Stream: t-compiler/wg-prioritization

Topic: pre-meeting triage 2020-05-07 #54818


Santiago Pastorino (May 06 2020 at 18:55, on Zulip):

@WG-prioritization pre-triage starting in 5 minutes

Santiago Pastorino (May 06 2020 at 18:58, on Zulip):

The triage meeting will happen in #t-compiler > weekly meeting 2020-05-07 #54818

Santiago Pastorino (May 06 2020 at 18:58, on Zulip):

I have created the agenda we will be building during this meeting

Santiago Pastorino (May 06 2020 at 18:59, on Zulip):

@WG-prioritization I guess we can start, one thing is that I have an ugly but somewhat working script to run the meeting now :)

Santiago Pastorino (May 06 2020 at 19:00, on Zulip):

anyway, bear a bit with me because I'm gonna be trying this, at some point would be nice to automate this a bit more and place it here as a zulip bot

DPC (May 06 2020 at 19:00, on Zulip):

@Bastian Kauschke will like to join this WG ;)

Santiago Pastorino (May 06 2020 at 19:01, on Zulip):

@Wesley Wiser do you have access to add @Bastian Kauschke to our zulip group?

Santiago Pastorino (May 06 2020 at 19:01, on Zulip):

send a PR on github to be added to the github group too

Santiago Pastorino (May 06 2020 at 19:01, on Zulip):

so let's start

pnkfelix (May 06 2020 at 19:01, on Zulip):

@Santiago Pastorino have you put the source for your script on a public repository for us to see?

Wesley Wiser (May 06 2020 at 19:02, on Zulip):

Hmm... It appears that I do not

Santiago Pastorino (May 06 2020 at 19:02, on Zulip):

yeah but it's ashame of code :P

DPC (May 06 2020 at 19:02, on Zulip):

i think only admins can

Santiago Pastorino (May 06 2020 at 19:02, on Zulip):

gonna clean it up and show

pnkfelix (May 06 2020 at 19:02, on Zulip):

(let me see if I can add @Bastian Kauschke )

Santiago Pastorino (May 06 2020 at 19:02, on Zulip):

I will add after the meeting if you can't

Santiago Pastorino (May 06 2020 at 19:02, on Zulip):

anyway let's start

Santiago Pastorino (May 06 2020 at 19:02, on Zulip):

[santiago@galago triagebot (meetings-automation-gross-wip)]$ GITHUB_API_TOKEN=XYZ cargo run --bin prioritization

Santiago Pastorino (May 06 2020 at 19:02, on Zulip):

:P

Santiago Pastorino (May 06 2020 at 19:02, on Zulip):

the script returns this ...

Santiago Pastorino (May 06 2020 at 19:03, on Zulip):

Pre-triage

Remove I-nominated leftovers

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

Issues

pnkfelix (May 06 2020 at 19:03, on Zulip):

hup you should delete that

pnkfelix (May 06 2020 at 19:03, on Zulip):

GITHUB API TOKEN was just leaked

Santiago Pastorino (May 06 2020 at 19:03, on Zulip):

doh :P

Bastian Kauschke (May 06 2020 at 19:04, on Zulip):

Still viewable :/

Santiago Pastorino (May 06 2020 at 19:04, on Zulip):

no worries I've revoked it

pnkfelix (May 06 2020 at 19:04, on Zulip):

It will probably show up on the zulip-archive too. Long term (i.e. after meeting) you should revoke that token

pnkfelix (May 06 2020 at 19:04, on Zulip):

Okay good

Santiago Pastorino (May 06 2020 at 19:04, on Zulip):

yeah I've created something with very limited access

Santiago Pastorino (May 06 2020 at 19:05, on Zulip):

and was just to test it

Santiago Pastorino (May 06 2020 at 19:05, on Zulip):

but anyway, no good to leak that :P

Santiago Pastorino (May 06 2020 at 19:05, on Zulip):

so let me paste that back

Santiago Pastorino (May 06 2020 at 19:05, on Zulip):

Santiago Pastorino said:

Pre-triage

Remove I-nominated leftovers

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

Issues

:point_up:

Santiago Pastorino (May 06 2020 at 19:05, on Zulip):

let me explain briefly what I'm gonna be doing

Santiago Pastorino (May 06 2020 at 19:06, on Zulip):

I have each section of the procedure, as we had originally

Santiago Pastorino (May 06 2020 at 19:06, on Zulip):

the script automatically runs, it shows the query we used to run, that's the Unnominate leftover I-nominated part

bjorn3 (May 06 2020 at 19:06, on Zulip):

I am going to use my bisect bot for some issues that are not yet bisected.

Santiago Pastorino (May 06 2020 at 19:07, on Zulip):

@bjorn3 great

Santiago Pastorino (May 06 2020 at 19:07, on Zulip):

but it runs possibly a different query

Santiago Pastorino (May 06 2020 at 19:07, on Zulip):

because I found errors and also because github API is not exactly the same you have through that kind of queries

Santiago Pastorino (May 06 2020 at 19:07, on Zulip):

so it's good to click that link and check if we have the same

Santiago Pastorino (May 06 2020 at 19:08, on Zulip):

so meanwhile we go over things we can check if this script is working correctly

Santiago Pastorino (May 06 2020 at 19:08, on Zulip):

in this particular case we got those two

Santiago Pastorino (May 06 2020 at 19:08, on Zulip):

so it's ok

Santiago Pastorino (May 06 2020 at 19:08, on Zulip):

anyway is there some I-nominated leftover there?

Santiago Pastorino (May 06 2020 at 19:09, on Zulip):

checking ...

