Stream: t-compiler/wg-prioritization

Topic: pre-meeting triage 2020-04-02 #54818


Santiago Pastorino (Apr 01 2020 at 16:09, on Zulip):

@WG-prioritization we have our prioritization meeting in ~2hs

Santiago Pastorino (Apr 01 2020 at 16:11, on Zulip):

I've gave another look at our procedure document (pre-triage section) and organized a bit better

Wesley Wiser (Apr 01 2020 at 18:04, on Zulip):

:wave:

centril (Apr 01 2020 at 18:06, on Zulip):

:wave:

Paul Faria (Apr 01 2020 at 18:06, on Zulip):

:wave:

Santiago Pastorino (Apr 01 2020 at 18:07, on Zulip):

hi, I'm a little bit late

Santiago Pastorino (Apr 01 2020 at 18:15, on Zulip):

The @WG-prioritization will be doing pre-triage in this channel

Santiago Pastorino (Apr 01 2020 at 18:19, on Zulip):

so let's get started

Santiago Pastorino (Apr 01 2020 at 18:20, on Zulip):

have created the agenda

Santiago Pastorino (Apr 01 2020 at 18:20, on Zulip):

Unprioritized I-nominated

Santiago Pastorino (Apr 01 2020 at 18:21, on Zulip):

Go over unprioritized I-nominated issues. Nominations are sometimes an implicit request for prioritization.

Santiago Pastorino (Apr 01 2020 at 18:21, on Zulip):
  1. No team assigned
    - Add T-compiler tag when it corresponds.
Santiago Pastorino (Apr 01 2020 at 18:21, on Zulip):

Empty

Santiago Pastorino (Apr 01 2020 at 18:22, on Zulip):
  1. T-compiler
    - Prioritize issues and remove nomination of the ones not worth discussing.
    - Tag regressions accordingly.
    - Ping appropriate people and/or ICE-breakers.
Santiago Pastorino (Apr 01 2020 at 18:22, on Zulip):

we have 4

Santiago Pastorino (Apr 01 2020 at 18:22, on Zulip):

unpri nom 1/4: "Panic: non-eager expansion without a parent scope" #70624

Santiago Pastorino (Apr 01 2020 at 18:23, on Zulip):

this is an ICE, we can ping cleanup ICE breakers

centril (Apr 01 2020 at 18:23, on Zulip):

petrochenkov notes this might be a dupe

Santiago Pastorino (Apr 01 2020 at 18:23, on Zulip):

also possibly a dup of #55414

centril (Apr 01 2020 at 18:24, on Zulip):

and the dupe has p-high

Santiago Pastorino (Apr 01 2020 at 18:24, on Zulip):

so we can p-high this one too

Santiago Pastorino (Apr 01 2020 at 18:24, on Zulip):

and let me ping ICE breakers

centril (Apr 01 2020 at 18:24, on Zulip):

I'm skeptical this should be P-high though

centril (Apr 01 2020 at 18:24, on Zulip):

https://github.com/rust-lang/rust/issues/55414#issuecomment-554005412 basically requires language design and is unlikely to be fixed soon

Wesley Wiser (Apr 01 2020 at 18:25, on Zulip):

P-high from Aug 2019 doesn't effectively seem to be high priority

centril (Apr 01 2020 at 18:25, on Zulip):

So I would P-medium the other issue this is a dupe of as well

Santiago Pastorino (Apr 01 2020 at 18:26, on Zulip):

yeah makes sense

Santiago Pastorino (Apr 01 2020 at 18:26, on Zulip):

This was discussed briefly on Zulip.

Removing I-nominated, prioritizing as P-medium.

@rustbot ping cleanup

Would be nice if we figure out if this is a dupe or not.

DPC (Apr 01 2020 at 18:26, on Zulip):

do you watn to keep the E-mentor on?

Santiago Pastorino (Apr 01 2020 at 18:28, on Zulip):

done with the priority of those

Santiago Pastorino (Apr 01 2020 at 18:29, on Zulip):

unpri nom 2/4: " internal compiler error: cannot relate region: LUB(ReErased, ReErased)" #70608

Santiago Pastorino (Apr 01 2020 at 18:30, on Zulip):

another ICE and stable to nightly regression

Santiago Pastorino (Apr 01 2020 at 18:30, on Zulip):

this one is not assigned

Santiago Pastorino (Apr 01 2020 at 18:30, on Zulip):

found by fuzzing but hmm

centril (Apr 01 2020 at 18:31, on Zulip):

input doesn't seem crazy to me

Santiago Pastorino (Apr 01 2020 at 18:31, on Zulip):

yeah doesn't seem crazy

centril (Apr 01 2020 at 18:31, on Zulip):

but this is diagnostics recovery code

Santiago Pastorino (Apr 01 2020 at 18:32, on Zulip):

and also the errors says

Santiago Pastorino (Apr 01 2020 at 18:32, on Zulip):
1 | const e = & & 39;
  |       ^ help: provide a type for the item: `e: &&i32`
centril (Apr 01 2020 at 18:32, on Zulip):

I lean towards P-high if only cause it might reveal borrowck issues

Santiago Pastorino (Apr 01 2020 at 18:32, on Zulip):

so there's a hint

centril (Apr 01 2020 at 18:32, on Zulip):

yeah I wrote that diagnostics code :P

Santiago Pastorino (Apr 01 2020 at 18:32, on Zulip):

:)

Santiago Pastorino (Apr 01 2020 at 18:33, on Zulip):

let's go with P-high then?

