Stream: t-compiler/meetings

Topic: [weekly meeting] 2020-09-17 #54818


Santiago Pastorino (Sep 16 2020 at 19:38, on Zulip):

Hi @T-compiler/meeting; the triage meeting will happen tomorrow at

Santiago Pastorino (Sep 16 2020 at 19:38, on Zulip):

The @WG-prioritization have done pre-triage in #t-compiler/wg-prioritization/alerts

Santiago Pastorino (Sep 16 2020 at 19:38, on Zulip):

@WG-prioritization have prepared the meeting agenda

Santiago Pastorino (Sep 16 2020 at 19:40, on Zulip):

We will have checkins from @WG-polymorphization and @WG-polonius which both were pre-filled by @davidtwco @eddyb @lcnr @nikomatsakis and @lqd

Santiago Pastorino (Sep 16 2020 at 19:41, on Zulip):

they've also followed the pre-set formatting, well done :tada:

Santiago Pastorino (Sep 17 2020 at 13:15, on Zulip):

Hi @T-compiler/meeting, triage meeting will be starting in ~ 45 minutes, exactly at

Santiago Pastorino (Sep 17 2020 at 13:15, on Zulip):

Check out the meeting agenda

pnkfelix (Sep 17 2020 at 14:01, on Zulip):

Hi @T-compiler/meeting! Add a :wave: emoji to show you're here :)

pnkfelix (Sep 17 2020 at 14:03, on Zulip):

we will start off with 5 minutes for ...

Announcements

pnkfelix (Sep 17 2020 at 14:03, on Zulip):
pnkfelix (Sep 17 2020 at 14:04, on Zulip):
nikomatsakis (Sep 17 2020 at 14:05, on Zulip):

Reminder that there is a survey

nikomatsakis (Sep 17 2020 at 14:05, on Zulip):

Also, I think I will spend some time going over the data the hour before

nikomatsakis (Sep 17 2020 at 14:05, on Zulip):

Probably in a public zulip stream :)

nikomatsakis (Sep 17 2020 at 14:05, on Zulip):

survey link

nikomatsakis (Sep 17 2020 at 14:05, on Zulip):

we've had 7 responses so far

pnkfelix (Sep 17 2020 at 14:06, on Zulip):
nikomatsakis (Sep 17 2020 at 14:06, on Zulip):

I realize I never responded :)

pnkfelix (Sep 17 2020 at 14:06, on Zulip):
pnkfelix (Sep 17 2020 at 14:07, on Zulip):

(some members had to leave the WG for various reasons, so I figured I'd mention we're happy to have more help)

pnkfelix (Sep 17 2020 at 14:07, on Zulip):
pnkfelix (Sep 17 2020 at 14:08, on Zulip):

/me wonders if the "move compiler/crates to stable Rust" is worthy of being promoted to a Friday steering meeting...

pnkfelix (Sep 17 2020 at 14:09, on Zulip):
pnkfelix (Sep 17 2020 at 14:09, on Zulip):
pnkfelix (Sep 17 2020 at 14:09, on Zulip):
pnkfelix (Sep 17 2020 at 14:09, on Zulip):
nikomatsakis (Sep 17 2020 at 14:10, on Zulip):

pnkfelix said:

/me wonders if the "move compiler/crates to stable Rust" is worthy of being promoted to a Friday steering meeting...

I was just looking at the thread and there are certainly a lot of comments; it feels ripe for a "summary"

Santiago Pastorino (Sep 17 2020 at 14:11, on Zulip):

pnkfelix said:

we've added other's teams Finalized FCPs, let us know after the meeting if this is interesting to be informed or if we should stick just to T-compiler

Santiago Pastorino (Sep 17 2020 at 14:11, on Zulip):

in general there are not that many, unsure what happened that we had 6 :)

pnkfelix (Sep 17 2020 at 14:11, on Zulip):

maybe T-libs had a meeting

pnkfelix (Sep 17 2020 at 14:12, on Zulip):

WG checkins

Polymorphization working group checkin by @davidtwco @eddyb @lcnr

I think the initial goal ought to be getting polymorphization enabled by default:

Joshua Nelson (Sep 17 2020 at 14:12, on Zulip):

@ecstatic-morse and @Christiaan Dirkx are working on making lots of functions const

pnkfelix (Sep 17 2020 at 14:12, on Zulip):

We've currently got the following open PRs:

Joshua Nelson (Sep 17 2020 at 14:12, on Zulip):

That's what all the T-libs FCPs are

pnkfelix (Sep 17 2020 at 14:13, on Zulip):

We've also got the following issues open:

I think that's everything. I think that landing the current PRs and addressing #75325 are the most important tasks right now, we can crater after that and hopefully try enabling polymorphization by default.

pnkfelix (Sep 17 2020 at 14:13, on Zulip):

(we probably need checkins to be a little more succinct, somehow)

Wesley Wiser (Sep 17 2020 at 14:13, on Zulip):

This is really awesome, great work @davidtwco @eddyb @lcnr!! :tada:

nikomatsakis (Sep 17 2020 at 14:14, on Zulip):

great summary, too

davidtwco (Sep 17 2020 at 14:14, on Zulip):

I didn’t write that intended for a meeting checkin, just happened to reuse it as one.

pnkfelix (Sep 17 2020 at 14:15, on Zulip):

@WG-polonius checkin by @nikomatsakis and @lqd

The polonius WG (Niko, Albin, and Rémy) had a sprint at the beginning of August, where:
- we continued the work on move errors: both in rustc to prepare the data needed to do the analysis and actually emit the errors computed by polonius, and in polonius itself, to fix potential false positives, and current efficiency issues in this part of the computation. Some new rules have been written but not yet fully implemented or tested, so this is something we'll continue to look at in the next sprint.
- we were able to prototype our different lifetime analysis, and evaluate it: there was an open question whether it'd be better to track subset relations or equality relations, and we were able to find some subtle flaws in our current thoughts about tracking equality. There's some interesting theoretical core that could be useful in the future, but also has foundational issues, so we decided to keep tracking subset relations for now and not change our existing lifetime analysis rules.
- we continued the work on universal region errors / illegal subset errors: we settled a tiny open question on whether we should compute these errors using subset relations or by looking for unexpected loans in universal regions. We decided to switch to using subset relations because it works everywhere: the polonius analysis is currently split into different parts for validation and efficiency reasons. Our existing subset errors computation had only been added to the main rules used for validation, and it used the "looking at unexpected loans" approach. We did the work necessary to update this, and also to compute the errors in the other parts dedicated to efficiency.
- we were able to work on a fix for the current OOMs in fact generation we see in a couple rustc UI tests: we have prototyped a fix in rustc and polonius. It is, however, quite invasive and requires subtle duplication of rules. So we also started looking at a different fix which could avoid that, and is something we'll continue to look at in the next sprint.
- we did our periodic reevaluation of rustc UI tests using the polonius compare mode: find which tests might have changed .stderr files, add the ones missing for the tests added recently, etc; since the compare mode is not run on CI yet, we do this manually every few months to keep up to date with diagnostics changes and the likes. This time, a new test OOMs for move errors and will be fixed as part of this work mentioned earlier.
- finally, we also looked at supporting profiling polonius' internals from rustc's -Zself-profile, and also when using the tracy profiler.

pnkfelix (Sep 17 2020 at 14:16, on Zulip):

"current OOMs in fact generation we see" ?

pnkfelix (Sep 17 2020 at 14:16, on Zulip):

(can you fix that in the agenda hackmd, niko?)

pnkfelix (Sep 17 2020 at 14:16, on Zulip):

anyway lets move along

pnkfelix (Sep 17 2020 at 14:17, on Zulip):

Beta-nominations

T-compiler

libs-impl

T-rustdoc

pnkfelix (Sep 17 2020 at 14:17, on Zulip):
pnkfelix (Sep 17 2020 at 14:18, on Zulip):

Note: #76529 is not a bug, its a PR

nikomatsakis (Sep 17 2020 at 14:18, on Zulip):

do we know it fixes those issues?

pnkfelix (Sep 17 2020 at 14:19, on Zulip):

it has landed, so presumably we could look at the generated docs of today...?

nikomatsakis (Sep 17 2020 at 14:19, on Zulip):

in any case, the diff looks a bit larger than it really is because it cleaned up the code a bit

nikomatsakis (Sep 17 2020 at 14:19, on Zulip):

it seems pretty safe

nikomatsakis (Sep 17 2020 at 14:19, on Zulip):

i.e., from what I can tell the only real addition is the final hunk

Esteban Küber (Sep 17 2020 at 14:21, on Zulip):

Theres another if condition that changed slightly.

pnkfelix (Sep 17 2020 at 14:21, on Zulip):

funny that it got rid of a comment

nikomatsakis (Sep 17 2020 at 14:22, on Zulip):

Esteban Küber said:

Theres another if condition that changed slightly.

oh I guess that's true

pnkfelix (Sep 17 2020 at 14:22, on Zulip):

anyway it seems fine to me

nikomatsakis (Sep 17 2020 at 14:22, on Zulip):

it only applies to unstable rustc_private features now, not all unstable features