Santiago Pastorino (May 06 2020 at 19:09, on Zulip):

about this one #71550, I've cooked a PR, I don't think this deserves more attention, we will mention it because it's P-critical

Santiago Pastorino (May 06 2020 at 19:10, on Zulip):

but I don't think it deserves a nomination anymore

Santiago Pastorino (May 06 2020 at 19:10, on Zulip):

and it was discussed

bjorn3 (May 06 2020 at 19:10, on Zulip):

https://github.com/bjorn3/cargo-bisect-rustc-bot-jobs/runs/650650115#step:6:107

I told cargo bisect-rustc to bisect up to 2020-04-25 and it did that, but in the final report it said that it bisected up to 2020-04-10

Santiago Pastorino (May 06 2020 at 19:11, on Zulip):

and the other one is nominated for t-lang

Santiago Pastorino (May 06 2020 at 19:11, on Zulip):

bjorn3 said:

https://github.com/bjorn3/cargo-bisect-rustc-bot-jobs/runs/650650115#step:6:107

I told cargo bisect-rustc to bisect up to 2020-04-25 and it did that, but in the final report it said that it bisected up to 2020-04-10

maybe that's an issue?

Santiago Pastorino (May 06 2020 at 19:11, on Zulip):

so back to nominations I'd say that we can leave or not the one that is P-critical

Santiago Pastorino (May 06 2020 at 19:12, on Zulip):

I'll leave it for now but unsure if worth, cc @pnkfelix

Santiago Pastorino (May 06 2020 at 19:12, on Zulip):

mainly based on that is was discussed and it's also P-critical and it will be reviewed again because of that

Santiago Pastorino (May 06 2020 at 19:12, on Zulip):

anyway, moving on ...

Santiago Pastorino (May 06 2020 at 19:14, on Zulip):

Unprioritized I-prioritize

Go over unprioritized I-prioritize issues. We should also go over I-nominated meanwhile we migrate, we were using I-nominated sometimes as an implicit request for prioritization.

No team assigned
Add T-compiler tag when it corresponds.

Issues

Empty

T-compiler
Prioritize issues and remove nomination of the ones not worth discussing.
Tag regressions accordingly. Ping appropriate people and/or ICE-breakers.

Issues

Empty

All teams
No action required. It's nice to look at this for comparison.

Issues

Empty

Regressions

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.

Issues

Empty

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.

Issues

Empty

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.

Issues

Empty

Beta nominations

No team assigned
Add T-compiler tag when it corresponds.

Issues

Empty

All teams
Add T-compiler tag when it corresponds.

Issues

Santiago Pastorino (May 06 2020 at 19:14, on Zulip):

you will notice there's a bunch of stuff in there, because this thing skips empty stuff

Santiago Pastorino (May 06 2020 at 19:15, on Zulip):

we should probably check that everything is correct

Santiago Pastorino (May 06 2020 at 19:16, on Zulip):

ok, everything looks correct

Santiago Pastorino (May 06 2020 at 19:17, on Zulip):

basically we have skipped all the Unprioritized I-prioritize section and the Regressions section

Santiago Pastorino (May 06 2020 at 19:17, on Zulip):

so we have just

Santiago Pastorino (May 06 2020 at 19:17, on Zulip):

Beta nominations

All teams
Add T-compiler tag when it corresponds.

Issues

Santiago Pastorino (May 06 2020 at 19:17, on Zulip):

let's add T-compiler to the ones that correspond

Santiago Pastorino (May 06 2020 at 19:18, on Zulip):

all of those are correctly labelled

Santiago Pastorino (May 06 2020 at 19:18, on Zulip):

let's move on

DPC (May 06 2020 at 19:18, on Zulip):

#71810 and #71682 are merged

Santiago Pastorino (May 06 2020 at 19:18, on Zulip):

yeah they can all be merged

Santiago Pastorino (May 06 2020 at 19:19, on Zulip):

but are tagged as beta nominated

DPC (May 06 2020 at 19:19, on Zulip):

okay :slight_smile:

Santiago Pastorino (May 06 2020 at 19:19, on Zulip):

so, right, please this is a meeting that happens maybe very fast

Santiago Pastorino (May 06 2020 at 19:19, on Zulip):

feel free to stop me and we can discuss

Santiago Pastorino (May 06 2020 at 19:19, on Zulip):

is a good time for that

DPC (May 06 2020 at 19:19, on Zulip):

sure :slight_smile:

Santiago Pastorino (May 06 2020 at 19:19, on Zulip):

so what happens here is that there are things that we may want to backport to beta

Santiago Pastorino (May 06 2020 at 19:19, on Zulip):

they are beta nominated

Santiago Pastorino (May 06 2020 at 19:20, on Zulip):

so those could be PRs already merged

bjorn3 (May 06 2020 at 19:20, on Zulip):

There was only one issue my bot could bisect. The rest are either compile time regressions or have dependencies. Opened bjorn3/cargo-bisect-rustc-bot#2 for the later.

Santiago Pastorino (May 06 2020 at 19:20, on Zulip):

we need to add those to our tomorrow's meeting agenda (next step in this meeting will be adding those to the agenda)

Santiago Pastorino (May 06 2020 at 19:20, on Zulip):

tomorrow we will discuss beta nominations and vote to see if we want to backport them or not

Santiago Pastorino (May 06 2020 at 19:21, on Zulip):

then we will tag as beta-accepted if we vote possitively

Santiago Pastorino (May 06 2020 at 19:21, on Zulip):

and at some point T-release will backport it

Santiago Pastorino (May 06 2020 at 19:21, on Zulip):

same happens with stable nominations

Santiago Pastorino (May 06 2020 at 19:21, on Zulip):

let's continue with the script ...