Santiago Pastorino (Apr 01 2020 at 18:33, on Zulip):

I'm fine with that

centril (Apr 01 2020 at 18:33, on Zulip):

Guess it will get fixed if @Matthew Jasper has the time to look at it :D

centril (Apr 01 2020 at 18:33, on Zulip):

and likely not otherwise

Santiago Pastorino (Apr 01 2020 at 18:35, on Zulip):

tagged as P-high, didn't remove nomination because it's an unassigned regression

Santiago Pastorino (Apr 01 2020 at 18:35, on Zulip):

unpri nom 3/4: " regression - temporary value dropped while borrowed with static slice (as used by rust-phf)" #70584

centril (Apr 01 2020 at 18:35, on Zulip):

If you are bold you could assign matthewjasper :D

Santiago Pastorino (Apr 01 2020 at 18:36, on Zulip):

let's wait for @Matthew Jasper to say something :)

Santiago Pastorino (Apr 01 2020 at 18:36, on Zulip):

this one is a stable to stable regression

Santiago Pastorino (Apr 01 2020 at 18:37, on Zulip):

assigned to oli

centril (Apr 01 2020 at 18:37, on Zulip):

let's P-high and unnominate

Santiago Pastorino (Apr 01 2020 at 18:38, on Zulip):

:+1:

Santiago Pastorino (Apr 01 2020 at 18:39, on Zulip):

done

Santiago Pastorino (Apr 01 2020 at 18:39, on Zulip):

unpri nom 4/4: " Should enum discriminants have generics in scope? " #70453

Santiago Pastorino (Apr 01 2020 at 18:39, on Zulip):

it's using #![feature(const_generics)]

centril (Apr 01 2020 at 18:40, on Zulip):

I don't think this is meant for prioritization

Santiago Pastorino (Apr 01 2020 at 18:40, on Zulip):

tagged with T-lang and T-compiler

centril (Apr 01 2020 at 18:40, on Zulip):

it's not a bug to be fixed

Santiago Pastorino (Apr 01 2020 at 18:40, on Zulip):

yeah, that's what I was going to say

centril (Apr 01 2020 at 18:40, on Zulip):

and more likely this is a question for T-lang

Santiago Pastorino (Apr 01 2020 at 18:41, on Zulip):

centril said:

it's not a bug to be fixed

well, it's an ICE I guess you meant that @eddyb didn't nominate the issue to be prioritized but rather to discuss it?

centril (Apr 01 2020 at 18:42, on Zulip):

oh it is an ICE

centril (Apr 01 2020 at 18:42, on Zulip):

let's add that label

Santiago Pastorino (Apr 01 2020 at 18:42, on Zulip):

added

Santiago Pastorino (Apr 01 2020 at 18:42, on Zulip):

oh you added it too

Santiago Pastorino (Apr 01 2020 at 18:43, on Zulip):

so ... should we prioritize this?

centril (Apr 01 2020 at 18:43, on Zulip):

medium

Santiago Pastorino (Apr 01 2020 at 18:43, on Zulip):

anyway if we were doing it I'd go with P-medium

centril (Apr 01 2020 at 18:43, on Zulip):

requires nightly

Santiago Pastorino (Apr 01 2020 at 18:43, on Zulip):

yeah

Santiago Pastorino (Apr 01 2020 at 18:43, on Zulip):

exactly

centril (Apr 01 2020 at 18:43, on Zulip):

keep nomination though

centril (Apr 01 2020 at 18:43, on Zulip):

for t-lang

Santiago Pastorino (Apr 01 2020 at 18:45, on Zulip):

yeah, kept the nomination and added a comment so @eddyb can be more explicit that the nomination intention was to discuss

Santiago Pastorino (Apr 01 2020 at 18:46, on Zulip):

we should split I-nominated into two things, one with let's discuss meaning and the other with the let's prioritize this one meaning

Santiago Pastorino (Apr 01 2020 at 18:46, on Zulip):

well, done with this part

Santiago Pastorino (Apr 01 2020 at 18:46, on Zulip):
  1. All teams
    - No action required. It's nice to look at this for comparison.
Santiago Pastorino (Apr 01 2020 at 18:47, on Zulip):

in case someone wants to look at that for comparison purposes

Santiago Pastorino (Apr 01 2020 at 18:47, on Zulip):
  1. Unnominate leftover I-nominated
    - Remove I-nominated tag from issues discussed on the last meeting.
centril (Apr 01 2020 at 18:47, on Zulip):