Joshua Nelson (Sep 17 2020 at 14:22, on Zulip):

I wrote a summary of this on the PR, one sec

nikomatsakis (Sep 17 2020 at 14:22, on Zulip):

in any case yeah I'm just judging on "likely to cause surprising breakage", I guess I'm assuming the "intent" is good from the original r+

Joshua Nelson (Sep 17 2020 at 14:23, on Zulip):

https://github.com/rust-lang/rust/pull/76571#issuecomment-690855610

Joshua Nelson (Sep 17 2020 at 14:23, on Zulip):

(although maybe this should be a T-rustdoc discussion? I don't know the process)

pnkfelix (Sep 17 2020 at 14:23, on Zulip):

T-rustdoc has basically delegated beta-nomination handling to T-compiler

pnkfelix (Sep 17 2020 at 14:24, on Zulip):

that summary from @Joshua Nelson gives the option of just backporting PR #76529

Esteban Küber (Sep 17 2020 at 14:24, on Zulip):

That last pr is smaller...

Joshua Nelson (Sep 17 2020 at 14:24, on Zulip):

note that PR is not on nightly

nikomatsakis (Sep 17 2020 at 14:25, on Zulip):

I vote for "just backport" :) seems like @Joshua Nelson and @GuillaumeGomez were both inclined to do so, and I judge the risk relatively low. At worst we get too many docs or a few are missing, right?

Joshua Nelson (Sep 17 2020 at 14:25, on Zulip):

correct

nikomatsakis (Sep 17 2020 at 14:25, on Zulip):

(In particular around unstable things)

Esteban Küber (Sep 17 2020 at 14:25, on Zulip):

Fair enough

pnkfelix (Sep 17 2020 at 14:25, on Zulip):

Is there any risk of it claiming a trait is implemented that is in fact not?

pnkfelix (Sep 17 2020 at 14:25, on Zulip):

seems impossible

Joshua Nelson (Sep 17 2020 at 14:25, on Zulip):

no, this is just determining whether to show the trait impl

nikomatsakis (Sep 17 2020 at 14:26, on Zulip):

I don't think so, esp. since trait impls are stable

Joshua Nelson (Sep 17 2020 at 14:26, on Zulip):

the implementation already has to exist

nikomatsakis (Sep 17 2020 at 14:26, on Zulip):

(the trait may of course be unstable)

pnkfelix (Sep 17 2020 at 14:26, on Zulip):

Stable-nominations

T-compiler

libs-impl

T-rustdoc

pnkfelix (Sep 17 2020 at 14:26, on Zulip):

PRs S-waiting-on-team

T-compiler

pnkfelix (Sep 17 2020 at 14:27, on Zulip):
pnkfelix (Sep 17 2020 at 14:27, on Zulip):

wasn't @simulacrum going to do something here...?

simulacrum (Sep 17 2020 at 14:27, on Zulip):

yes there's a PR assigned to you

pnkfelix (Sep 17 2020 at 14:27, on Zulip):

yes, see PR #76607

pnkfelix (Sep 17 2020 at 14:27, on Zulip):

ah ha

Santiago Pastorino (Sep 17 2020 at 14:27, on Zulip):

btw, it seemed to me that this is not waiting on team anymore and incorrectly labelled?

pnkfelix (Sep 17 2020 at 14:28, on Zulip):

yeah I agree with @Santiago Pastorino

pnkfelix (Sep 17 2020 at 14:28, on Zulip):

okay lets move along then

pnkfelix (Sep 17 2020 at 14:28, on Zulip):
Joshua Nelson (Sep 17 2020 at 14:29, on Zulip):

(deleted) (got this confused with https://github.com/rust-lang/rust/pull/75671)

pnkfelix (Sep 17 2020 at 14:29, on Zulip):

I'd say this is blocked on the aforementioned MCP

pnkfelix (Sep 17 2020 at 14:30, on Zulip):

I'll just second it, I think I agree with the intent here (of MCP#347 that is)

pnkfelix (Sep 17 2020 at 14:30, on Zulip):

so we can move on

pnkfelix (Sep 17 2020 at 14:30, on Zulip):

libs-impl

pnkfelix (Sep 17 2020 at 14:31, on Zulip):

Issues of Note

Short Summary

pnkfelix (Sep 17 2020 at 14:31, on Zulip):

P-critical

T-compiler

pnkfelix (Sep 17 2020 at 14:31, on Zulip):
Santiago Pastorino (Sep 17 2020 at 14:31, on Zulip):

one important thing to note, there have been like 5 P-critical issues related to LLVM 11

pnkfelix (Sep 17 2020 at 14:31, on Zulip):

so do we just need to cherry-pick the fix to our LLVM?

