Stream: t-compiler/meetings

Topic: [weekly meeting] 2020-07-23 #54818


Santiago Pastorino (Jul 22 2020 at 16:22, on Zulip):

Hi @T-compiler/meeting; the triage meeting will happen tomorrow at 2pm UTC

Santiago Pastorino (Jul 22 2020 at 16:22, on Zulip):

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

Santiago Pastorino (Jul 22 2020 at 16:23, on Zulip):

@WG-prioritization have prepared the meeting agenda

Santiago Pastorino (Jul 22 2020 at 16:24, on Zulip):

We will have checkins from @T-compiler/WG-meta and @WG-mir-opt

Santiago Pastorino (Jul 22 2020 at 16:24, on Zulip):

@nikomatsakis do you have something you want to share about @T-compiler/WG-meta?

Santiago Pastorino (Jul 22 2020 at 16:24, on Zulip):

it could also be me :), but I'm not sure what the current state of the working group actually is

Santiago Pastorino (Jul 22 2020 at 16:25, on Zulip):

@oli do you have something you want to share about @WG-mir-opt?

Santiago Pastorino (Jul 22 2020 at 16:25, on Zulip):

feel free to paste whatever you have directly in the agenda :heart:

oli (Jul 23 2020 at 08:19, on Zulip):

I don't have edit permissions for the agenda

lcnr (Jul 23 2020 at 08:21, on Zulip):

Wanna send the summary to me and I put it into the agenda until you get permissions?

oli (Jul 23 2020 at 08:24, on Zulip):

first gotta write it up :laughing:

oli (Jul 23 2020 at 08:24, on Zulip):

I was going to write it up in the hackmd

oli (Jul 23 2020 at 12:39, on Zulip):

I should start tagging PRs instead of having to find them manually again every time...

oli (Jul 23 2020 at 12:39, on Zulip):

https://gist.github.com/oli-obk/1b3b3e2d851a43696927ac2b469d9eef is the summary for today

Santiago Pastorino (Jul 23 2020 at 13:12, on Zulip):

@oli and others that want write access to all the compiler HackMD documents, all I need is your HackMD usernames :)

Santiago Pastorino (Jul 23 2020 at 13:15, on Zulip):

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

Santiago Pastorino (Jul 23 2020 at 13:15, on Zulip):

Check out the meeting agenda

pnkfelix (Jul 23 2020 at 14:03, on Zulip):

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

pnkfelix (Jul 23 2020 at 14:03, on Zulip):

we will start off with 5 minutes for ...

Announcements

pnkfelix (Jul 23 2020 at 14:03, on Zulip):
pnkfelix (Jul 23 2020 at 14:04, on Zulip):
pnkfelix (Jul 23 2020 at 14:04, on Zulip):
pnkfelix (Jul 23 2020 at 14:04, on Zulip):
simulacrum (Jul 23 2020 at 14:05, on Zulip):
nikomatsakis (Jul 23 2020 at 14:05, on Zulip):
pnkfelix (Jul 23 2020 at 14:07, on Zulip):

Heh, I was going to note that we need a @T-compiler/WG-meta checkin-text from @nikomatsakis

pnkfelix (Jul 23 2020 at 14:08, on Zulip):

but based on the previous note, we might not be getting that today

pnkfelix (Jul 23 2020 at 14:08, on Zulip):

WG checkins

pnkfelix (Jul 23 2020 at 14:08, on Zulip):

@WG-mir-opt checkin by @oli:

Not sure if we already had this last time, but even if, this is worth repeating: we have a totally awesome NRVO pass!

pnkfelix (Jul 23 2020 at 14:10, on Zulip):

@nikomatsakis feel free, if you can, to add the checkin notes for @T-compiler/WG-meta to the agenda hackmd, and I'll try to post them at the end of the meeting if there is time

pnkfelix (Jul 23 2020 at 14:10, on Zulip):

Beta-nominations

T-compiler

pnkfelix (Jul 23 2020 at 14:10, on Zulip):
pnkfelix (Jul 23 2020 at 14:11, on Zulip):

(I assume the "P-high nightly regression" is in fact now a stable-to-beta regression...)