@Santiago Pastorino easiest would be to keep I-nominated for "discuss" and I-prioritize for the other

Santiago Pastorino (Apr 01 2020 at 18:47, on Zulip):

yes, this is what I'd do

centril (Apr 01 2020 at 18:48, on Zulip):

maybe add that to the agenda

Santiago Pastorino (Apr 01 2020 at 18:48, on Zulip):

and it's already in the Plan, we can discuss next monday

Santiago Pastorino (Apr 01 2020 at 18:48, on Zulip):

you meant to the weekly meeting agenda?

Santiago Pastorino (Apr 01 2020 at 18:48, on Zulip):

or we should just discussing during the wg meeting this monday?

pnkfelix (Apr 01 2020 at 18:49, on Zulip):

I suspect @centril meant add it to the agenda for all of T-compiler

centril (Apr 01 2020 at 18:49, on Zulip):

@pnkfelix anything works for me :slight_smile:

pnkfelix (Apr 01 2020 at 18:49, on Zulip):

for tomorrows meeting. Just to let people know "Hey we are planning to add this new label. Here is what it is for."

centril (Apr 01 2020 at 18:49, on Zulip):

mostly some place it won't be forgotten

Santiago Pastorino (Apr 01 2020 at 18:50, on Zulip):

:+1:

Santiago Pastorino (Apr 01 2020 at 18:50, on Zulip):

Santiago Pastorino said:

  1. Unnominate leftover I-nominated
    - Remove I-nominated tag from issues discussed on the last meeting.

about this

Santiago Pastorino (Apr 01 2020 at 18:51, on Zulip):
Santiago Pastorino (Apr 01 2020 at 18:51, on Zulip):

those two are the ones that were on last meeting agenda

Santiago Pastorino (Apr 01 2020 at 18:51, on Zulip):

don't remember what we concluding about those

pnkfelix (Apr 01 2020 at 18:51, on Zulip):

I just skimmed those

pnkfelix (Apr 01 2020 at 18:51, on Zulip):

#69718 can be unnominated

pnkfelix (Apr 01 2020 at 18:52, on Zulip):

(it was discussed last week and action has been taken on it)

Santiago Pastorino (Apr 01 2020 at 18:53, on Zulip):

:+1:

Santiago Pastorino (Apr 01 2020 at 18:53, on Zulip):

should we keep #68304 ?

pnkfelix (Apr 01 2020 at 18:53, on Zulip):

we didn't have much time to discuss #68304 last week

pnkfelix (Apr 01 2020 at 18:53, on Zulip):

but I did make this note

centril (Apr 01 2020 at 18:53, on Zulip):

seems important to discuss

pnkfelix (Apr 01 2020 at 18:53, on Zulip):

I think we can let follow-up discussion on #68304 move forward on either a Zulip topic, or on the issue itself

Santiago Pastorino (Apr 01 2020 at 18:54, on Zulip):

let's keep it then

pnkfelix (Apr 01 2020 at 18:54, on Zulip):

so at the very least, the people present were notified

Santiago Pastorino (Apr 01 2020 at 18:54, on Zulip):

Regressions

Santiago Pastorino (Apr 01 2020 at 18:54, on Zulip):
  1. Beta regressions without P-label
    - Prioritize.
    - Ping appropriate people and/or ICE-breakers.
    - Assign if possible; if it remains unassigned, add it to agenda so we can assign during the meeting.
centril (Apr 01 2020 at 18:55, on Zulip):

@pnkfelix it would be good to keep momentum on fixing the issue though

centril (Apr 01 2020 at 18:55, on Zulip):

that's mainly why I want to keep it nominated

centril (Apr 01 2020 at 18:55, on Zulip):

otherwise no one may say anything and nothing happens

Santiago Pastorino (Apr 01 2020 at 18:55, on Zulip):

there's one beta regression without P-label

centril (Apr 01 2020 at 18:55, on Zulip):

(diffusion of responsibility)

Santiago Pastorino (Apr 01 2020 at 18:55, on Zulip):

beta regression 1/1: " regression: local ambiguity: multiple parsing options: built-in NTs lifetime " #70446

Santiago Pastorino (Apr 01 2020 at 18:56, on Zulip):

we have the culprit #69384

Santiago Pastorino (Apr 01 2020 at 18:56, on Zulip):

it's assigned to @Vadim Petrochenkov

centril (Apr 01 2020 at 18:57, on Zulip):

Not even sure it's a bug heh

pnkfelix (Apr 01 2020 at 18:57, on Zulip):

centril said:

otherwise no one may say anything and nothing happens

(A problem is that this approach doesn't scale to many bugs. We cannot afford to adopt this attitude for anything that is not P-critical, in my opinion.)