Santiago Pastorino (May 06 2020 at 19:22, on Zulip):

T-compiler
Add these issues to the meeting agenda.

Issues

Santiago Pastorino (May 06 2020 at 19:22, on Zulip):

we are still talking about beta nominations but now t-compiler

Santiago Pastorino (May 06 2020 at 19:22, on Zulip):

we need to add these to the meeting agenda

Santiago Pastorino (May 06 2020 at 19:22, on Zulip):

will do now

Santiago Pastorino (May 06 2020 at 19:23, on Zulip):

done

Santiago Pastorino (May 06 2020 at 19:23, on Zulip):

happily this is done faster now than previously thanks to the automation :)

Santiago Pastorino (May 06 2020 at 19:23, on Zulip):

let's continue ...

Santiago Pastorino (May 06 2020 at 19:23, on Zulip):

Stable nominations

No team assigned
Add T-compiler tag when it corresponds.

Issues

Empty

All teams
Add T-compiler tag when it corresponds.

Issues

Empty

T-compiler
Add these issues to the meeting agenda.

Issues

Empty

Santiago Pastorino (May 06 2020 at 19:23, on Zulip):

no stable nominations

Santiago Pastorino (May 06 2020 at 19:24, on Zulip):

PR's waiting for our team

PR's waiting for our team
Add them to the meeting agenda explaining what they are waiting for.
Explicitly nominate any that you think may be able to be resolved quickly in triage meeting.

Issues

Empty

Santiago Pastorino (May 06 2020 at 19:24, on Zulip):

no PR's waiting for our team

Santiago Pastorino (May 06 2020 at 19:24, on Zulip):

Critical and High priority issues

Santiago Pastorino (May 06 2020 at 19:24, on Zulip):

let's do this

Santiago Pastorino (May 06 2020 at 19:26, on Zulip):

this could be automated but it's actually not

Santiago Pastorino (May 06 2020 at 19:26, on Zulip):

at least the summary part

Santiago Pastorino (May 06 2020 at 19:26, on Zulip):

will fill this and be back in a bit

pnkfelix (May 06 2020 at 19:27, on Zulip):

hey @Santiago Pastorino

pnkfelix (May 06 2020 at 19:28, on Zulip):

is there T-libs stuff we need to cover?

pnkfelix (May 06 2020 at 19:28, on Zulip):

I haven't seen mention of any; that could because there isn't any T-libs stuff this week, or it could be that your script needs to incorporate T-libs?

Santiago Pastorino (May 06 2020 at 19:32, on Zulip):

we need to incorporate T-libs to the script I think, hmm

Santiago Pastorino (May 06 2020 at 19:32, on Zulip):

can re-check in a bit

Santiago Pastorino (May 06 2020 at 19:33, on Zulip):

well about T-libs I didn't have that incorporated last week

Santiago Pastorino (May 06 2020 at 19:33, on Zulip):

yeah, we should be able to see those in the All teams part

Santiago Pastorino (May 06 2020 at 19:33, on Zulip):

anyway, I will check properly after the meeting and try to be sure

Santiago Pastorino (May 06 2020 at 19:33, on Zulip):

let's continue, I've added the summary to the agenda

Santiago Pastorino (May 06 2020 at 19:34, on Zulip):

next ...

Santiago Pastorino (May 06 2020 at 19:34, on Zulip):

Stable to beta regressions

Stable-to-beta regressions (all! not just T-compiler)
Check if there are relevant issues that are worth raising awareness.
Assign if possible; if it remains unassigned, add it to agenda so we can assign during the meeting.

Issues

Santiago Pastorino (May 06 2020 at 19:34, on Zulip):

@pnkfelix here we have some T-libs things :)

Santiago Pastorino (May 06 2020 at 19:35, on Zulip):

so the only assigned thing from there is the P-high one

Santiago Pastorino (May 06 2020 at 19:35, on Zulip):

do people think it may worth adding some of the P-medium regressions to the agenda?

Santiago Pastorino (May 06 2020 at 19:36, on Zulip):

or should we just try to add some general P-high

Santiago Pastorino (May 06 2020 at 19:36, on Zulip):

?

Santiago Pastorino (May 06 2020 at 19:36, on Zulip):

the thing is that seems like there's almost in the agenda this time :)

pnkfelix (May 06 2020 at 19:36, on Zulip):

(it would be good if your script added meta-data like "unassigned" to its issue summary; I'm assuming that wouldn't be that hard.)

Santiago Pastorino (May 06 2020 at 19:36, on Zulip):

@pnkfelix yes

Santiago Pastorino (May 06 2020 at 19:37, on Zulip):

I'm figuring that a bunch of things can be improved

pnkfelix (May 06 2020 at 19:37, on Zulip):

:thumbs_up:

Santiago Pastorino (May 06 2020 at 19:37, on Zulip):

and it's literally not finished :)

pnkfelix (May 06 2020 at 19:37, on Zulip):

okay normally I push back against just adding material to the meeting because the agenda "seems light"

Wesley Wiser (May 06 2020 at 19:38, on Zulip):

I think it might be good to talk about #71359. It seems like the kind of thing where we might decide that these are WONTFIX issues.

pnkfelix (May 06 2020 at 19:38, on Zulip):

but in this case I think it seems reasonable to try to burn down our backlog of P-high issues

pnkfelix (May 06 2020 at 19:38, on Zulip):

and also, I think some of those beta-nom's might qualify for me to go ahead and just approve their backports...