Santiago Pastorino (Sep 17 2020 at 14:32, on Zulip):

some were fixed and some are still up

pnkfelix (Sep 17 2020 at 14:32, on Zulip):

sounds doable.

Santiago Pastorino (Sep 17 2020 at 14:32, on Zulip):

pnkfelix said:

so do we just need to cherry-pick the fix to our LLVM?

it seems that it will be included in 11.0

nikomatsakis (Sep 17 2020 at 14:32, on Zulip):

when will 11.0 be released?

Santiago Pastorino (Sep 17 2020 at 14:32, on Zulip):

saw some comments saying that they were a bit delayed already

Santiago Pastorino (Sep 17 2020 at 14:33, on Zulip):

in another P-critical we will see next :)

Santiago Pastorino (Sep 17 2020 at 14:33, on Zulip):

but I'm not 100% sure when they will release or when they're supposed to release

pnkfelix (Sep 17 2020 at 14:33, on Zulip):
Santiago Pastorino (Sep 17 2020 at 14:34, on Zulip):

that one yeah, we would want to cherry-pick to our llvm branch

Santiago Pastorino (Sep 17 2020 at 14:34, on Zulip):

unless they approve and include in 11.0

Santiago Pastorino (Sep 17 2020 at 14:34, on Zulip):

all this seems a bit risky

nikomatsakis (Sep 17 2020 at 14:34, on Zulip):

it seems like cherry picking would be good

nikomatsakis (Sep 17 2020 at 14:35, on Zulip):

I mean in general

nikomatsakis (Sep 17 2020 at 14:35, on Zulip):

both for the previous and this one, potentially

Santiago Pastorino (Sep 17 2020 at 14:35, on Zulip):

there are a bunch of P-critical and I'd say, we're just seeing the ones that were found, there may be more undiscovered yet

nikomatsakis (Sep 17 2020 at 14:35, on Zulip):

I guess I can't quite judge how risky an LLVM PR is

nikomatsakis (Sep 17 2020 at 14:35, on Zulip):

but waiting for LLVM 11 could be a long wait

nikomatsakis (Sep 17 2020 at 14:35, on Zulip):

we've often cherry-picked things in advance in the past, afaik

simulacrum (Sep 17 2020 at 14:35, on Zulip):

LLVM upgrades with just cherry-picks are usually pretty safe, I think

simulacrum (Sep 17 2020 at 14:35, on Zulip):

but LLVM 11 is always going to fix (and not fix) some bugs

Santiago Pastorino (Sep 17 2020 at 14:36, on Zulip):

what I'm referring about risky is that after LLVM 11 upgrade there were a lot of P-critical issues

Santiago Pastorino (Sep 17 2020 at 14:36, on Zulip):

and probably more to discover

nikomatsakis (Sep 17 2020 at 14:36, on Zulip):

ok, I see. I think that's also not unusual. :)

simulacrum (Sep 17 2020 at 14:36, on Zulip):

yeah, that's my point

simulacrum (Sep 17 2020 at 14:36, on Zulip):

like we can delay it for a cycle but I don't personally expect that to "stop"

pnkfelix (Sep 17 2020 at 14:36, on Zulip):

I would like to just point out that there's a difference between cherry-picking something that has landed on LLVM upstream, vs cherry-picking something one of our volunteers wrote that has not yet landed on LLVM upstream

pnkfelix (Sep 17 2020 at 14:36, on Zulip):

I personally feel pretty confident about landing code that LLVM itself has signed off on

Santiago Pastorino (Sep 17 2020 at 14:37, on Zulip):

is it skipping one stable cycle an option, so we don't have to bake LLVM out? :P

Aaron Hill (Sep 17 2020 at 14:37, on Zulip):

I'm not very familiar with LLVM, so it's certainly possible that I missed something

simulacrum (Sep 17 2020 at 14:37, on Zulip):

we can revert the upgrade in beta, I suppose

simulacrum (Sep 17 2020 at 14:37, on Zulip):

but I personally do not feel like that's an obvious win

pnkfelix (Sep 17 2020 at 14:37, on Zulip):

can/should we decide this here?

