15:41, on Zulip):

(Though if someone else wants to, that's fine with me too)

nagisa (Jan 03 2019 at 15:41, on Zulip):

What priority do you reckon we give it?

nagisa (Jan 03 2019 at 15:41, on Zulip):

P-high?

nikomatsakis (Jan 03 2019 at 15:41, on Zulip):

Seems like P-high

nikomatsakis (Jan 03 2019 at 15:41, on Zulip):

at least to bisect

nikomatsakis (Jan 03 2019 at 15:41, on Zulip):

ps @Pietro Albini have you ever considered a bisection bot? :)

nikomatsakis (Jan 03 2019 at 15:41, on Zulip):

dang but that would be cool

Pietro Albini (Jan 03 2019 at 15:42, on Zulip):

sure I have considered it, but I need someone to patch my clock to add a few extra hours :P

Pietro Albini (Jan 03 2019 at 15:42, on Zulip):

and we have too many infra nightmares right now

nikomatsakis (Jan 03 2019 at 15:42, on Zulip):

haha fair enough

nagisa (Jan 03 2019 at 15:42, on Zulip):

@nikomatsakis I found cargo bisect to work fairly nice already

nikomatsakis (Jan 03 2019 at 15:42, on Zulip):

as long as it's on a wishlist somewhere :)

nikomatsakis (Jan 03 2019 at 15:43, on Zulip):

do you mean the cargo-bisect-rustc tool that @simulacrum made, @nagisa ?

nagisa (Jan 03 2019 at 15:43, on Zulip):

being able to request it on github comments would be :star: though.

nagisa (Jan 03 2019 at 15:43, on Zulip):

yeah.

nikomatsakis (Jan 03 2019 at 15:43, on Zulip):

if so, I agree it works well, which is why we are tantalizingly close to a bot :)

nikomatsakis (Jan 03 2019 at 15:43, on Zulip):

although it always takes me a bit to figure out how to configure it again

nagisa (Jan 03 2019 at 15:43, on Zulip):

Okay, so that was it for the stable-to-beta regressions

Pietro Albini (Jan 03 2019 at 15:44, on Zulip):

@nagisa please set both the nominated and accepted labels when you accept something though

nagisa (Jan 03 2019 at 15:44, on Zulip):

# stable-to-nightly regressions

6 of them. 1 is a P-high and has been looked at before. This is also where my notepad ends for all but two more issues.

nagisa (Jan 03 2019 at 15:44, on Zulip):

@nagisa please set both the nominated and accepted labels when you accept something though

ah right...

nagisa (Jan 03 2019 at 15:44, on Zulip):

/me did a blunder.

nagisa (Jan 03 2019 at 15:45, on Zulip):
nagisa (Jan 03 2019 at 15:45, on Zulip):
nagisa (Jan 03 2019 at 15:45, on Zulip):

@Alex Crichton made a PR for this, it seems.

nagisa (Jan 03 2019 at 15:45, on Zulip):

(https://github.com/rust-lang/rust/pull/57287)

Pietro Albini (Jan 03 2019 at 15:45, on Zulip):

(when you have time, could someone check if https://github.com/rust-lang/rust/pull/56838 actually needs to be backported?)

nagisa (Jan 03 2019 at 15:46, on Zulip):

So the issue is in good hands.

nagisa (Jan 03 2019 at 15:47, on Zulip):

@Pietro Albini do you mean "needs" as in whether it has been backported yet or not?