Santiago Pastorino (May 06 2020 at 19:39, on Zulip):

Wesley Wiser said:

I think it might be good to talk about #71359. It seems like the kind of thing where we might decide that these are WONTFIX issues.

that one was discussed if I'm not wrong, it's also nominated so it should have been like two weeklys ago

Santiago Pastorino (May 06 2020 at 19:40, on Zulip):

pnkfelix said:

but in this case I think it seems reasonable to try to burn down our backlog of P-high issues

I wonder in some cases if we should consider just P-high or maybe some P-medium regressions?

Wesley Wiser (May 06 2020 at 19:40, on Zulip):

Thanks! Totally forgot

pnkfelix (May 06 2020 at 19:40, on Zulip):

@Santiago Pastorino wait the set of nominated topics is empty in the agenda, but there's a PR I had nominated ...

pnkfelix (May 06 2020 at 19:40, on Zulip):

oh wait, do we not included closed things as part of nominations?

Santiago Pastorino (May 06 2020 at 19:41, on Zulip):

Wesley Wiser said:

Thanks! Totally forgot

can you check if we discussed and what was the outcome?

Santiago Pastorino (May 06 2020 at 19:41, on Zulip):

pnkfelix said:

Santiago Pastorino wait the set of nominated topics is empty in the agenda, but there's a PR I had nominated ...

let me check that :)

pnkfelix (May 06 2020 at 19:41, on Zulip):

Santiago Pastorino said:

I wonder in some cases if we should consider just P-high or maybe some P-medium regressions?

Let's see. I would prioritize: 1. focusing on the unassigned P-high, then 2. try to identify assigned P-high's that haven't made recent progress, and then 3. P-medium regressions...

Santiago Pastorino (May 06 2020 at 19:42, on Zulip):

@pnkfelix the nominations part is at the end

pnkfelix (May 06 2020 at 19:42, on Zulip):

PR i nominated was "fix error code in E0751.md" #71426

pnkfelix (May 06 2020 at 19:42, on Zulip):

oh I see

Santiago Pastorino (May 06 2020 at 19:42, on Zulip):

yeah those will show up

Santiago Pastorino (May 06 2020 at 19:42, on Zulip):

maybe I should sneak into that part

pnkfelix (May 06 2020 at 19:42, on Zulip):

we shouldn't be padding out the agenda until we finsih

pnkfelix (May 06 2020 at 19:42, on Zulip):

going through all the standard filler

Santiago Pastorino (May 06 2020 at 19:42, on Zulip):

or maybe we should even re arrange the order of this

Santiago Pastorino (May 06 2020 at 19:43, on Zulip):

pnkfelix said:

we shouldn't be padding out the agenda until we finsih

agree

Santiago Pastorino (May 06 2020 at 19:43, on Zulip):

let's move on then

pnkfelix (May 06 2020 at 19:43, on Zulip):

my memory is that historically, I tried to handle going through regressions and P-high's before nominations is because I was doing it on the fly during the meeting, and it seemed more important to make sure to prioritize regressions and P-high first

pnkfelix (May 06 2020 at 19:43, on Zulip):

so its sort of an accident of history that we're still following that order now for filling out the agenda

pnkfelix (May 06 2020 at 19:44, on Zulip):

(that is, historically we often didn't have time to get to the nominations...)

Santiago Pastorino (May 06 2020 at 19:44, on Zulip):

yeah, I will probably re-sort this

pnkfelix (May 06 2020 at 19:44, on Zulip):

but if wg-prioritize can be good about constructing the agenda this far in advance, we shouldn't have that problem.

Santiago Pastorino (May 06 2020 at 19:44, on Zulip):

btw, was checking and anyway our queries do not include closed I-nominated stuff

Santiago Pastorino (May 06 2020 at 19:45, on Zulip):

pnkfelix said:

but if wg-prioritize can be good about constructing the agenda this far in advance, we shouldn't have that problem.

definitely

pnkfelix (May 06 2020 at 19:45, on Zulip):

yeah, probably because we forget to remove I-nominated from closed things.

pnkfelix (May 06 2020 at 19:45, on Zulip):

"forget" or "deliberately don't bother"

Santiago Pastorino (May 06 2020 at 19:45, on Zulip):

Santiago Pastorino said:

btw, was checking and anyway our queries do not include closed I-nominated stuff

my guess is that we are not cleaning those and the list would be huge?

Santiago Pastorino (May 06 2020 at 19:45, on Zulip):

yes

pnkfelix (May 06 2020 at 19:45, on Zulip):

right exactly

pnkfelix (May 06 2020 at 19:45, on Zulip):

okay well that's fine; that was my mistake for relying on the I-nominated label. I should know better

pnkfelix (May 06 2020 at 19:45, on Zulip):

(but also, maybe we can/should clean those out? Not sure.)

Santiago Pastorino (May 06 2020 at 19:46, on Zulip):

we could clean up everything or we can consider that one for this time :)

pnkfelix (May 06 2020 at 19:46, on Zulip):

Anyway the only reason PR #71426 was nominated was because I wanted to broadcast that I unilaterally accepted it for beta-backport.

Santiago Pastorino (May 06 2020 at 19:46, on Zulip):

I meant, we can clean up all the issues and change our search

pnkfelix (May 06 2020 at 19:46, on Zulip):

I basically wanted to add a new summary section to the meeting to have a list of all such issues

Santiago Pastorino (May 06 2020 at 19:47, on Zulip):

pnkfelix said:

Anyway the only reason PR #71426 was nominated was because I wanted to broadcast that I unilaterally accepted it for beta-backport.

can't we search for those by just beta-nominated and beta-accepted?