Aaron Hill (Sep 17 2020 at 14:37, on Zulip):

If we do cherry-pick before meeting, I'd like someone more familiar with LLVM to review it first

simulacrum (Sep 17 2020 at 14:37, on Zulip):

@pnkfelix we've never had a good process for these decisions tbh

pnkfelix (Sep 17 2020 at 14:37, on Zulip):

@simulacrum when would you say is deadline for deciding on whether to revert LLVM 11 from beta?

simulacrum (Sep 17 2020 at 14:38, on Zulip):

I would probably say roughly 1.5 weeks would be good

pnkfelix (Sep 17 2020 at 14:38, on Zulip):

1.5 weeks from today?

simulacrum (Sep 17 2020 at 14:38, on Zulip):

yeah

pnkfelix (Sep 17 2020 at 14:38, on Zulip):

okay

nikomatsakis (Sep 17 2020 at 14:38, on Zulip):

Yeah, so -- I think we should only cherry-pick things that LLVM has reviewed unless very urgent

simulacrum (Sep 17 2020 at 14:38, on Zulip):

(that's very just "glanced at a calendar and subtracted a bit over a week from release day)

pnkfelix (Sep 17 2020 at 14:38, on Zulip):

That to me means that we should make sure we discuss this again, at next week's meeting at lhe latest.

pnkfelix (Sep 17 2020 at 14:39, on Zulip):

(or next friday maybe; I don't remember if we have a steering meeting already scheduled for next friday...)

simulacrum (Sep 17 2020 at 14:39, on Zulip):

my personal opinion is that discussing seems very vague -- at least, for me, I don't have any basis on which to make this call either way

Santiago Pastorino (Sep 17 2020 at 14:39, on Zulip):

pnkfelix said:

(or next friday maybe; I don't remember if we have a steering meeting already scheduled for next friday...)

was thinking the same but there's a planning meeting yeah

pnkfelix (Sep 17 2020 at 14:39, on Zulip):

oh next friday is a planning meeting, right

pnkfelix (Sep 17 2020 at 14:39, on Zulip):

okay we might be able to steal from time from that if necessary

Wesley Wiser (Sep 17 2020 at 14:40, on Zulip):

Is not a regression, this never worked so it was also discussed if this was P-critical or P-high

If this never worked, I'm not sure this is critical enough to revert the LLVM 11 upgrade.

pnkfelix (Sep 17 2020 at 14:40, on Zulip):

the job in the meantime should be to brainstorm if we can figure out what data we would need to gather to make the decision

Aaron Hill (Sep 17 2020 at 14:40, on Zulip):

Also, I think this may only affect fat pointers

Aaron Hill (Sep 17 2020 at 14:40, on Zulip):

I don't know the details of how we lower things to LLVM types/return values, but this bug only affects things that LLVM considers to be 'multiple return values'

pnkfelix (Sep 17 2020 at 14:41, on Zulip):

Wesley Wiser said:

Is not a regression, this never worked so it was also discussed if this was P-critical or P-high

If this never worked, I'm not sure this is critical enough to revert the LLVM 11 upgrade.

yeah, my musings about backing out LLVM upgrade is more about @Santiago Pastorino 's concerns about how many critical bugs have shown up

Aaron Hill (Sep 17 2020 at 14:41, on Zulip):

which I believe does not include named structs. I don't know how it relates to our tuple types

Santiago Pastorino (Sep 17 2020 at 14:41, on Zulip):

@Wesley Wiser yeah, maybe this is not the best example, but there are more issues

pnkfelix (Sep 17 2020 at 14:41, on Zulip):

I do actually think that rust#76387 seems like its "under control"

Wesley Wiser (Sep 17 2020 at 14:41, on Zulip):

Ok, that makes sense

pnkfelix (Sep 17 2020 at 14:41, on Zulip):

we may want to see if we can prod the LLVM-devs to review the patch upstream

pnkfelix (Sep 17 2020 at 14:42, on Zulip):

if only to give us confidence about a backport here.

pnkfelix (Sep 17 2020 at 14:42, on Zulip):

lets move along

Santiago Pastorino (Sep 17 2020 at 14:42, on Zulip):

my fear is that if in 2 months 5 critical issues were discovered, how many more there really are that we may hit live

Santiago Pastorino (Sep 17 2020 at 14:42, on Zulip):

but I'm just guessing and being conservative in this case

nikomatsakis (Sep 17 2020 at 14:43, on Zulip):

I guess I feel wary of a big change like backing out LLVM upgrade

nikomatsakis (Sep 17 2020 at 14:43, on Zulip):

that it will cause more disruption than it fixes

pnkfelix (Sep 17 2020 at 14:43, on Zulip):

I suspect something we must do before making a decision about a revert is to make a summary of every known bug that we know a revert would re-inject

pnkfelix (Sep 17 2020 at 14:43, on Zulip):

(which sort of gets at @nikomatsakis 's point)

pnkfelix (Sep 17 2020 at 14:43, on Zulip):

hopefully including performance issues as well as soundness ones...

pnkfelix (Sep 17 2020 at 14:44, on Zulip):

but okay, lets move along

pnkfelix (Sep 17 2020 at 14:44, on Zulip):
Santiago Pastorino (Sep 17 2020 at 14:44, on Zulip):

pnkfelix said:

I suspect something we must do before making a decision about a revert is to make a summary of every known bug that we know a revert would re-inject

and also a list of the issues that've just shown up as P-critical after release

pnkfelix (Sep 17 2020 at 14:44, on Zulip):

we did decide what to do about #76479 last week, didn't we?