Wesley Wiser (Jul 23 2020 at 14:12, on Zulip):

Yes

Santiago Pastorino (Jul 23 2020 at 14:12, on Zulip):

right :)

pnkfelix (Jul 23 2020 at 14:13, on Zulip):

yeah; beta-accepted

Wesley Wiser (Jul 23 2020 at 14:13, on Zulip):

This PR missed the beta cutoff by a few hours

pnkfelix (Jul 23 2020 at 14:14, on Zulip):
pnkfelix (Jul 23 2020 at 14:15, on Zulip):

Okay beta-accepted

pnkfelix (Jul 23 2020 at 14:16, on Zulip):
pnkfelix (Jul 23 2020 at 14:17, on Zulip):

Oh, I guess we were never setting the pre-existing implicit_region_bound to anything but None?

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

Probably a historical artifact

pnkfelix (Jul 23 2020 at 14:18, on Zulip):

beta-accepted

pnkfelix (Jul 23 2020 at 14:18, on Zulip):
Santiago Pastorino (Jul 23 2020 at 14:18, on Zulip):

pnkfelix said:

note that is not yet merged

DPC (Jul 23 2020 at 14:19, on Zulip):

(it's being retried)

pnkfelix (Jul 23 2020 at 14:19, on Zulip):

yeah I'll probably want to bring up that detail more when we get to the stable-nom

pnkfelix (Jul 23 2020 at 14:20, on Zulip):

okay #74557 is beta-accepted

pnkfelix (Jul 23 2020 at 14:21, on Zulip):
simulacrum (Jul 23 2020 at 14:21, on Zulip):

Do we have higher confidence in this patch than before?

simulacrum (Jul 23 2020 at 14:21, on Zulip):

A similar patch was already reverted once

pnkfelix (Jul 23 2020 at 14:22, on Zulip):

@simulacrum sorry which PR is your comment in reference to? #74557, or #74650 ?

simulacrum (Jul 23 2020 at 14:22, on Zulip):

rust#74650

Esteban Küber (Jul 23 2020 at 14:22, on Zulip):

(not yet merged)

Esteban Küber (Jul 23 2020 at 14:22, on Zulip):

I believe it addresses the problems brought up

pnkfelix (Jul 23 2020 at 14:22, on Zulip):

Oh, @Esteban Küber , I was reading the title as literally just talking about empty blocks on the LHS of &&

Esteban Küber (Jul 23 2020 at 14:23, on Zulip):

Because it was only with && being both and and ref

pnkfelix (Jul 23 2020 at 14:23, on Zulip):

but this is the more general <block> && <bool>, right?

Esteban Küber (Jul 23 2020 at 14:23, on Zulip):

Not just empty blocks but any kind of block

Esteban Küber (Jul 23 2020 at 14:23, on Zulip):

Yes

Esteban Küber (Jul 23 2020 at 14:23, on Zulip):

And not just book, really (although bool is the case that is a regression)

simulacrum (Jul 23 2020 at 14:23, on Zulip):

if we have more confidence then I'm fine with it, just wanted to raise that this area does seem bug-prone

Esteban Küber (Jul 23 2020 at 14:24, on Zulip):

For backport we could just bring in a single line removal that makes everything work with worse diagnostics, most of the code in the PR is to regain the diagnostics we had before.

pnkfelix (Jul 23 2020 at 14:24, on Zulip):

the regression itself is a stable-to-stable one, introduced in 1.36, right? that's how I read the text of #74233

Esteban Küber (Jul 23 2020 at 14:24, on Zulip):

Correct

Esteban Küber (Jul 23 2020 at 14:25, on Zulip):

It's an old regression (introduced by me)

simulacrum (Jul 23 2020 at 14:25, on Zulip):

(that's part of why maybe not rushing with this is fine -- it's been a while since we introduced the bug)

pnkfelix (Jul 23 2020 at 14:25, on Zulip):

lets maybe let this ride the trains

pnkfelix (Jul 23 2020 at 14:25, on Zulip):

okay then, declining to beta-backport.

pnkfelix (Jul 23 2020 at 14:26, on Zulip):

(and likewise I'll treat that as a declining to stable-backport as well)

pnkfelix (Jul 23 2020 at 14:26, on Zulip):

libs-impl

T-rustdoc

pnkfelix (Jul 23 2020 at 14:26, on Zulip):

Stable-nominations

T-compiler

pnkfelix (Jul 23 2020 at 14:26, on Zulip):
pnkfelix (Jul 23 2020 at 14:27, on Zulip):

This is the one that hasn't merged yet

pnkfelix (Jul 23 2020 at 14:27, on Zulip):

so we hit a question of: How confident are we in this. :)

simulacrum (Jul 23 2020 at 14:28, on Zulip):

