Stream: t-compiler

Topic: pre-meeting triage 2019-11-07 #54818


pnkfelix (Nov 07 2019 at 13:57, on Zulip):

I will be doing pre-triage in this channel.

pnkfelix (Nov 07 2019 at 13:58, on Zulip):

we have 8 unprioritized nominated T-compiler issues

pnkfelix (Nov 07 2019 at 13:59, on Zulip):

unpri nom 1/8: "Panics when compiling" #66187.

pnkfelix (Nov 07 2019 at 13:59, on Zulip):

definitely needs a better title

lqd (Nov 07 2019 at 14:01, on Zulip):

(@pnkfelix hopefully you saw the paragraph for this meeting’s polonius wg check-in I sent you and niko yesterday)

pnkfelix (Nov 07 2019 at 14:01, on Zulip):

@lqd yep, thanks! sorry I should have sent you a response there

lqd (Nov 07 2019 at 14:01, on Zulip):

np

pnkfelix (Nov 07 2019 at 14:02, on Zulip):

triage: #66187 P-high; removing nomination label. Will likely close as duplicate of #62649 if reporter confirms issue goes away after a cargo clean.

pnkfelix (Nov 07 2019 at 14:03, on Zulip):

unpri nom 2/8: "miri no longer builds after rust-lang/rust#66175" #66179

pnkfelix (Nov 07 2019 at 14:04, on Zulip):

triage #66179 as P-medium, removing nomination label.

pnkfelix (Nov 07 2019 at 14:04, on Zulip):

unpri nom 3/8: "ICE in const fn evaluation - Disagreement between legacy and dataflow-based const validators when a const fn takes &mut self" #66167

pnkfelix (Nov 07 2019 at 14:09, on Zulip):

@ecstatic-morse has written a comment with the opinion that we should not spend time trying to fix this. I don't have much of an opinion myself on that matter.

pnkfelix (Nov 07 2019 at 14:10, on Zulip):

triage: leaving nominated for discussion

pnkfelix (Nov 07 2019 at 14:10, on Zulip):

unpri nom 4/8: "rustc error inside chroot" #66153

pnkfelix (Nov 07 2019 at 14:13, on Zulip):

this does not seem like a high priority item to address

pnkfelix (Nov 07 2019 at 14:13, on Zulip):

triage: P-medium. Assigning to self.

nagisa (Nov 07 2019 at 14:15, on Zulip):

/me pops out

nagisa (Nov 07 2019 at 14:15, on Zulip):

the chroot one I think should be just closed until the reporter proves it is not PEBCAK

pnkfelix (Nov 07 2019 at 14:16, on Zulip):

PEBCAK ?

pnkfelix (Nov 07 2019 at 14:16, on Zulip):

is that some acronym for user-error?

pnkfelix (Nov 07 2019 at 14:16, on Zulip):

okay I looked it up

nagisa (Nov 07 2019 at 14:17, on Zulip):

oh I thought it is already time for a meeting, forgot US also shifted time

pnkfelix (Nov 07 2019 at 14:17, on Zulip):

(I wrote a comment saying that the diagnostics look pretty bad. That was the main reason I left it open but made it P-medium.)

pnkfelix (Nov 07 2019 at 14:17, on Zulip):

unpri nom: "librustc_mir panics on structure" #66152

pnkfelix (Nov 07 2019 at 14:17, on Zulip):

another not-so-great title

centril (Nov 07 2019 at 14:18, on Zulip):

couldn't minimize beyond that, but it's fairly small

centril (Nov 07 2019 at 14:19, on Zulip):

maybe @nikomatsakis can look at it as part of the universes stuff

pnkfelix (Nov 07 2019 at 14:20, on Zulip):

there are other bugs that seem similar, like #64173

pnkfelix (Nov 07 2019 at 14:21, on Zulip):

since #64173 is P-medium, I'm going to call #66152 P-medium as well

centril (Nov 07 2019 at 14:22, on Zulip):

oh yeah, it seems to have identical structure

pnkfelix (Nov 07 2019 at 14:23, on Zulip):

but I'll assign myself in hopes that I don't lose track of it

pnkfelix (Nov 07 2019 at 14:23, on Zulip):

unpri nom 6/8: "clippy-driver no longer builds after rust-lang/rust#66143" #66150

pnkfelix (Nov 07 2019 at 14:25, on Zulip):

unpri nom 7/8: "Document query modifiers" #65929

pnkfelix (Nov 07 2019 at 14:25, on Zulip):

we discussed this last week, let me see...