Santiago Pastorino (May 06 2020 at 19:47, on Zulip):

it really would be, beta-nominated, beta-accepted and I-nominated

Santiago Pastorino (May 06 2020 at 19:47, on Zulip):

regardless if closed or not

Wesley Wiser (May 06 2020 at 19:47, on Zulip):

Santiago Pastorino said:

Wesley Wiser said:

Thanks! Totally forgot

can you check if we discussed and what was the outcome?

Actually, it looks like it wasn't discussed in either of the past two meetings.

Santiago Pastorino (May 06 2020 at 19:47, on Zulip):

that's weird, was it nominated for t-compiler or t-lang?

Wesley Wiser (May 06 2020 at 19:48, on Zulip):

Ah

Santiago Pastorino (May 06 2020 at 19:48, on Zulip):

ohh I see, it's T-libs

Wesley Wiser (May 06 2020 at 19:48, on Zulip):

It was changed to libs

Wesley Wiser (May 06 2020 at 19:48, on Zulip):

Which I guess means it falls back to us again.

Santiago Pastorino (May 06 2020 at 19:49, on Zulip):

ok, so that one should show up in the list of nominations at the end I guess but if not, we can just add it

Santiago Pastorino (May 06 2020 at 19:49, on Zulip):

well I guess we can continue :)

Santiago Pastorino (May 06 2020 at 19:49, on Zulip):

Stable to nightly regressions

Stable-to-nightly regressions (all! not just T-compiler)
Check if there are relevant issues that are worth raising awareness.
Assign if possible; if it remains unassigned, add it to agenda so we can assign during the meeting.

Issues

Santiago Pastorino (May 06 2020 at 19:49, on Zulip):

ok, the only P-high is assigned :)

Santiago Pastorino (May 06 2020 at 19:50, on Zulip):

let's move on

Santiago Pastorino (May 06 2020 at 19:50, on Zulip):

I-nominated T-compiler

I-nominated T-compiler
Do a sanity check on them.
Add them to the agenda.

Issues

Santiago Pastorino (May 06 2020 at 19:50, on Zulip):

well, I guess here we may want to include T-libs too?

Santiago Pastorino (May 06 2020 at 19:51, on Zulip):

next part is about all teams anyway, so we will see those but maybe it's more clear if we have T-libs included or a separate section for T-libs

Santiago Pastorino (May 06 2020 at 19:51, on Zulip):

cc @pnkfelix

pnkfelix (May 06 2020 at 19:52, on Zulip):

I guess separate T-libs is good

Santiago Pastorino (May 06 2020 at 19:52, on Zulip):

so according to this #71550 is nominated but it's also P-critical, will add that one but it's clear is gonna be discussed because of being P-critical

Santiago Pastorino (May 06 2020 at 19:52, on Zulip):

and #27060 is t-lang

pnkfelix (May 06 2020 at 19:53, on Zulip):

if nothing else, keeping separate T-libs may help us keep rough track of how much overhead supporting T-libs is adding

Santiago Pastorino (May 06 2020 at 19:53, on Zulip):

Check I-nominated all-teams.

Issues

Santiago Pastorino (May 06 2020 at 19:54, on Zulip):

looking at that list there are 7 t-libs nominations

Santiago Pastorino (May 06 2020 at 19:55, on Zulip):

@pnkfelix should I add them all? :smile:

pnkfelix (May 06 2020 at 19:55, on Zulip):

um

pnkfelix (May 06 2020 at 19:55, on Zulip):

I don't know. I doubt it

Santiago Pastorino (May 06 2020 at 19:55, on Zulip):

maybe it's not fair to judge based on this, once we can "catch" up it's gonna be less on a week by week basis

pnkfelix (May 06 2020 at 19:55, on Zulip):

keep in mind that our T-libs support is meant to center around implementation issues

Santiago Pastorino (May 06 2020 at 19:56, on Zulip):

one good strategy is to add them all, sorted by priority and try to cover the ones we can

Santiago Pastorino (May 06 2020 at 19:56, on Zulip):

maybe at some point we are even :)

pnkfelix (May 06 2020 at 19:56, on Zulip):

I don't think it necessarily works out well that way

pnkfelix (May 06 2020 at 19:56, on Zulip):

when there's N things on the agenda, for large N

Santiago Pastorino (May 06 2020 at 19:56, on Zulip):

pnkfelix said:

keep in mind that our T-libs support is meant to center around implementation issues

yes, I didn't have time to check what those are really about, maybe neither of those are about implementation details

pnkfelix (May 06 2020 at 19:57, on Zulip):

large N means that I (or whomever is facillitating the meeting) is pressured to dedicated O(1/N) amount of time to each issue

Santiago Pastorino (May 06 2020 at 19:57, on Zulip):

I guess we would need to check what's the reason for those nominations

Santiago Pastorino (May 06 2020 at 19:57, on Zulip):

will do that now :)

Santiago Pastorino (May 06 2020 at 19:57, on Zulip):

hmm this may take a lot of time from the meeting

pnkfelix (May 06 2020 at 19:58, on Zulip):

e.g. its pretty unclear whether #68709 would fall under our domain, unless we actually "fixed" it (which I don't think we're going to do)

Santiago Pastorino (May 06 2020 at 19:59, on Zulip):

maybe is good to discuss that one?

pnkfelix (May 06 2020 at 19:59, on Zulip):

heh that was meant to be an example of something I figured we'd ignore

pnkfelix (May 06 2020 at 20:00, on Zulip):

but I guess we could bring it up to see if we can see a path for closing it

Santiago Pastorino (May 06 2020 at 20:00, on Zulip):

yeap

pnkfelix (May 06 2020 at 20:00, on Zulip):

but honestly, I suspect that issue is more an issue for T-libs

Santiago Pastorino (May 06 2020 at 20:00, on Zulip):