I will note that release is on Thursday next week and we can always delay it if something comes up. So we could provisionally approve and if regressions are noted not release with it (either delaying or just removing this PR from the release).

pnkfelix (Jul 23 2020 at 14:29, on Zulip):

@simulacrum will there be a crater run with this PR in it (on either stable or beta branch) before the stable release takes place?

simulacrum (Jul 23 2020 at 14:29, on Zulip):

No

simulacrum (Jul 23 2020 at 14:29, on Zulip):

(Or at least none was planned)

simulacrum (Jul 23 2020 at 14:29, on Zulip):

I guess if we p=1 a cargo check run it'd finish in time

pnkfelix (Jul 23 2020 at 14:29, on Zulip):

I'll admit that it may be a relative waste of resources

pnkfelix (Jul 23 2020 at 14:30, on Zulip):

My reading of the situation is that @nikomatsakis and @Matthew Jasper are confident that this doesn't introduce any soundness bug

pnkfelix (Jul 23 2020 at 14:30, on Zulip):

("this" meaning "this PR")

pnkfelix (Jul 23 2020 at 14:32, on Zulip):

I guess we can stable-approve it. I don't have any well-motivated objection apart from general distrust of landing something on stable that hasn't seen testing on the nightly channel

simulacrum (Jul 23 2020 at 14:32, on Zulip):

Yeah, my point was that we should have a week or so of testing

pnkfelix (Jul 23 2020 at 14:32, on Zulip):

okay

pnkfelix (Jul 23 2020 at 14:32, on Zulip):

Stable-accepted then

pnkfelix (Jul 23 2020 at 14:32, on Zulip):

and we'll all just try to keep our eye out for problems that could be related to this...

Wesley Wiser (Jul 23 2020 at 14:33, on Zulip):

simulacrum said:

I will note that release is on Thursday next week and we can always delay it if something comes up. So we could provisionally approve and if regressions are noted not release with it (either delaying or just removing this PR from the release).

Uh wait. What release?

pnkfelix (Jul 23 2020 at 14:33, on Zulip):

a point release

simulacrum (Jul 23 2020 at 14:33, on Zulip):

1.45.1 is being released next thursday by current planning

Wesley Wiser (Jul 23 2020 at 14:33, on Zulip):

Ah, got it

pnkfelix (Jul 23 2020 at 14:33, on Zulip):

this was part of annoucements, let me see

pnkfelix (Jul 23 2020 at 14:33, on Zulip):

yeah that

Wesley Wiser (Jul 23 2020 at 14:33, on Zulip):

Sorry

pnkfelix (Jul 23 2020 at 14:34, on Zulip):

okay moving on to rest of stable noms

pnkfelix (Jul 23 2020 at 14:34, on Zulip):
simulacrum (Jul 23 2020 at 14:35, on Zulip):

I will say that the ICE here doesn't look terrible (the error is still emitted)

simulacrum (Jul 23 2020 at 14:36, on Zulip):

so maybe worth being conservative and not backporting

simulacrum (Jul 23 2020 at 14:36, on Zulip):

But I'm fine either way

pnkfelix (Jul 23 2020 at 14:37, on Zulip):

the injection is only in the most recent release

nikomatsakis (Jul 23 2020 at 14:38, on Zulip):

(ok, I rebooted and did some other things, and I think my internet is working a bit better--- regarding rust#74590, I am fairly confident it's fine, but not like 99.999%, it's a subtle area of the code)

pnkfelix (Jul 23 2020 at 14:38, on Zulip):

so I guess I'm inclined to go ahead and stable-approve

simulacrum (Jul 23 2020 at 14:38, on Zulip):

seems good, yeah, since it's just in 1.45

pnkfelix (Jul 23 2020 at 14:38, on Zulip):

I did have a thought

pnkfelix (Jul 23 2020 at 14:39, on Zulip):

the PR #74557 has a comment, in the code where it adds the if guard on the match arm, where it says "... It will be handled as an error by the AST lowering"

pnkfelix (Jul 23 2020 at 14:39, on Zulip):

I'm trying to remember if delay_span_bug would handle a case like this well or not

pnkfelix (Jul 23 2020 at 14:40, on Zulip):

in terms of issuing an ICE if we don't issue an error?

pnkfelix (Jul 23 2020 at 14:40, on Zulip):

Am I misremembering? Does delay_span_bug not do that for this case?

simulacrum (Jul 23 2020 at 14:40, on Zulip):

(Something like that would increase confidence that we're not introducing a soundness bug at least)

Vadim Petrochenkov (Jul 23 2020 at 14:40, on Zulip):

resolve is a weird place to fix this, AST lowering would indeed look better

Vadim Petrochenkov (Jul 23 2020 at 14:41, on Zulip):

Also invalid binding @ ..s can occur in various other non-tuplestruct contexts.