pnkfelix (Nov 07 2019 at 14:29, on Zulip):

#65929: P-medium, removing nomination and assigned to eddyb.

pnkfelix (Nov 07 2019 at 14:29, on Zulip):

unpri nom 8/8: "Reduce the places where stable annotations are needed" #65515

pnkfelix (Nov 07 2019 at 14:31, on Zulip):

#65515: Marking P-medium. Removing nomination label. Assigning to @Esteban Küber

pnkfelix (Nov 07 2019 at 14:32, on Zulip):

okay, next prepass: beta-regressions without P-label.

pnkfelix (Nov 07 2019 at 14:33, on Zulip):

there's one: "Lifetime error when using Self instead of <StructName>" #66176

centril (Nov 07 2019 at 14:33, on Zulip):

not clear why it is a regression; I tried it on a few compilers back and it doesn't seem to be

centril (Nov 07 2019 at 14:33, on Zulip):

cc @eddyb

centril (Nov 07 2019 at 14:33, on Zulip):

and cc @Esteban Küber

pnkfelix (Nov 07 2019 at 14:34, on Zulip):

I was just trusting the labels. Lets see why @Esteban Küber tagged it as stable-to-beta regression

pnkfelix (Nov 07 2019 at 14:34, on Zulip):

oh it showed up in a crater run?

centril (Nov 07 2019 at 14:35, on Zulip):

it's a function call in the crater run, so that might explain things

centril (Nov 07 2019 at 14:35, on Zulip):

alas, this is expected behavior

pnkfelix (Nov 07 2019 at 14:36, on Zulip):

right, yeah. Lets close as not-a-bug

eddyb (Nov 07 2019 at 14:36, on Zulip):

I wrote a comment suggesting that maybe Self {...} was incompletely implemented and didn't have parameters applied or something, before realizing it doesn't even repro as a regression

centril (Nov 07 2019 at 14:36, on Zulip):

@pnkfelix do we want to convert to a diagnostics improvement issue?

pnkfelix (Nov 07 2019 at 14:37, on Zulip):

hmm

eddyb (Nov 07 2019 at 14:38, on Zulip):

I wonder if it's even possible for the current system to point to the S path in S {...} if it's the reason for constraining the lifetime

pnkfelix (Nov 07 2019 at 14:38, on Zulip):

I'm not going to reopen it myself because it would delay further triage. But I won't object if someone else reopens.

pnkfelix (Nov 07 2019 at 14:38, on Zulip):

(presuming that someone includes description of what the actual issue to be fixed is.)

pnkfelix (Nov 07 2019 at 14:39, on Zulip):

next, unprioritized stable-to-nightly prepass

pnkfelix (Nov 07 2019 at 14:39, on Zulip):

there's just one here too: "-Zsanitizer=address no longer works on macOS as of past 9 days" #66140

nagisa (Nov 07 2019 at 14:40, on Zulip):

That’s mine.

pnkfelix (Nov 07 2019 at 14:40, on Zulip):

seems P-medium at most?

nagisa (Nov 07 2019 at 14:40, on Zulip):

Did not prioritize or nominate because this is nightly-only

centril (Nov 07 2019 at 14:40, on Zulip):

@nagisa shouldn't the regression label be removed?

nagisa (Nov 07 2019 at 14:40, on Zulip):

I guess.

nagisa (Nov 07 2019 at 14:41, on Zulip):

