Stream: t-compiler

Topic: pre-meeting triage 2019-03-14 #54818


pnkfelix (Mar 14 2019 at 12:01, on Zulip):

This week's triage will be tracked here

pnkfelix (Mar 14 2019 at 12:03, on Zulip):

(I'm actually about to take part in an IRL meeting that may take up the next hour so triage may not begin in earnest for another 60 minutes or so)

pnkfelix (Mar 14 2019 at 12:04, on Zulip):

prepass: unprioritized but I-nominated T-compiler issues

pnkfelix (Mar 14 2019 at 12:08, on Zulip):

"Implement "pipelined" rustc compilation (compiler side)" #58465

pnkfelix (Mar 14 2019 at 12:08, on Zulip):

we need to dedicate time somewhere to talk about this. right now the thursday meeting is divided between triage and WG-checkin.

pnkfelix (Mar 14 2019 at 12:09, on Zulip):

its possible that it could be part of steering discussion. Or we could just allocate time for it in a thursday meeting

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

I'll switch to talking about that in the meeting channel. Moving on to more triage here...

pnkfelix (Mar 14 2019 at 13:08, on Zulip):

next: "Building Firefox with rustc 1.34.0-nightly fails with LLVM ERROR: broken function found." #58881

pnkfelix (Mar 14 2019 at 13:13, on Zulip):

I'm going to tag this as P-high since it came from Firefox

pnkfelix (Mar 14 2019 at 13:14, on Zulip):

next: "ICE with trait alias when alias refers to bare Self" #59029

pnkfelix (Mar 14 2019 at 13:15, on Zulip):

initially triaging as P-high since its an ICE

pnkfelix (Mar 14 2019 at 13:15, on Zulip):

also assigning to @Sean McArthur since they have already posted a PR for it

pnkfelix (Mar 14 2019 at 13:16, on Zulip):

next: "ICE on beta and nightly: index out of bounds in RestoreSubsliceArrayMoveOut MIR pass" #59048

pnkfelix (Mar 14 2019 at 13:17, on Zulip):

triaging as P-high as this is a stable-to-beta regression and an ICE

pnkfelix (Mar 14 2019 at 13:18, on Zulip):

next: "miri no longer builds after rust-lang/rust#56732" #59059

pnkfelix (Mar 14 2019 at 13:19, on Zulip):

triaging as P-high since IIUC miri is important piece of infrastructure

pnkfelix (Mar 14 2019 at 13:20, on Zulip):

one potential issue is that this was auto-assigned to four different people, along with a Q from highfive to @Zoxc asking if they are willing to address the fallout. I hope that the issue being assigned to four different people does not lead to everyone assuming its someone else's problem.

pnkfelix (Mar 14 2019 at 13:21, on Zulip):

next: "cargo fix --edition failure (probably macros related)" #59065

pnkfelix (Mar 14 2019 at 13:21, on Zulip):

hmm do we have a team dedicated to rustfix maintenance?

pnkfelix (Mar 14 2019 at 13:23, on Zulip):

it looks like rustfix rewrote this line:

) -> Box<dyn Iterator<Item = $crate::errors::Result<Self>> + 'a> {

to say instead

) -> Box<r#dyn Iterator<Item = $crate::errors::Result<Self>> + 'a> {

which seems wrong to me.

pnkfelix (Mar 14 2019 at 13:25, on Zulip):

I'm going to leave the nominated label but not assign a priority yet. Lets discuss in meeting.

pnkfelix (Mar 14 2019 at 13:26, on Zulip):

next: "Specific code layout can cause compiler panic with lto=true" #59137

pnkfelix (Mar 14 2019 at 13:28, on Zulip):

looks like setting lto = true on a project that uses dylib can lead ICE's.

pnkfelix (Mar 14 2019 at 13:28, on Zulip):

triaging as P-high. Leaving nominated for discussion and hopefully assignment to someone to address in some manner. (e.g. Is one supposed to disable lto when using dylib? If so, should we report a conflict error (or at least warning) explicitly then?)

pnkfelix (Mar 14 2019 at 13:29, on Zulip):

(though I personally assume that such things are supposed to "just work", even if "just work" means "you don't get the optimization.")

pnkfelix (Mar 14 2019 at 13:31, on Zulip):

second prepass, nominated issues without a team tag

pnkfelix (Mar 14 2019 at 13:36, on Zulip):

okay there are none of those (I've just been adding some refined links to the agenda)

pnkfelix (Mar 14 2019 at 13:38, on Zulip):

third prepass: beta-regressions without P-label :zero:

pnkfelix (Mar 14 2019 at 13:38, on Zulip):

fourth prepass: nightly-regressions without P-label :one:

pnkfelix (Mar 14 2019 at 13:39, on Zulip):

the issue: "Rust's libLLVM-8.so doesn't work with clang" #59034

pnkfelix (Mar 14 2019 at 13:39, on Zulip):

triage: P-high. Assigning to @emilio who has already posted PR #59173

pnkfelix (Mar 14 2019 at 13:40, on Zulip):

okay that's all the prepasses. Next up, all P-high T-compiler issues

pnkfelix (Mar 14 2019 at 13:41, on Zulip):

first, "Specific code layout can cause compiler panic with lto=true" #59137: this was discussed during nominated prepass. I left it nominated. We'll discuss it at the meeting itself.