Esteban Küber (Jul 23 2020 at 14:42, on Zulip):

@pnkfelix I believe you're correct

pnkfelix (Jul 23 2020 at 14:42, on Zulip):

I guess my question is: Does delay_span_bug do what I describe (issue an ICE if we get through static analysis without emitting an error diagnostic)? And if it doesn't, do we have something that does?

pnkfelix (Jul 23 2020 at 14:42, on Zulip):

the more I think about it the more I think delay_span_bug is for precisely this scenario, right? :)

pnkfelix (Jul 23 2020 at 14:43, on Zulip):

I'd be curious to know if we can add that to this PR in some fashion before we backport

Esteban Küber (Jul 23 2020 at 14:43, on Zulip):

I _think_ it isn't checked/emitted on every pass but only at the end, right?

Esteban Küber (Jul 23 2020 at 14:43, on Zulip):

Otherwise we have a couple of bugs in some places :grimacing:

pnkfelix (Jul 23 2020 at 14:43, on Zulip):

yes I think so.

pnkfelix (Jul 23 2020 at 14:44, on Zulip):

which is the right thing for this case, right?

pnkfelix (Jul 23 2020 at 14:44, on Zulip):

I.e. I mainly want to ensure that we don't end up having the compiler produce object code

Esteban Küber (Jul 23 2020 at 14:44, on Zulip):

If that matches is all needed to identify the case, yes

pnkfelix (Jul 23 2020 at 14:45, on Zulip):

well there is that sub.as_ref().filter(|p| p.is_rest()).is_some()) business too ...

pnkfelix (Jul 23 2020 at 14:45, on Zulip):

anyway

pnkfelix (Jul 23 2020 at 14:45, on Zulip):

I'm not saying I'd require the addition of a delay_span_bug

simulacrum (Jul 23 2020 at 14:46, on Zulip):

I'm happy to insert it if someone wants to either provide a patch or tell me where to put it

pnkfelix (Jul 23 2020 at 14:46, on Zulip):

so I'll just say "stable-approved, but if the person doing the backport, preferably the PR author, has the chance to make a variant with delay_span_bug added on, I'd prefer to see that in the backport.

pnkfelix (Jul 23 2020 at 14:46, on Zulip):

thanks @simulacrum . Let's follow-up on this after the meeting.

pnkfelix (Jul 23 2020 at 14:47, on Zulip):

finally, we already established that we're declining to backport "Correctly parse {} && false in tail expression" rust#74650

pnkfelix (Jul 23 2020 at 14:47, on Zulip):

libs-impl

T-rustdoc

pnkfelix (Jul 23 2020 at 14:47, on Zulip):

PRs S-waiting-on-team

T-compiler

libs-impl

pnkfelix (Jul 23 2020 at 14:48, on Zulip):

Issues of Note

Short Summary

pnkfelix (Jul 23 2020 at 14:48, on Zulip):

P-critical

T-compiler

pnkfelix (Jul 23 2020 at 14:48, on Zulip):
pnkfelix (Jul 23 2020 at 14:49, on Zulip):

(this is the one we were discussing up above. I'm going to assign it to @Matthew Jasper )

pnkfelix (Jul 23 2020 at 14:49, on Zulip):
pnkfelix (Jul 23 2020 at 14:50, on Zulip):

do we know if rust#74498 continues to reproduce when one uses a rust with LLVM 11?

pnkfelix (Jul 23 2020 at 14:50, on Zulip):

oh never mind

pnkfelix (Jul 23 2020 at 14:50, on Zulip):

(it reproduces on LLVM master)

pnkfelix (Jul 23 2020 at 14:51, on Zulip):

has a bug been filed against LLVM itself then?

pnkfelix (Jul 23 2020 at 14:52, on Zulip):

I'm assuming not

pnkfelix (Jul 23 2020 at 14:52, on Zulip):

I'll self-assign and try to make sure that a bug against LLVM master either exists or gets filed

Santiago Pastorino (Jul 23 2020 at 14:52, on Zulip):

I'm not sure neither, maybe we should ask in the issue?