:+1:

pnkfelix (May 06 2020 at 20:00, on Zulip):

in terms of them deciding how to document the lack of specification in cases like this

Santiago Pastorino (May 06 2020 at 20:00, on Zulip):

one thing I wonder is how do we separate the T-libs things that are for T-libs from the ones that are for T-compiler

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

do we want to tag those just as T-compiler?, should we tag those kind of things with both labels?

pnkfelix (May 06 2020 at 20:01, on Zulip):

I want to go back and review the meeting where we agreed to take on T-libs implementation work

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

differentiating between T-libs things for just T-libs from things that are meant for T-compiler under this new schema would make all this decisions a bit easier :)

pnkfelix (May 06 2020 at 20:02, on Zulip):

I'm pretty sure it was implicitly established that the I-nominated T-libs issues were not intended to be our problem

Santiago Pastorino (May 06 2020 at 20:02, on Zulip):

ok, if you're going to check that please let us know here about that

Santiago Pastorino (May 06 2020 at 20:02, on Zulip):

let's continue with the meeting then :)

Santiago Pastorino (May 06 2020 at 20:02, on Zulip):

Announcements

pnkfelix (May 06 2020 at 20:03, on Zulip):

Santiago Pastorino said:

differentiating between T-libs things for just T-libs from things that are meant for T-compiler under this new schema would make all this decisions a bit easier :)

note that we did add a libs-impl label to github, which presumably could be used as a differentiating tag here

Santiago Pastorino (May 06 2020 at 20:04, on Zulip):

ok

Santiago Pastorino (May 06 2020 at 20:05, on Zulip):

would need to add that to this process

Santiago Pastorino (May 06 2020 at 20:05, on Zulip):

so we are saying that we should not consider T-libs nominations?

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

but I'd say we should consider T-libs regressions

pnkfelix (May 06 2020 at 20:06, on Zulip):

Santiago Pastorino said:

ok, if you're going to check that please let us know here about that

okay, conversation at end strongly implied to me that this WG should not be using T-libs in every case as basis for its input workload

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

or should regressions be tagged as libs-impl?

pnkfelix (May 06 2020 at 20:06, on Zulip):

I think we should look at everything tagged libs-impl

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

pnkfelix said:

Santiago Pastorino said:

ok, if you're going to check that please let us know here about that

okay, conversation at end strongly implied to me that this WG should not be using T-libs in every case as basis for its input workload

but I think some impl things are still tagged as T-libs?

pnkfelix (May 06 2020 at 20:06, on Zulip):

and we should also probably look at regressions tagged T-libs

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

like regressions

pnkfelix (May 06 2020 at 20:06, on Zulip):

but not nominations

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

ok ok

pnkfelix (May 06 2020 at 20:06, on Zulip):

the nominations is too much workload for now

Santiago Pastorino (May 06 2020 at 20:07, on Zulip):

agreed then

pnkfelix (May 06 2020 at 20:07, on Zulip):

I may (or may not) try to review the T-libs issues (at least the nominated ones) to see which are candidates to be tagged with libs-impl

pnkfelix (May 06 2020 at 20:07, on Zulip):

and then that way it would end up as part of input for this WG

Santiago Pastorino (May 06 2020 at 20:07, on Zulip):

:+1:

pnkfelix (May 06 2020 at 20:07, on Zulip):

but (importantly) I think anyone can do that

Santiago Pastorino (May 06 2020 at 20:08, on Zulip):

yes

pnkfelix (May 06 2020 at 20:08, on Zulip):

Another detail

pnkfelix (May 06 2020 at 20:08, on Zulip):

the main reason the libs design team asked us to take on work

pnkfelix (May 06 2020 at 20:08, on Zulip):

was to deal with PR queue

pnkfelix (May 06 2020 at 20:08, on Zulip):

but we don't actually review the PR queue here, apart from looking at nominated PR's or PR's marked S-waiting-on-team, right?

pnkfelix (May 06 2020 at 20:09, on Zulip):

anyway its just a thought

pnkfelix (May 06 2020 at 20:10, on Zulip):

as in, I don't know how to solve that problem, and I don't want to attempt to solve it right now in this meeting

Santiago Pastorino (May 06 2020 at 20:10, on Zulip):

pnkfelix said:

was to deal with PR queue

what do you meant by deal with PR queue?

pnkfelix (May 06 2020 at 20:12, on Zulip):

making sure things get reviewed or closed as "not gonna happen"

pnkfelix (May 06 2020 at 20:12, on Zulip):

(or "needs MCP", "needs RFC", etc)

Santiago Pastorino (May 06 2020 at 20:12, on Zulip):

:+1:

pnkfelix (May 06 2020 at 20:12, on Zulip):

we tend to let that happen organically

Santiago Pastorino (May 06 2020 at 20:12, on Zulip):

yes

pnkfelix (May 06 2020 at 20:12, on Zulip):

and T-libs told us they didn't have workforce to deal with PR's

Santiago Pastorino (May 06 2020 at 20:13, on Zulip):

:+1:

Santiago Pastorino (May 06 2020 at 20:13, on Zulip):

Santiago Pastorino said:

Announcements

have added things to the agenda

Santiago Pastorino (May 06 2020 at 20:13, on Zulip):

tidying it up a little bit

DPC (May 06 2020 at 20:14, on Zulip):

just fyi there's a point release happening this week (tomorrow)

Santiago Pastorino (May 06 2020 at 20:15, on Zulip):

:+1:

Santiago Pastorino (May 06 2020 at 20:15, on Zulip):

@pnkfelix can you check https://hackmd.io/CEVZBzg7T8OmQaIeLf6GTw?both#Announcements ?