Santiago Pastorino (Sep 17 2020 at 14:45, on Zulip):

what I can clearly tell is that after last meeting nothing new has happened

pnkfelix (Sep 17 2020 at 14:45, on Zulip):

to at least alpha-rename the new method?

pnkfelix (Sep 17 2020 at 14:45, on Zulip):

(to something harder to hit)

pnkfelix (Sep 17 2020 at 14:46, on Zulip):

I'm trying to remember if we were proposing something more robust than that in the end ...

Santiago Pastorino (Sep 17 2020 at 14:46, on Zulip):

yes, I remember that

Santiago Pastorino (Sep 17 2020 at 14:46, on Zulip):

Santiago Pastorino said:

what I can clearly tell is that after last meeting nothing new has happened

I meant, that I don't think there's much to discuss given there are no news

pnkfelix (Sep 17 2020 at 14:47, on Zulip):

found it: https://rust-lang.zulipchat.com/#narrow/stream/238009-t-compiler.2Fmeetings/topic/.5Bweekly.20meeting.5D.202020-09-10.20.2354818/near/209659769

pnkfelix (Sep 17 2020 at 14:47, on Zulip):

I know I've skipped lang team meetings ( dealing with child care and job search stuff at moment)

pnkfelix (Sep 17 2020 at 14:47, on Zulip):

so I myself didn't bring it up there

pnkfelix (Sep 17 2020 at 14:47, on Zulip):

but renaming method to something more obscure was what we settled on as short term fix

pnkfelix (Sep 17 2020 at 14:48, on Zulip):

I'll assume @Matthew Jasper doesn't have time. It seems like something straight-forward, i.e. good for mentorship

pnkfelix (Sep 17 2020 at 14:48, on Zulip):

So I'll see about writing up an issue with mentorship instructions, see if that gets us anywhere?

pnkfelix (Sep 17 2020 at 14:49, on Zulip):

lets keep going thnen

pnkfelix (Sep 17 2020 at 14:49, on Zulip):

(oh and I'll self-assign to do those instructions)

pnkfelix (Sep 17 2020 at 14:49, on Zulip):
pnkfelix (Sep 17 2020 at 14:50, on Zulip):

wow

pnkfelix (Sep 17 2020 at 14:50, on Zulip):

well sounds like its being handled by @Wesley Wiser

pnkfelix (Sep 17 2020 at 14:50, on Zulip):

libs-impl

pnkfelix (Sep 17 2020 at 14:50, on Zulip):

sounds under control too

pnkfelix (Sep 17 2020 at 14:50, on Zulip):

T-rustdoc

pnkfelix (Sep 17 2020 at 14:50, on Zulip):

Unassigned P-high regressions

Beta regressions

pnkfelix (Sep 17 2020 at 14:50, on Zulip):

Nightly regressions

pnkfelix (Sep 17 2020 at 14:51, on Zulip):
pnkfelix (Sep 17 2020 at 14:51, on Zulip):

hmm yet another LLVM 11 issue (maybe)

pnkfelix (Sep 17 2020 at 14:52, on Zulip):

not tagged with the LLVM ice breakers though?

pnkfelix (Sep 17 2020 at 14:52, on Zulip):

(because windows ice breakers is more focused? not sure.)

pnkfelix (Sep 17 2020 at 14:52, on Zulip):

lets keep going

pnkfelix (Sep 17 2020 at 14:52, on Zulip):

Performance logs

Triage done by @simulacrum .
Revision range: 35fc8359868e65a0970094f648ba87fce634e8c7..a53f449516f23486d2dfd4e5685d4e869e8591d9

0 Regressions, 2 Improvements, 0 of them in rollups.

pnkfelix (Sep 17 2020 at 14:52, on Zulip):

Regressions

No regressions noted.