It does break people’s fuzzers so there’s that, didn’t think of a better way to bring an attention to the fact that this broke

pnkfelix (Nov 07 2019 at 14:41, on Zulip):

there's always nomination, I suppose

pnkfelix (Nov 07 2019 at 14:42, on Zulip):

It does break people’s fuzzers so there’s that, didn’t think of a better way to bring an attention to the fact that this broke

do we include discussion of breakage of nightly-only features (or at least "popular" ones, whatever that means) in the release notes?

pnkfelix (Nov 07 2019 at 14:42, on Zulip):

Probably not worth while since people using nightlies are .. using nightlies...

pnkfelix (Nov 07 2019 at 14:43, on Zulip):

next, skim over nominated issues.

pnkfelix (Nov 07 2019 at 14:43, on Zulip):

we already discussed #66167

pnkfelix (Nov 07 2019 at 14:44, on Zulip):

the only other nominated issue is "Some features can no longer be controlled by conditional compilation" #65860

pnkfelix (Nov 07 2019 at 14:44, on Zulip):

(which I believe we discussed last week. so I'll now review for myself where the discussion has gone since then...)

pnkfelix (Nov 07 2019 at 14:45, on Zulip):

if nothing else, PR #66004 (eddyb's surgical backout) was merged ...

centril (Nov 07 2019 at 14:46, on Zulip):

#65974 was also merged

pnkfelix (Nov 07 2019 at 14:48, on Zulip):

anyway, regarding #65860 itself

pnkfelix (Nov 07 2019 at 14:51, on Zulip):

there has been another crater run, this one on a "stable/stable" delta. (Though looking at the commits, I think beta/beta would maybe be more accurate?) In any case, this new crater run is meant to sidestep the scenario where a crate detects the compiler version and enables nightly-only features that this was breaking

pnkfelix (Nov 07 2019 at 14:51, on Zulip):

@simulacrum wrote a comment summarizing the outcome there (only a few crates regressed, so that is good).

pnkfelix (Nov 07 2019 at 14:52, on Zulip):

but @simulacrum made it clear that they think we should have more discussion around the change too

pnkfelix (Nov 07 2019 at 14:52, on Zulip):

I guess I'll leave it nominated for this week. We'll see whether we can come up with any kind of plan going forward during the meeting itself.

simulacrum (Nov 07 2019 at 14:53, on Zulip):

Josh left some thoughts on the zulip channel as well from last week

pnkfelix (Nov 07 2019 at 14:54, on Zulip):

right, here's a link to that discussion

pnkfelix (Nov 07 2019 at 14:54, on Zulip):

(and here's where josh jumped in at the end)

pnkfelix (Nov 07 2019 at 14:56, on Zulip):

I'm looking over the nominated PR's too

pnkfelix (Nov 07 2019 at 14:56, on Zulip):

there is "PowerPC C ABI fixes" #64259

pnkfelix (Nov 07 2019 at 14:57, on Zulip):

I guess @eddyb nominated it nine days ago but we did not get to it last week

pnkfelix (Nov 07 2019 at 14:58, on Zulip):

(it doesn't help that the existing links for viewing nominations tend to include "is:issue". I'll fix that now.)

pnkfelix (Nov 07 2019 at 14:59, on Zulip):

the other one is "[WIP] Make a table of trait object type_ids and vtable pointers available to programs" #66113

eddyb (Nov 07 2019 at 14:59, on Zulip):

I didn't say anything on the PR for the latter but I think it's "needs RFC"?

pnkfelix (Nov 07 2019 at 15:00, on Zulip):

yeah it might

pnkfelix (Nov 07 2019 at 15:00, on Zulip):

or a design meeting. Or both.

centril (Nov 07 2019 at 15:00, on Zulip):

seems quite probable -- I haven't looked deeply into it yet

centril (Nov 07 2019 at 15:02, on Zulip):

It sounds like the PR wants to offer language guarantees somehow but just to tier 1 targets

centril (Nov 07 2019 at 15:03, on Zulip):

Seems like this needs a combined T-Lang + T-Compiler RFC

Last update: Nov 16 2019 at 01:05UTC