pnkfelix (Jul 23 2020 at 14:53, on Zulip):

the other question is whether this can/should remain P-critical

pnkfelix (Jul 23 2020 at 14:53, on Zulip):

miscompilation is pretty nasty

pnkfelix (Jul 23 2020 at 14:53, on Zulip):

I'm willing to keep it at P-critical for now

pnkfelix (Jul 23 2020 at 14:53, on Zulip):

we'll see if we get annoyed with it being there over next few weeks

pnkfelix (Jul 23 2020 at 14:53, on Zulip):

ideal thing would be to push for a fix in LLVM in time for the LLVM 11 release

pnkfelix (Jul 23 2020 at 14:54, on Zulip):

so that we'd pull it in "naturally"

simulacrum (Jul 23 2020 at 14:54, on Zulip):

We could presumably fix it on our end by passing that LLVM arg someone posted?

simulacrum (Jul 23 2020 at 14:54, on Zulip):

(perhaps only on MSVC)

Santiago Pastorino (Jul 23 2020 at 14:54, on Zulip):

simulacrum said:

(perhaps only on MSVC)

it is MSVC only?

pnkfelix (Jul 23 2020 at 14:54, on Zulip):

oh interesting

pnkfelix (Jul 23 2020 at 14:54, on Zulip):

the known rust repro is MSVC only

pnkfelix (Jul 23 2020 at 14:54, on Zulip):

but there is evidence that it may occur on other targets

simulacrum (Jul 23 2020 at 14:54, on Zulip):

doesn't on mingw, at least

pnkfelix (Jul 23 2020 at 14:54, on Zulip):

it depends on whether anyone might generate LLVM IR that was posted

Santiago Pastorino (Jul 23 2020 at 14:55, on Zulip):

right

Santiago Pastorino (Jul 23 2020 at 14:55, on Zulip):

that's why I had doubts about being MSVC only

pnkfelix (Jul 23 2020 at 14:55, on Zulip):

@simulacrum there was a follow-up comment from AlyoshaVasilieva that says that the LLVM flag didn't fix it for them

simulacrum (Jul 23 2020 at 14:55, on Zulip):

ah, okay

pnkfelix (Jul 23 2020 at 14:55, on Zulip):

anyway I will follow up on this

pnkfelix (Jul 23 2020 at 14:56, on Zulip):

libs-impl

T-rustdoc

pnkfelix (Jul 23 2020 at 14:56, on Zulip):

Unassigned P-high regressions

Beta regressions

Nightly regressions

pnkfelix (Jul 23 2020 at 14:56, on Zulip):
pnkfelix (Jul 23 2020 at 14:57, on Zulip):

@nikomatsakis this is apparently due to removal of leak-check?

pnkfelix (Jul 23 2020 at 14:57, on Zulip):

can I assign this to you, @nikomatsakis , for further investigation?

pnkfelix (Jul 23 2020 at 14:58, on Zulip):

(removal of leak check was PR #73643)

pnkfelix (Jul 23 2020 at 14:59, on Zulip):

I'm going to go ahead and assign to @nikomatsakis , I think they are best person to take point on regressions here, at least for now. They can delegate if necessary.

pnkfelix (Jul 23 2020 at 14:59, on Zulip):

Performance logs

Triage done by njn

This week was a disaster, perf-wise. 28 revisions checked. 7 regressions, several of them ranging from large to huge, many in rollups. Some additional regressions may have occurred in rollups that were masked by other regressions/improvements. 3 improvements, one of which was a reversion of a regression. Thanks for Mark-Simulacrum and eddyb for follow-up measurements and adding new tooling to help investigate regressions in rollups. A follow-up thread on Zulip is here

In better news, rustdoc performance is now being benchmarked, thanks to the
efforts of Joshua Nelson.

Triage done by njn. Revision range: 9d09331e00b02f81c714b0c41ce3a38380dd36a2..71384101ea3b030b80f7def80a37f67e148518b0.