Santiago Pastorino (Sep 17 2020 at 14:52, on Zulip):

I've pinged windows first but yeah, may worth at some point to ping llvm crew too

pnkfelix (Sep 17 2020 at 14:53, on Zulip):

Improvements

BTreeMap mutable iterators should not take any reference to visited nodes during iteration #73971

make ConstEvaluatable more strict #74595

pnkfelix (Sep 17 2020 at 14:53, on Zulip):

wow

pnkfelix (Sep 17 2020 at 14:53, on Zulip):

very cool

pnkfelix (Sep 17 2020 at 14:53, on Zulip):

Nags requiring follow up

None

simulacrum (Sep 17 2020 at 14:53, on Zulip):

we noted last night that there's actually a regression, https://perf.rust-lang.org/compare.html?start=a055c5a1bd95e029e9b31891db63b6dc8258b472&end=7402a394471a6738a40fea7d4f1891666e5a80c5&stat=wall-time

simulacrum (Sep 17 2020 at 14:53, on Zulip):

I've yet to follow-up, been busy this morning

pnkfelix (Sep 17 2020 at 14:53, on Zulip):

hmm

pnkfelix (Sep 17 2020 at 14:53, on Zulip):

okay

pnkfelix (Sep 17 2020 at 14:53, on Zulip):

Nominated Issues

T-compiler

simulacrum (Sep 17 2020 at 14:53, on Zulip):

but an interesting case of "wall times are not matching instructions"

pnkfelix (Sep 17 2020 at 14:54, on Zulip):
pnkfelix (Sep 17 2020 at 14:54, on Zulip):

(this is " RFC: Promote aarch64-unknown-linux-gnu to a Tier-1 Rust target" #2959 )

pnkfelix (Sep 17 2020 at 14:55, on Zulip):

this is nominated because we need to keep our eye on it, decide what to do?

pnkfelix (Sep 17 2020 at 14:55, on Zulip):

i.e. decide when to actually suggest fcp merge?

Pietro Albini (Sep 17 2020 at 14:56, on Zulip):

hey

Pietro Albini (Sep 17 2020 at 14:56, on Zulip):

like, whether to block the FCP or not is something y'all should decide

Pietro Albini (Sep 17 2020 at 14:56, on Zulip):

I guess we could just open the FCP and any of you can add a concern

pnkfelix (Sep 17 2020 at 14:56, on Zulip):

yes I think that is best

nikomatsakis (Sep 17 2020 at 14:56, on Zulip):

I would be ready to move to FCP

pnkfelix (Sep 17 2020 at 14:56, on Zulip):

lets keep going

pnkfelix (Sep 17 2020 at 14:57, on Zulip):
simulacrum (Sep 17 2020 at 14:57, on Zulip):

this is looking for help from t-compiler, doesn't need discussion sync though

pnkfelix (Sep 17 2020 at 14:57, on Zulip):

okay. so just looking for help

pnkfelix (Sep 17 2020 at 14:57, on Zulip):

lets keep going

pnkfelix (Sep 17 2020 at 14:57, on Zulip):

(probably need a decidated section for "looking for help!!!")

pnkfelix (Sep 17 2020 at 14:58, on Zulip):
pnkfelix (Sep 17 2020 at 14:58, on Zulip):

I take it progress wasn't made?

Esteban Küber (Sep 17 2020 at 14:58, on Zulip):

I had a mitigation change, but didn't post the pr

Esteban Küber (Sep 17 2020 at 14:59, on Zulip):

Wanted to make something better but lacked the time

pnkfelix (Sep 17 2020 at 14:59, on Zulip):

okay

pnkfelix (Sep 17 2020 at 14:59, on Zulip):

but you do have something?

pnkfelix (Sep 17 2020 at 14:59, on Zulip):

I'm happy to move along in that case

Esteban Küber (Sep 17 2020 at 14:59, on Zulip):

If we're ok with something along the lines of this https://github.com/rust-lang/rust/issues/75791#issuecomment-686842188 in stable, I can get that up quickly

pnkfelix (Sep 17 2020 at 15:00, on Zulip):

hmm

Esteban Küber (Sep 17 2020 at 15:00, on Zulip):

It's not ideal, but again, a mitigation that mentions the lifetimes

Esteban Küber (Sep 17 2020 at 15:00, on Zulip):

The ticket would remain open, imo

pnkfelix (Sep 17 2020 at 15:00, on Zulip):

yeah, @nikomatsakis any thoughts on thaT?

pnkfelix (Sep 17 2020 at 15:00, on Zulip):

I think we're willing to live with '1 and '2 in the output

pnkfelix (Sep 17 2020 at 15:01, on Zulip):

if it helps people get some hint as to what's going on

pnkfelix (Sep 17 2020 at 15:01, on Zulip):
pnkfelix (Sep 17 2020 at 15:01, on Zulip):

(actually lets leave this for another meeting, we're already over time)

pnkfelix (Sep 17 2020 at 15:02, on Zulip):

we already discussed rust#76387 above I think

pnkfelix (Sep 17 2020 at 15:02, on Zulip):

finally

pnkfelix (Sep 17 2020 at 15:02, on Zulip):
pnkfelix (Sep 17 2020 at 15:02, on Zulip):

hmm

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

don't really have time for that either I think

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

but maybe another "need help" case anyway, not ready for sync discussion?