Santiago Pastorino (May 06 2020 at 20:15, on Zulip):

in particular about the form

pnkfelix (May 06 2020 at 20:16, on Zulip):

oh yeah thanks for adding that

Santiago Pastorino (May 06 2020 at 20:16, on Zulip):

DPC said:

just fyi there's a point release happening this week (tomorrow)

1.43.1?

Santiago Pastorino (May 06 2020 at 20:16, on Zulip):

pnkfelix said:

oh yeah thanks for adding that

phrase it the way you prefer

pnkfelix (May 06 2020 at 20:16, on Zulip):

might as well remind people that there are two forms, one for username-identified and one for anonymous responses

pnkfelix (May 06 2020 at 20:16, on Zulip):

okay I'll rephrase

Santiago Pastorino (May 06 2020 at 20:17, on Zulip):

we are items away to finish and I'd need to check a bunch of things

Santiago Pastorino (May 06 2020 at 20:17, on Zulip):

:)

Santiago Pastorino (May 06 2020 at 20:17, on Zulip):

Toolstate

Santiago Pastorino (May 06 2020 at 20:18, on Zulip):

we are far from release :)

Santiago Pastorino (May 06 2020 at 20:18, on Zulip):

anyway cc @DPC

DPC (May 06 2020 at 20:18, on Zulip):

yeh

Santiago Pastorino (May 06 2020 at 20:18, on Zulip):

and lastly

Santiago Pastorino (May 06 2020 at 20:18, on Zulip):

Performance regressions

Check perf regressions.

Santiago Pastorino (May 06 2020 at 20:19, on Zulip):

image.png

Santiago Pastorino (May 06 2020 at 20:19, on Zulip):

what has happened there?

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

https://perf.rust-lang.org/compare.html?start=19ae74d0d0e864bfc5c809e47af3a903d3fc2cf1&end=dae90c195989b09475b6c0225a3018cbd7afa587&stat=instructions:u

Santiago Pastorino (May 06 2020 at 20:21, on Zulip):

does someone understand this? O_O

o0Ignition0o - Jeremy Lempereur (May 06 2020 at 20:21, on Zulip):

i dont

Santiago Pastorino (May 06 2020 at 20:21, on Zulip):

https://github.com/rust-lang/rust/pull/71716

Santiago Pastorino (May 06 2020 at 20:22, on Zulip):

how renaming that generates that performance regression?

Santiago Pastorino (May 06 2020 at 20:22, on Zulip):

or am I looking at the wrong thing?

Bastian Kauschke (May 06 2020 at 20:22, on Zulip):

or it's https://github.com/rust-lang/rust/commit/19e5da902bf6ade2c0558383051215459754b73d

Bastian Kauschke (May 06 2020 at 20:22, on Zulip):

which makes even less sense afaict

DPC (May 06 2020 at 20:23, on Zulip):

it's in a rollup so could be something from the rollup?

pnkfelix (May 06 2020 at 20:23, on Zulip):

Santiago Pastorino said:

how renaming that generates that performance regression?

if its an option renaming, maybe something in benchmark suite is still invoking with old option?

Santiago Pastorino (May 06 2020 at 20:23, on Zulip):

or this starts in https://github.com/rust-lang/rust/pull/71949 ?

Santiago Pastorino (May 06 2020 at 20:24, on Zulip):

so I'm not 100% sure how to go from that perf link to the proper PR

Santiago Pastorino (May 06 2020 at 20:25, on Zulip):

it should be the second link

Santiago Pastorino (May 06 2020 at 20:25, on Zulip):

in our case

Santiago Pastorino (May 06 2020 at 20:25, on Zulip):

https://github.com/rust-lang/rust/commit/dae90c195989b09475b6c0225a3018cbd7afa587

Santiago Pastorino (May 06 2020 at 20:26, on Zulip):

and from there you see the PR

Santiago Pastorino (May 06 2020 at 20:26, on Zulip):

but how that PR causes this issue?

pnkfelix (May 06 2020 at 20:26, on Zulip):

the summary is an average, right?

DPC (May 06 2020 at 20:26, on Zulip):

do you want to do a revert+perf test?

pnkfelix (May 06 2020 at 20:26, on Zulip):

so we should be able to see which cases below are correlated

Santiago Pastorino (May 06 2020 at 20:27, on Zulip):

I guess it may be what Felix is saying

Santiago Pastorino (May 06 2020 at 20:27, on Zulip):

pnkfelix said:

Santiago Pastorino said:

how renaming that generates that performance regression?

if its an option renaming, maybe something in benchmark suite is still invoking with old option?

:point_up:

Santiago Pastorino (May 06 2020 at 20:27, on Zulip):

unsure

Santiago Pastorino (May 06 2020 at 20:28, on Zulip):

cc @simulacrum

Santiago Pastorino (May 06 2020 at 20:29, on Zulip):

actually cc @Alex Crichton

pnkfelix (May 06 2020 at 20:29, on Zulip):

if nothing else

pnkfelix (May 06 2020 at 20:29, on Zulip):

lets file a bug, right?

Santiago Pastorino (May 06 2020 at 20:29, on Zulip):

given that Alex seems to be online

pnkfelix (May 06 2020 at 20:30, on Zulip):

like, this meeting is not the place to try to debug this. :)

Santiago Pastorino (May 06 2020 at 20:30, on Zulip):

yeah

Santiago Pastorino (May 06 2020 at 20:30, on Zulip):

definitely

Santiago Pastorino (May 06 2020 at 20:30, on Zulip):

all I'm trying to do is routing it :)

simulacrum (May 06 2020 at 20:30, on Zulip):

what do you need from me?