centril (Apr 01 2020 at 18:57, on Zulip):

@pnkfelix yeah but it's a soundness hole

pnkfelix (Apr 01 2020 at 18:57, on Zulip):

So do you think we should be revisiting every soundness hole every week?

pnkfelix (Apr 01 2020 at 18:58, on Zulip):

I'm actually not opposed

centril (Apr 01 2020 at 18:58, on Zulip):

It's a recent one

pnkfelix (Apr 01 2020 at 18:58, on Zulip):

(not immediately, at least)

centril (Apr 01 2020 at 18:58, on Zulip):

but it wouldn't be the worst idea :slight_smile:

centril (Apr 01 2020 at 18:58, on Zulip):

(so the potential for fixing it is greater if it is recent)

Santiago Pastorino (Apr 01 2020 at 18:58, on Zulip):

centril said:

Not even sure it's a bug heh

I guess that gives a low priority according to you?

centril (Apr 01 2020 at 18:59, on Zulip):

@Santiago Pastorino not necessarily; it just means I don't know what fixing it means

centril (Apr 01 2020 at 18:59, on Zulip):

but petrochenkov hopefully will know

Santiago Pastorino (Apr 01 2020 at 19:00, on Zulip):

should we leave it unprioritized or we can start with a priority and change it later?

centril (Apr 01 2020 at 19:01, on Zulip):

sure, maybe p-high?

Santiago Pastorino (Apr 01 2020 at 19:02, on Zulip):

yeah

Santiago Pastorino (Apr 01 2020 at 19:04, on Zulip):
  1. Nightly regressions without P-label
    - Prioritize.
    - Ping appropriate people and/or ICE-breakers.
    - Assign if possible; if it remains unassigned, add it to agenda so we can assign during the meeting.
Santiago Pastorino (Apr 01 2020 at 19:04, on Zulip):

Empty

Santiago Pastorino (Apr 01 2020 at 19:04, on Zulip):

Beta nominations

Santiago Pastorino (Apr 01 2020 at 19:04, on Zulip):
  1. No team assigned
    - Add T-compiler tag when it corresponds.
Santiago Pastorino (Apr 01 2020 at 19:04, on Zulip):

Empty

Santiago Pastorino (Apr 01 2020 at 19:05, on Zulip):
  1. All teams
    - Add T-compiler tag when it corresponds.
Santiago Pastorino (Apr 01 2020 at 19:06, on Zulip):

labels seems fine, I'd say

centril (Apr 01 2020 at 19:06, on Zulip):

@pnkfelix I feel we already declined https://github.com/rust-lang/rust/pull/70369 ?

centril (Apr 01 2020 at 19:06, on Zulip):

and accepted some others

Santiago Pastorino (Apr 01 2020 at 19:07, on Zulip):

well there's #70018

centril (Apr 01 2020 at 19:07, on Zulip):

did we forget to apply beta-accepted labels and remove beta-nominated on declined ones?

Santiago Pastorino (Apr 01 2020 at 19:07, on Zulip):

which is T-lib

Santiago Pastorino (Apr 01 2020 at 19:07, on Zulip):

centril said:

did we forget to apply beta-accepted labels and remove beta-nominated on declined ones?

I think we need to check last meeting list just in case

Santiago Pastorino (Apr 01 2020 at 19:07, on Zulip):

let me see

centril (Apr 01 2020 at 19:07, on Zulip):

might as well add t-compiler on #70018

centril (Apr 01 2020 at 19:07, on Zulip):

and its trivial

Santiago Pastorino (Apr 01 2020 at 19:08, on Zulip):

last meeting we had:

Santiago Pastorino (Apr 01 2020 at 19:08, on Zulip):

[#70369]: https://github.com/rust-lang/rust/pull/70369
[#70294]: https://github.com/rust-lang/rust/pull/70294
[#70126]: https://github.com/rust-lang/rust/pull/70126
[#69956]: https://github.com/rust-lang/rust/pull/69956

Santiago Pastorino (Apr 01 2020 at 19:09, on Zulip):

yeah all those show up in this search

Santiago Pastorino (Apr 01 2020 at 19:09, on Zulip):

going to check last meeting

Santiago Pastorino (Apr 01 2020 at 19:10, on Zulip):

we said

Santiago Pastorino (Apr 01 2020 at 19:10, on Zulip):

pnkfelix: #70369 declined for beta-backport
pnkfelix: #70294 beta-accepted

Santiago Pastorino (Apr 01 2020 at 19:10, on Zulip):

pnkfelix: #70126 declined for beta backport

Santiago Pastorino (Apr 01 2020 at 19:12, on Zulip):

pnkfelix: #69956 beta-accepted

Santiago Pastorino (Apr 01 2020 at 19:13, on Zulip):

@pnkfelix @centril should I tag beta accept on the ones beta-accepted and remove beta-nominations from the rejected ones?