Pietro Albini (Jan 03 2019 at 15:47, on Zulip):

it's backported, but the changed code doesn't exist on beta

varkor (Jan 03 2019 at 15:47, on Zulip):

should #56736 have been marked with T-infra?

nagisa (Jan 03 2019 at 15:47, on Zulip):

@Pietro Albini you want to ask for a backport PR then, instead of doing it yourself.

nagisa (Jan 03 2019 at 15:47, on Zulip):

@Pietro Albini and that’s T-rustdoc

Pietro Albini (Jan 03 2019 at 15:48, on Zulip):

@nagisa it changes only compiler files though

nagisa (Jan 03 2019 at 15:48, on Zulip):

@varkor it is not too obvious whether it falls onto infra hands or not either

nagisa (Jan 03 2019 at 15:48, on Zulip):

@varkor but I marked it as T-infra nevertheless

varkor (Jan 03 2019 at 15:49, on Zulip):

it seems good to pass the baton somewhere if it's a regression, so it can either be closed or fixed

Pietro Albini (Jan 03 2019 at 15:49, on Zulip):

(please also put the nomination when you tag infra, we don't check regressions on our meeting -- already did that for this)

nagisa (Jan 03 2019 at 15:49, on Zulip):

@Pietro Albini for the backport I would then ping relevant-seeming people on the ticket (git blame perhaps helps here?)

nagisa (Jan 03 2019 at 15:50, on Zulip):

Okay, next up is:

nagisa (Jan 03 2019 at 15:50, on Zulip):
nagisa (Jan 03 2019 at 15:50, on Zulip):

oh wait

nagisa (Jan 03 2019 at 15:50, on Zulip):

Here we go, fixed.

nagisa (Jan 03 2019 at 15:51, on Zulip):

Any P-flag suggestions?

nikomatsakis (Jan 03 2019 at 15:51, on Zulip):

P-high

nikomatsakis (Jan 03 2019 at 15:51, on Zulip):

probably due to some of the recent PRs in this area

nikomatsakis (Jan 03 2019 at 15:52, on Zulip):

not sure who to assign it to exactly

nikomatsakis (Jan 03 2019 at 15:52, on Zulip):

I'd volunteer but I don't know if I will realistically have time

nikomatsakis (Jan 03 2019 at 15:52, on Zulip):

probably a good idea to start with a bisection

nagisa (Jan 03 2019 at 15:53, on Zulip):

@Alexander Regueiro since you have been commenting on it, are you perhaps interested in looking into this?

nagisa (Jan 03 2019 at 15:53, on Zulip):

I could bisect this too, this seems fairly straightforward I guess.

nagisa (Jan 03 2019 at 15:54, on Zulip):

I’ll assign this to myself.

nikomatsakis (Jan 03 2019 at 15:54, on Zulip):

@nagisa cc me once you figure out the cause if you remember :)

nagisa (Jan 03 2019 at 15:54, on Zulip):

Next up is

Alexander Regueiro (Jan 03 2019 at 15:54, on Zulip):

@nagisa If you can pin down the commit that introduced it, I can work on a fix if you like.

Alexander Regueiro (Jan 03 2019 at 15:55, on Zulip):

Only that build times are agonisingly long on my machine, so I don’t want to bisect!

nagisa (Jan 03 2019 at 15:55, on Zulip):

@Taylor Cramer has identified a possible cause here.

nikomatsakis (Jan 03 2019 at 15:55, on Zulip):

@Taylor Cramer suspects #57264 is caused by a @Vadim Petrochenkov commit, maybe we should bisect it to be sure?

nikomatsakis (Jan 03 2019 at 15:55, on Zulip):

but we can also assign to @Vadim Petrochenkov I think

nikomatsakis (Jan 03 2019 at 15:55, on Zulip):

i'd call it P-high

nikomatsakis (Jan 03 2019 at 15:56, on Zulip):

(my default stance for a regression)

nikomatsakis (Jan 03 2019 at 15:56, on Zulip):

Only that build times are agonisingly long on my machine, so I don’t want to bisect!