Santiago Pastorino (May 06 2020 at 20:31, on Zulip):

I wanted to know if you were aware of this performance regression

pnkfelix (May 06 2020 at 20:31, on Zulip):

wait, the PR itself discusses a performance regression

simulacrum (May 06 2020 at 20:31, on Zulip):

oh yeah that should be fixed soon

pnkfelix (May 06 2020 at 20:31, on Zulip):

see here: https://github.com/rust-lang/rust/pull/71716#issuecomment-623734683

simulacrum (May 06 2020 at 20:31, on Zulip):

we're just getting cargo and rustc synced

Santiago Pastorino (May 06 2020 at 20:31, on Zulip):

and actually it seems to be cause by https://github.com/rust-lang/rust/pull/71716

simulacrum (May 06 2020 at 20:31, on Zulip):

(on the name of the option)

Santiago Pastorino (May 06 2020 at 20:31, on Zulip):

:+1:

simulacrum (May 06 2020 at 20:31, on Zulip):

so the perf graphs look a bit wonky but I wouldn't worry about it

Santiago Pastorino (May 06 2020 at 20:31, on Zulip):

:+1:

Santiago Pastorino (May 06 2020 at 20:32, on Zulip):

ok, so we've executed all the procedure of the meeting

Santiago Pastorino (May 06 2020 at 20:32, on Zulip):

we have this agenda https://hackmd.io/CEVZBzg7T8OmQaIeLf6GTw?view

Santiago Pastorino (May 06 2020 at 20:32, on Zulip):

but it looks really empty :)

Santiago Pastorino (May 06 2020 at 20:32, on Zulip):

@pnkfelix this time for real :)

Santiago Pastorino (May 06 2020 at 20:33, on Zulip):

should I go over and nominate some P-high stuff or do you want to keep the meeting short?

Alex Crichton (May 06 2020 at 20:34, on Zulip):

@Santiago Pastorino hey how can I help?

Santiago Pastorino (May 06 2020 at 20:35, on Zulip):

there seems to be a regression starting on #71716 but don't worry simulacrum has explained that you're syncronizing rustc and cargo

pnkfelix (May 06 2020 at 20:43, on Zulip):

Santiago Pastorino said:

should I go over and nominate some P-high stuff or do you want to keep the meeting short?

lets nominate some unassigned P-high issues, yeah

Santiago Pastorino (May 06 2020 at 20:44, on Zulip):

:+1:

pnkfelix (May 06 2020 at 21:01, on Zulip):

@Santiago Pastorino anyway I'll take care of nominating some issues

pnkfelix (May 06 2020 at 21:02, on Zulip):

we can end this meeting now, unless people have issues to raise

Santiago Pastorino (May 06 2020 at 21:02, on Zulip):

pnkfelix said:

Santiago Pastorino anyway I'll take care of nominating some issues

please go ahead if you want, I was planning on doing so too

Santiago Pastorino (May 06 2020 at 21:02, on Zulip):

yep this meeting is over :wave:

pnkfelix (May 06 2020 at 21:03, on Zulip):

@Santiago Pastorino this query yielded a nice set of just four issues. (I just added "-label:I-ICE -label:E-needs-mcve"

pnkfelix (May 06 2020 at 21:03, on Zulip):

so I'll put those on the list

Santiago Pastorino (May 06 2020 at 21:03, on Zulip):

cool

Santiago Pastorino (May 06 2020 at 21:04, on Zulip):

pnkfelix said:

so I'll put those on the list

you meant, you will nominate those? or we should just place those in the agenda?

Bastian Kauschke (May 06 2020 at 21:04, on Zulip):

@Santiago Pastorino Can you add me to the zulip group :sparkles:

send a PR on github to be added to the github group too

^ This is something I should do afaict.

Santiago Pastorino (May 06 2020 at 21:07, on Zulip):

@Bastian Kauschke :+1:, you're already in the zulip group, send a PR to rust-lang/team to add yourself to this wg and mention me in the PR

pnkfelix (May 06 2020 at 21:10, on Zulip):

Santiago Pastorino said:

you meant, you will nominate those? or we should just place those in the agenda?

I'm going to do both: nominate and update the agenda accordingly.

Santiago Pastorino (May 06 2020 at 21:13, on Zulip):

cool

Bastian Kauschke (May 06 2020 at 21:15, on Zulip):

Looks like I don't exist :shrug: https://github.com/rust-lang/team/pull/333

DPC (May 06 2020 at 21:16, on Zulip):

@Bastian Kauschke did you add yourself to the people section?

DPC (May 06 2020 at 21:16, on Zulip):

and look at you, you got PR #333 xD

Bastian Kauschke (May 06 2020 at 21:16, on Zulip):

snice

Bastian Kauschke (May 06 2020 at 21:17, on Zulip):

no, will do :thumbs_up:

DPC (May 06 2020 at 21:17, on Zulip):

add yourself here: https://github.com/rust-lang/team/tree/master/people

Alex Crichton (May 06 2020 at 23:29, on Zulip):

@Santiago Pastorino sorry got distracted

Alex Crichton (May 06 2020 at 23:29, on Zulip):

Yes that regression will be fixed with cargo update

Bastian Kauschke (May 07 2020 at 14:49, on Zulip):

rust-lang now has 256 members :sparkles: something may break by adding one more :laughing:

Last update: Jun 05 2020 at 23:05UTC