centril (Apr 01 2020 at 19:13, on Zulip):

Would be good, but write something also

centril (Apr 01 2020 at 19:13, on Zulip):

that's what Felix usually does

pnkfelix (Apr 01 2020 at 19:14, on Zulip):

yeah sorry I obviously got distracted and didn't follow up

pnkfelix (Apr 01 2020 at 19:15, on Zulip):

I do try to link to the conversation itself from the publicly accessible zulip archive

pnkfelix (Apr 01 2020 at 19:15, on Zulip):

let me go look now

Santiago Pastorino (Apr 01 2020 at 19:15, on Zulip):

:+1:

Santiago Pastorino (Apr 01 2020 at 19:15, on Zulip):

ok, we can ignore those for this meeting

Santiago Pastorino (Apr 01 2020 at 19:15, on Zulip):

let's continue then

centril (Apr 01 2020 at 19:15, on Zulip):

@Santiago Pastorino so you're adding a comment to those right?

Santiago Pastorino (Apr 01 2020 at 19:16, on Zulip):

pnkfelix said:

let me go look now

unsure what Felix meant by this ^^^

Santiago Pastorino (Apr 01 2020 at 19:16, on Zulip):

but I guess he will?

pnkfelix (Apr 01 2020 at 19:16, on Zulip):

I meant that I would go add the comments with links, yes

Santiago Pastorino (Apr 01 2020 at 19:16, on Zulip):

anyway, I was going to ignore for now

pnkfelix (Apr 01 2020 at 19:16, on Zulip):

since I don't have to drive this meeting, thanks to @Santiago Pastorino

Santiago Pastorino (Apr 01 2020 at 19:16, on Zulip):

and add the comment ... ahh ok ok, if Felix can better :)

centril (Apr 01 2020 at 19:16, on Zulip):

great; let's continue

Santiago Pastorino (Apr 01 2020 at 19:17, on Zulip):

so, if we ignore those 4 and #70320 which is already beta accepted

Santiago Pastorino (Apr 01 2020 at 19:17, on Zulip):

we have 3

Santiago Pastorino (Apr 01 2020 at 19:18, on Zulip):

beta nom 1/3: "parse_and_disallow_postfix_after_cast: account for ExprKind::Err." #70556

Santiago Pastorino (Apr 01 2020 at 19:19, on Zulip):

well hold on

Santiago Pastorino (Apr 01 2020 at 19:19, on Zulip):
  1. T-compiler
    - Add these issues to the meeting agenda.
Santiago Pastorino (Apr 01 2020 at 19:19, on Zulip):

we are doing step 2 and 3 together

Santiago Pastorino (Apr 01 2020 at 19:19, on Zulip):

gonna add those 3 into agenda, even the one that is T-lib

Santiago Pastorino (Apr 01 2020 at 19:19, on Zulip):

correct @centril?

Wesley Wiser (Apr 01 2020 at 19:20, on Zulip):

Yeah, that sounds right

Wesley Wiser (Apr 01 2020 at 19:21, on Zulip):

based on the earlier discussion

centril (Apr 01 2020 at 19:22, on Zulip):

yea

Santiago Pastorino (Apr 01 2020 at 19:23, on Zulip):

done, there are 2 actually

Santiago Pastorino (Apr 01 2020 at 19:23, on Zulip):

Stable nominations

Santiago Pastorino (Apr 01 2020 at 19:24, on Zulip):
  1. No team assigned
    - Add T-compiler tag when it corresponds.
centril (Apr 01 2020 at 19:24, on Zulip):

Empty

Santiago Pastorino (Apr 01 2020 at 19:24, on Zulip):
  1. All teams
    - Add T-compiler tag when it corresponds.
Santiago Pastorino (Apr 01 2020 at 19:24, on Zulip):

one that is T-infra

Santiago Pastorino (Apr 01 2020 at 19:25, on Zulip):

I guess we shouldn't do anything there

Santiago Pastorino (Apr 01 2020 at 19:25, on Zulip):
  1. T-compiler
    - Add these issues to the meeting agenda.
Santiago Pastorino (Apr 01 2020 at 19:25, on Zulip):

Empty

Santiago Pastorino (Apr 01 2020 at 19:25, on Zulip):

PR's waiting for our team

Santiago Pastorino (Apr 01 2020 at 19:25, on Zulip):
Santiago Pastorino (Apr 01 2020 at 19:25, on Zulip):

we have 2

Santiago Pastorino (Apr 01 2020 at 19:26, on Zulip):

adding those to the agenda

Santiago Pastorino (Apr 01 2020 at 19:26, on Zulip):

do we feel that some of those can be quickly resolved?

centril (Apr 01 2020 at 19:26, on Zulip):

What's the purpose of nominating https://github.com/rust-lang/rust/pull/70420 for t-compiler...?