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

libs-impl

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

okay so that's everything

Esteban Küber (Sep 17 2020 at 15:03, on Zulip):

Yeah, the position is "this is expected regression", but we should have team sign off on that

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

We obviously left a few nominated issues undiscussed

pnkfelix (Sep 17 2020 at 15:05, on Zulip):

seems like a notice regarding #76480 is most realistic outcome?

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

ah but its already in release notes? Need to check that

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

i.e. look if its solely that PR is mentioned, or if known expected breakage is also mentioned.

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

(release notes are #76101)

Aaron Hill (Sep 17 2020 at 15:07, on Zulip):

Unfortunately, a compiler warning would be extremely difficult to implement

Aaron Hill (Sep 17 2020 at 15:08, on Zulip):

if you mean a mention in the release notes, that sounds like a good idea

pnkfelix (Sep 17 2020 at 15:08, on Zulip):

yeah, that's all I meant by notice

pnkfelix (Sep 17 2020 at 15:08, on Zulip):

I understand a warning cycle is not tenable here

pnkfelix (Sep 17 2020 at 15:08, on Zulip):

okay I need to go. THank you to everyone in @T-compiler/meeting for attending! :wave:

Santiago Pastorino (Sep 17 2020 at 15:10, on Zulip):

next week we will have checkins from @WG-prioritization and @WG-rfc-2229

Santiago Pastorino (Sep 17 2020 at 15:10, on Zulip):

I can provide the one about prioritization, would like to discuss with @pnkfelix and @nikomatsakis if it's worth to continue doing a checkin given that things has settled

Santiago Pastorino (Sep 17 2020 at 15:11, on Zulip):

@nikomatsakis @blitzerr can any of you provide a checkin for @WG-rfc-2229 ?

Santiago Pastorino (Sep 17 2020 at 15:11, on Zulip):

next week's agenda will live here https://hackmd.io/rtJNgRt3Sz2N5jX8c2-m3A?both

blitzerr (Sep 17 2020 at 15:13, on Zulip):

Hi @Santiago Pastorino ,

I am no longer working on rfc-2229.

Aman Arora (Sep 17 2020 at 17:03, on Zulip):

Santiago Pastorino said:

nikomatsakis blitzerr can any of you provide a checkin for @WG-rfc-2229 ?

Feel free to ping me for 2229, going forward :)

Santiago Pastorino (Sep 17 2020 at 17:04, on Zulip):

Aman Arora said:

Santiago Pastorino said:

nikomatsakis blitzerr can any of you provide a checkin for @WG-rfc-2229 ?

Feel free to ping me for 2229, going forward :)

do you want to provide a PR for compiler-team adding you and maybe removing @blitzerr ?

Santiago Pastorino (Sep 17 2020 at 17:04, on Zulip):

or a PR in general updating the group would be nice

Aman Arora (Sep 17 2020 at 17:06, on Zulip):

Can you point me to where this is, I can do that. I thought when we started the project , the rust-lang/team repo was updated
https://github.com/rust-lang/team/pull/354/files

lqd (Sep 17 2020 at 17:10, on Zulip):

in the compiler-team repo, here

Aman Arora (Sep 17 2020 at 17:15, on Zulip):

thanks :)

Santiago Pastorino (Sep 17 2020 at 17:17, on Zulip):

there are some groups that are pointing directly to the corresponding file on the team repo

Santiago Pastorino (Sep 17 2020 at 17:18, on Zulip):

maybe we should make a PR that modifies all the groups to do that

Aman Arora (Sep 17 2020 at 17:19, on Zulip):

Santiago Pastorino said:

maybe we should make a PR that modifies all the groups to do that

From what I understand the governance page for compiler team renders the team repo, might just be easier to point there

nvm, they don't have labels to each grp to hyperlink there.

apiraino (Sep 17 2020 at 17:32, on Zulip):

apropos, one more small thing. when we prepare the compiler team agenda we look at the WG checkins of the week and ping people in various WG groups.
Sometimes I get the feeling that some of these pages might not be updated
https://rust-lang.github.io/compiler-team/#meeting-calendar

apiraino (Sep 17 2020 at 17:32, on Zulip):

I'm offering a bit of my time to help update them (in case this is something useful)

Last update: Nov 25 2020 at 02:30UTC