(btw note that cargo-bisect-rustc downloads prebuilt binaries)

nikomatsakis (Jan 03 2019 at 15:56, on Zulip):

though I'm often amazed how long that takes regardless :)

nikomatsakis (Jan 03 2019 at 15:57, on Zulip):

I've also got the bisection thing up and going now so I can bisect #57264

nagisa (Jan 03 2019 at 15:57, on Zulip):

That makes it all regressions.

nikomatsakis (Jan 03 2019 at 15:57, on Zulip):

we're about out of time looks like

Alexander Regueiro (Jan 03 2019 at 15:57, on Zulip):

Speaking of builds, @centril has suggested SSH boxes for building available to T-compiler and possibly regular regular contributors, to speed things up. I liked the idea. Not sure if Mozilla could do this for us though. Anyway, don’t want to derail the conversation really. Maybe talk about it another time.

nagisa (Jan 03 2019 at 15:58, on Zulip):

@Alexander Regueiro there is a cool thing called GCC Farm.

nikomatsakis (Jan 03 2019 at 15:58, on Zulip):

it's been suggested before; would be nice to make something available

nikomatsakis (Jan 03 2019 at 15:58, on Zulip):

not sure what format it should take

Alexander Regueiro (Jan 03 2019 at 15:58, on Zulip):

Oh yeah? Not heard of it.

nikomatsakis (Jan 03 2019 at 15:58, on Zulip):

(hopefully it could be funded by more than just mozilla, too)

nikomatsakis (Jan 03 2019 at 15:58, on Zulip):

I wanted to call attention to the question of tomorrow's scheduled steering meeting

nikomatsakis (Jan 03 2019 at 15:58, on Zulip):

I posted this comment on internals -- not sure if enough people will be around or what

nagisa (Jan 03 2019 at 15:58, on Zulip):

We could achieve a lot by just donating some compute resources to that project.

nikomatsakis (Jan 03 2019 at 15:59, on Zulip):

(this meeting for example seemed a bit lightly attended)

Alexander Regueiro (Jan 03 2019 at 15:59, on Zulip):

@nikomatsakis Yeah. I don’t care who funds it, but if someone or multiple ones could, would be great.

nagisa (Jan 03 2019 at 16:00, on Zulip):

Right, so I want to call to attention that there is an ongoing FCP

nikomatsakis (Jan 03 2019 at 16:00, on Zulip):

ah, true, I hadn't noticed that somehow, must have missed it

nagisa (Jan 03 2019 at 16:01, on Zulip):

That’s it. Adjourned!

nikomatsakis (Jan 03 2019 at 16:01, on Zulip):

but I'm excited about that PR ( :tada: @Oli)

nikomatsakis (Jan 03 2019 at 16:01, on Zulip):

That’s it. Adjourned!

Thanks @nagisa for running the meeting

Alexander Regueiro (Jan 03 2019 at 16:01, on Zulip):

I have an idea for a build service that infra could run given a set of computing resources... if they have the time to manage it. I might even be able to write the software for it if there’s interest. Should I make an internals thread?

Alexander Regueiro (Jan 03 2019 at 16:02, on Zulip):

@nikomatsakis Cosmetics PR on topic for next meeting, if there aren’t more important things? :-)

nagisa (Jan 03 2019 at 16:03, on Zulip):

I have an idea for a build service that infra could run given a set of computing resources... if they have the time to manage it. I might even be able to write the software for it if there’s interest. Should I make an internals thread?

Sure! Although I strongly encourage you to play with GCC Farm as well. The accounts there are given out on-request to contributors of free (as in freedom) software, which rustc is.

nikomatsakis (Jan 03 2019 at 16:03, on Zulip):

@nikomatsakis Cosmetics PR on topic for next meeting, if there aren’t more important things? :-)

Yep! I was thinking that there were some nominated PRs we really should discuss, but I didn't raise it because it seemed like we didn't have that many folks anyway (plus out of time)

nagisa (Jan 03 2019 at 16:04, on Zulip):

Given the difficulties we’re having with procuring some non-standard non-x86_64 hardware for our CI, I would not expect much success in getting ssh boxes infrastructure going on our own.

nikomatsakis (Jan 03 2019 at 16:05, on Zulip):

(sharing the infra with GCC farm does seem appealing, if they'd be interested -- I'd not heard of that project before)

Last update: Nov 16 2019 at 01:50UTC