Regressions

Improvements

pnkfelix (Jul 23 2020 at 14:59, on Zulip):

hmm

pnkfelix (Jul 23 2020 at 14:59, on Zulip):

I was worried about that

pnkfelix (Jul 23 2020 at 15:00, on Zulip):

(the Performance Log info is so long that it was condensed by Zulip into one message that is cut off with a [More ...])

pnkfelix (Jul 23 2020 at 15:00, on Zulip):

let me try again

pnkfelix (Jul 23 2020 at 15:00, on Zulip):

Performance logs

Triage done by njn

This week was a disaster, perf-wise. 28 revisions checked. 7 regressions, several of them ranging from large to huge, many in rollups. Some additional regressions may have occurred in rollups that were masked by other regressions/improvements. 3 improvements, one of which was a reversion of a regression. Thanks for Mark-Simulacrum and eddyb for follow-up measurements and adding new tooling to help investigate regressions in rollups. A follow-up thread on Zulip is here

In better news, rustdoc performance is now being benchmarked, thanks to the
efforts of Joshua Nelson.

pnkfelix (Jul 23 2020 at 15:00, on Zulip):

Triage done by njn. Revision range: 9d09331e00b02f81c714b0c41ce3a38380dd36a2..71384101ea3b030b80f7def80a37f67e148518b0.

pnkfelix (Jul 23 2020 at 15:01, on Zulip):

Regressions

pnkfelix (Jul 23 2020 at 15:01, on Zulip):

Improvements

pnkfelix (Jul 23 2020 at 15:02, on Zulip):

so rollup #74461 caused big time losses (up to 73.0%), and PR #74416 was identified as culprit, but the reversion of #74416 in #74478 was in another rollup that then "only" provided 33.5% improvements

simulacrum (Jul 23 2020 at 15:02, on Zulip):

we're trying to track that down still

pnkfelix (Jul 23 2020 at 15:03, on Zulip):

/me remembers/realizes that they prefer scaling ratios over percentages to ease comparsions like this

simulacrum (Jul 23 2020 at 15:03, on Zulip):

(see https://github.com/rust-lang/rust/pull/74678)

pnkfelix (Jul 23 2020 at 15:03, on Zulip):

i.e. saying 0.25x slower can be better than 75% slower because then the inverse is clearly 4x faster

pnkfelix (Jul 23 2020 at 15:04, on Zulip):

vs ... 300% faster, right? depends on what you choose for denominator

pnkfelix (Jul 23 2020 at 15:04, on Zulip):

anyway

pnkfelix (Jul 23 2020 at 15:05, on Zulip):

this is sort of terrible because this was super important thing to look into but we are out of time

pnkfelix (Jul 23 2020 at 15:05, on Zulip):

but I take it that people are actively looking into it

simulacrum (Jul 23 2020 at 15:05, on Zulip):

I think we reverted everything that was a regression at this point

simulacrum (Jul 23 2020 at 15:06, on Zulip):

(except for the uncertainty around rollup #74461)

pnkfelix (Jul 23 2020 at 15:06, on Zulip):

just to wrap up meeting, there were two nominated issues, but they can wait until next week. I'll cut and paste them here but like I said< i think they can wait

pnkfelix (Jul 23 2020 at 15:06, on Zulip):

Nominated Issues

T-compiler

libs-impl

Santiago Pastorino (Jul 23 2020 at 15:06, on Zulip):

pnkfelix said:

this is sort of terrible because this was super important thing to look into but we are out of time

yeah it was important to discuss, maybe we should keep around a reference to this info for next meeting?

pnkfelix (Jul 23 2020 at 15:06, on Zulip):

right, there are two things

pnkfelix (Jul 23 2020 at 15:07, on Zulip):
  1. lets make sure we discuss this weeks performance issues in next week's meeting
pnkfelix (Jul 23 2020 at 15:07, on Zulip):

my hope is that we will see amazing performance wins over this week

pnkfelix (Jul 23 2020 at 15:07, on Zulip):

and I will just say "well lets not pat ourselves on back too easily"

pnkfelix (Jul 23 2020 at 15:07, on Zulip):

but also

pnkfelix (Jul 23 2020 at 15:07, on Zulip):
  1. When performance triage results are this troubling, probably would be good to at least include a one or two line summary in the announcements
pnkfelix (Jul 23 2020 at 15:08, on Zulip):

to ensure that we won't let it get missed if we happen to get bogged down in nominations

pnkfelix (Jul 23 2020 at 15:08, on Zulip):

okay that's all folks!

pnkfelix (Jul 23 2020 at 15:08, on Zulip):

thanks for attending, everyone in @T-compiler/meeting ! Stay safe, stay distant, but don't be a stranger! :mask:

pnkfelix (Jul 23 2020 at 15:09, on Zulip):

:dancer: :ruler: :ruler: :ruler: :ruler: :ruler: :ruler: :dancing:

lcnr (Jul 23 2020 at 15:09, on Zulip):

1 inch rulers? so a total of ~1.5 meters :laughing: nice

Last update: Nov 25 2020 at 02:45UTC