Santiago Pastorino (Apr 01 2020 at 19:27, on Zulip):

unsure :)

Santiago Pastorino (Apr 01 2020 at 19:28, on Zulip):

it's possibly just for T-lang?

centril (Apr 01 2020 at 19:28, on Zulip):

Well David did nominate it for both

simulacrum (Apr 01 2020 at 19:28, on Zulip):

T-compiler tag was added right away, T-lang added much later

simulacrum (Apr 01 2020 at 19:28, on Zulip):

er I-nominated much later

centril (Apr 01 2020 at 19:28, on Zulip):

oh yeah true

simulacrum (Apr 01 2020 at 19:28, on Zulip):

indeed, you @centril nominated it

simulacrum (Apr 01 2020 at 19:29, on Zulip):

David had no intent to do so that I can see :)

centril (Apr 01 2020 at 19:29, on Zulip):

I only intended to nominate it for lang

Santiago Pastorino (Apr 01 2020 at 19:29, on Zulip):

simulacrum said:

T-compiler tag was added right away, T-lang added much later

image.png

centril (Apr 01 2020 at 19:29, on Zulip):

Alright, dropping t-compiler then

Santiago Pastorino (Apr 01 2020 at 19:29, on Zulip):

:+1:

centril (Apr 01 2020 at 19:30, on Zulip):

(this labeling system is frequently confusing :D )

Santiago Pastorino (Apr 01 2020 at 19:30, on Zulip):

yep

Santiago Pastorino (Apr 01 2020 at 19:30, on Zulip):

the only one is #70175

Santiago Pastorino (Apr 01 2020 at 19:30, on Zulip):

so jumping to the question is it possible to discuss this briefly during the meeting?

Santiago Pastorino (Apr 01 2020 at 19:31, on Zulip):

ahh sorry this is already FCPed

Santiago Pastorino (Apr 01 2020 at 19:32, on Zulip):

nothing to do then, I guess :)

Santiago Pastorino (Apr 01 2020 at 19:32, on Zulip):

High priority issues

Santiago Pastorino (Apr 01 2020 at 19:32, on Zulip):
Santiago Pastorino (Apr 01 2020 at 19:33, on Zulip):

50 p-high and 28 unassigned

Santiago Pastorino (Apr 01 2020 at 19:35, on Zulip):

I think we have enough to discuss during this meeting so I wouldn't bother try finding more, thoughts?

Santiago Pastorino (Apr 01 2020 at 19:37, on Zulip):

Stable to beta regressions

Santiago Pastorino (Apr 01 2020 at 19:37, on Zulip):
Santiago Pastorino (Apr 01 2020 at 19:38, on Zulip):

#70098 is not assigned

Santiago Pastorino (Apr 01 2020 at 19:38, on Zulip):

I guess we can add to the agenda to assign it

Santiago Pastorino (Apr 01 2020 at 19:40, on Zulip):

are there some of those someone consider worth raising awareness?

Santiago Pastorino (Apr 01 2020 at 19:40, on Zulip):

I'm checking ...

Santiago Pastorino (Apr 01 2020 at 19:41, on Zulip):

about #70445 @simulacrum how can we check if it's already fixed as Matthew Jasper pointed out?

simulacrum (Apr 01 2020 at 19:42, on Zulip):

ah we need a beta backport for that PR (I think it's accepted?) and then we check beta

simulacrum (Apr 01 2020 at 19:42, on Zulip):

I was planning on doing beta backports tomorrow or friday

simulacrum (Apr 01 2020 at 19:42, on Zulip):

obviously someone else can do that though

Santiago Pastorino (Apr 01 2020 at 19:43, on Zulip):

:+1:

Santiago Pastorino (Apr 01 2020 at 19:43, on Zulip):

ok then we do not need to add anything else with this set

Santiago Pastorino (Apr 01 2020 at 19:43, on Zulip):

Stable to nightly regressions

Santiago Pastorino (Apr 01 2020 at 19:43, on Zulip):
Santiago Pastorino (Apr 01 2020 at 19:44, on Zulip):

there are a bunch not assigned

Santiago Pastorino (Apr 01 2020 at 19:48, on Zulip):

added those to the agenda

Santiago Pastorino (Apr 01 2020 at 19:48, on Zulip):

Stable to stable regressions

Santiago Pastorino (Apr 01 2020 at 19:48, on Zulip):
Santiago Pastorino (Apr 01 2020 at 19:48, on Zulip):

71 open stable to stable regressions

Santiago Pastorino (Apr 01 2020 at 19:49, on Zulip):

27 not prioritized

Santiago Pastorino (Apr 01 2020 at 19:49, on Zulip):

I-nominated T-compiler

Santiago Pastorino (Apr 01 2020 at 19:49, on Zulip):
Santiago Pastorino (Apr 01 2020 at 19:50, on Zulip):

there are 5

Santiago Pastorino (Apr 01 2020 at 19:50, on Zulip):

we will need to have 4 or 5 meetings to accomplish all the stuff we have in the agenda :)

eddyb (Apr 01 2020 at 19:54, on Zulip):

uh oh I missed that the pretriage moved to a different stream

Santiago Pastorino (Apr 01 2020 at 19:54, on Zulip):

welcome :)

Santiago Pastorino (Apr 01 2020 at 19:54, on Zulip):

Toolstate

Santiago Pastorino (Apr 01 2020 at 19:55, on Zulip):

Check toolstate for outstanding tool breakage.

Santiago Pastorino (Apr 01 2020 at 19:55, on Zulip):

clippy, miri, rls and rustfmt are failing

DPC (Apr 01 2020 at 19:55, on Zulip):

i think today's nightly fixes them

Santiago Pastorino (Apr 01 2020 at 19:56, on Zulip):

we are far from the release so I guess we are fine

DPC (Apr 01 2020 at 19:56, on Zulip):

as in the one that will be released in a few hours

Santiago Pastorino (Apr 01 2020 at 19:56, on Zulip):

DPC said:

i think today's nightly fixes them

that's cool :)

eddyb (Apr 01 2020 at 19:56, on Zulip):

Santiago Pastorino said:

we should split I-nominated into two things, one with let's discuss meaning and the other with the let's prioritize this one meaning

I don't think I ever use I-nominated for prioritization, can we have P-needs-prioritization or something similar for that?
I only use I-nominated for "needs discussion"

Santiago Pastorino (Apr 01 2020 at 19:56, on Zulip):

eddyb said:

Santiago Pastorino said:

we should split I-nominated into two things, one with let's discuss meaning and the other with the let's prioritize this one meaning

I don't think I ever use I-nominated for prioritization, can we have P-needs-prioritization or something similar for that?
I only use I-nominated for "needs discussion"

yes, we were discussing that and we will be introducing a new label

Santiago Pastorino (Apr 01 2020 at 19:56, on Zulip):

I-prioritize

eddyb (Apr 01 2020 at 19:56, on Zulip):

fair enough, that's short. I-discuss could be used for the other usecase

Santiago Pastorino (Apr 01 2020 at 19:56, on Zulip):

I also thought if the I prefix makes sense

eddyb (Apr 01 2020 at 19:57, on Zulip):

right, the prefixes can get silly at times

Santiago Pastorino (Apr 01 2020 at 19:57, on Zulip):

I meant I-prioritize P-prioritize or something else?

DPC (Apr 01 2020 at 19:57, on Zulip):

ah clippy failed but others passed: https://github.com/rust-lang/rust/pull/70638#issuecomment-606973331

Santiago Pastorino (Apr 01 2020 at 19:57, on Zulip):

and last

Santiago Pastorino (Apr 01 2020 at 19:57, on Zulip):

Performance regressions

Santiago Pastorino (Apr 01 2020 at 19:57, on Zulip):

Check perf regressions.

eddyb (Apr 01 2020 at 19:58, on Zulip):

anyway I left comments on #70453 to explain the situation: it's not an ICE issue

Santiago Pastorino (Apr 01 2020 at 19:59, on Zulip):

image.png

Santiago Pastorino (Apr 01 2020 at 19:59, on Zulip):

wonder what happened there?

Santiago Pastorino (Apr 01 2020 at 19:59, on Zulip):

eddyb said:

anyway I left comments on #70453 to explain the situation: it's not an ICE issue

cool, thanks!

Santiago Pastorino (Apr 01 2020 at 19:59, on Zulip):

Santiago Pastorino said:

wonder what happened there?

@centril ?

centril (Apr 01 2020 at 20:00, on Zulip):

perfbot says:

no data yet, please wait

centril (Apr 01 2020 at 20:01, on Zulip):

well https://perf.rust-lang.org/dashboard.html looks good

centril (Apr 01 2020 at 20:01, on Zulip):

@Santiago Pastorino I have no idea what happened there

eddyb (Apr 01 2020 at 20:02, on Zulip):

"no data yet, please wait" happens during uploads, just wait a minute

Santiago Pastorino (Apr 01 2020 at 20:03, on Zulip):

eddyb said:

"no data yet, please wait" happens during uploads, just wait a minute

?

Santiago Pastorino (Apr 01 2020 at 20:03, on Zulip):

you're not seeing the image?

eddyb (Apr 01 2020 at 20:03, on Zulip):

I was replying to @centril

Santiago Pastorino (Apr 01 2020 at 20:03, on Zulip):

centril said:

well https://perf.rust-lang.org/dashboard.html looks good

the screenshot I've passed doesn't look good

eddyb (Apr 01 2020 at 20:03, on Zulip):

isn't it going down?

Santiago Pastorino (Apr 01 2020 at 20:03, on Zulip):

that's number of instructions and were reduced, or am I wrong?

eddyb (Apr 01 2020 at 20:03, on Zulip):

lower = better

centril (Apr 01 2020 at 20:04, on Zulip):

its looks very good in fact

eddyb (Apr 01 2020 at 20:04, on Zulip):

ofc because rollups, good luck finding it in here https://github.com/rust-lang/rust/compare/1057dc97afce39ff6a224966ece3ed438af4c1f5...5f13820478907b09d50baf74f3ff2b78499ecd6c

centril (Apr 01 2020 at 20:04, on Zulip):

except it is going a bit up again, but meh

Santiago Pastorino (Apr 01 2020 at 20:04, on Zulip):

eddyb said:

lower = better

what is the meaning then?

centril (Apr 01 2020 at 20:04, on Zulip):

eddyb said:

ofc because rollups, good luck finding it in here https://github.com/rust-lang/rust/compare/1057dc97afce39ff6a224966ece3ed438af4c1f5...5f13820478907b09d50baf74f3ff2b78499ecd6c

please stop complaining about rollups...

centril (Apr 01 2020 at 20:05, on Zulip):

we need them, period.

eddyb (Apr 01 2020 at 20:05, on Zulip):

please stop including perf-sensitive PRs in rollups

centril (Apr 01 2020 at 20:05, on Zulip):

folks should mark sensitive PRs with rollup=never

DPC (Apr 01 2020 at 20:05, on Zulip):

if people mark PRs as perf sensitive it will be easier for us to skip them during rollup

eddyb (Apr 01 2020 at 20:05, on Zulip):

(but nobody is telling anyone to)

centril (Apr 01 2020 at 20:05, on Zulip):

we have done it several times

eddyb (Apr 01 2020 at 20:06, on Zulip):

@Santiago Pastorino I don't understand the question, if it's the number of instructions rustc executed, less instructions is better

Santiago Pastorino (Apr 01 2020 at 20:06, on Zulip):

Santiago Pastorino said:

eddyb said:

lower = better

what is the meaning then?

I read it like less instructions it's worse, like in the same period of time

Santiago Pastorino (Apr 01 2020 at 20:06, on Zulip):

but does it means less instructions to accomplish something?

eddyb (Apr 01 2020 at 20:06, on Zulip):

it's not instructions/time, it's instructions total

Santiago Pastorino (Apr 01 2020 at 20:06, on Zulip):

ahh right

Santiago Pastorino (Apr 01 2020 at 20:06, on Zulip):

cool

eddyb (Apr 01 2020 at 20:06, on Zulip):

we don't measure instruction throughput I don't think :P

Santiago Pastorino (Apr 01 2020 at 20:06, on Zulip):

well that was the last thing of the meeting

Santiago Pastorino (Apr 01 2020 at 20:07, on Zulip):

let's close the meeting :wave:, you can keep talking on this topic but please do not fight :smile: :love:

eddyb (Apr 01 2020 at 20:07, on Zulip):

my guess is the perf win is from #69033, but I'm not seeing a perf run on that PR

centril (Apr 01 2020 at 20:08, on Zulip):

uh; why would generator resume arguments in lowering affect the ctfe stress tests?

eddyb (Apr 01 2020 at 20:12, on Zulip):

CTFE? I only saw async

eddyb (Apr 01 2020 at 20:13, on Zulip):

I guess I didn't scroll too much oops

eddyb (Apr 01 2020 at 20:14, on Zulip):

the CTFE one is a different range: https://github.com/rust-lang/rust/compare/3c1d9adb3cb3aad4233075fa296fc3c70b42cdb8...a5fb9ae5b2ed3cb011ada9dc1e8633aa0927f279

eddyb (Apr 01 2020 at 20:14, on Zulip):

probably #70226

eddyb (Apr 01 2020 at 20:15, on Zulip):

not sure if @RalfJ intentionally optimized anything though in that PR?

RalfJ (Apr 03 2020 at 09:39, on Zulip):

eddyb said:

not sure if RalfJ intentionally optimized anything though in that PR?

I did not

RalfJ (Apr 03 2020 at 09:39, on Zulip):

in fact I turned a bunch of wrapping casts/arithmetic into checked casts/arithmetic. if anything that should make things slower.^^

RalfJ (Apr 03 2020 at 09:40, on Zulip):

Maybe splitting mplace_field into two methods helped? I don't see how, though. In particular not on a 64bit host where usize == u64.

RalfJ (Apr 03 2020 at 09:40, on Zulip):

centril said:

folks should mark sensitive PRs with rollup=never

so do we have a crystall ball now that will tell me in advance if a PR is perf sensitive? :D
for the PR that @eddyb asked about, never would I have expected a perf impact.

centril (Apr 03 2020 at 11:15, on Zulip):

@RalfJ yeah that's fine; I don't blame you -- but then complaining about rollups for things no one could have forseen is not good

Last update: Jun 05 2020 at 23:10UTC