pnkfelix (Mar 14 2019 at 13:41, on Zulip):

next: "Unions with ZSTs are no longer usable" #59126

pnkfelix (Mar 14 2019 at 13:42, on Zulip):

seems like @oli has posted PR #59139 which meant to address this.

pnkfelix (Mar 14 2019 at 13:42, on Zulip):

assigning #59126 to @oli

pnkfelix (Mar 14 2019 at 13:43, on Zulip):

next: "miri no longer builds after rust-lang/rust#56732" #59059

pnkfelix (Mar 14 2019 at 13:43, on Zulip):

I'd like to narrow down the assignee list here to one person. @Zoxc or @oli , is one of you game?

pnkfelix (Mar 14 2019 at 13:44, on Zulip):

ah I think PR https://github.com/rust-lang/miri/pull/608 is in fact the thing meant to address this, which was rebased in https://github.com/rust-lang/miri/pull/659

pnkfelix (Mar 14 2019 at 13:44, on Zulip):

and rust-lang/miri#659 landed

davidtwco (Mar 14 2019 at 13:45, on Zulip):

(I've just added a miri linkifier: miri#659)

pnkfelix (Mar 14 2019 at 13:46, on Zulip):

closing #59059 as fixed in face of miri#659

pnkfelix (Mar 14 2019 at 13:47, on Zulip):

next: "ICE on beta and nightly: index out of bounds in RestoreSubsliceArrayMoveOut MIR pass" #59048

pnkfelix (Mar 14 2019 at 13:48, on Zulip):

this needs a volunteer to look at it

pnkfelix (Mar 14 2019 at 13:48, on Zulip):

I'll post a request-for-volunteer in the meeting topic

pnkfelix (Mar 14 2019 at 13:50, on Zulip):

next: "Rust's libLLVM-8.so doesn't work with clang" #59034 -- we just discussed this. it seems under control.

pnkfelix (Mar 14 2019 at 13:53, on Zulip):

its worth nothing that PR #59034 has been beta-nominated and beta-accepted ... because ... its necessary to get the bootrapping itself fixed? Do I understand that right, @simulacrum ?

simulacrum (Mar 14 2019 at 13:53, on Zulip):

That's my understanding, yes

simulacrum (Mar 14 2019 at 13:53, on Zulip):

(this shouldn't really be a problem for anyone -- my understanding is the patch does very little except maybe fix things)

simulacrum (Mar 14 2019 at 13:53, on Zulip):

It's also small. :)

pnkfelix (Mar 14 2019 at 13:54, on Zulip):

but we won't witness the bug nor the fix in our CI ?

pnkfelix (Mar 14 2019 at 13:54, on Zulip):

(I mean "won't witness" in the sense of "this scenario is not being replicated and thus is not tested" )

simulacrum (Mar 14 2019 at 13:54, on Zulip):

I believe so, yes -- for reasons not quite well understood by me, but mostly down to "we've fixed it when noticing it by manual hacks"

pnkfelix (Mar 14 2019 at 13:55, on Zulip):

okay

simulacrum (Mar 14 2019 at 13:55, on Zulip):

this is the presumably more general fix

simulacrum (Mar 14 2019 at 13:55, on Zulip):

hacks like https://github.com/rust-lang/rust/commit/2d21df8a3fd7a68ba9f52389ead7f06f13190c12 basically

pnkfelix (Mar 14 2019 at 13:55, on Zulip):

okay thanks

pnkfelix (Mar 14 2019 at 13:56, on Zulip):

next: "ICE with trait alias when alias refers to bare Self" #59029

pnkfelix (Mar 14 2019 at 13:56, on Zulip):

appears under control, thanks to @Sean McArthur and mentoring effort from @Alexander Regueiro

pnkfelix (Mar 14 2019 at 13:56, on Zulip):

next: "Compiler panic when using a slice pattern" #59016

pnkfelix (Mar 14 2019 at 13:57, on Zulip):

this looks like a const-eval issue to me

pnkfelix (Mar 14 2019 at 13:57, on Zulip):

or is it an impl Trait issue?

pnkfelix (Mar 14 2019 at 13:59, on Zulip):

posted request for volunteer in main meeting topic

pnkfelix (Mar 14 2019 at 14:00, on Zulip):

next: "Building Firefox with rustc 1.34.0-nightly fails with LLVM ERROR: broken function found." #58881

pnkfelix (Mar 14 2019 at 14:00, on Zulip):

I discussed this briefly up above. There's been some narrowing and some investigation but no one is assigned.

pnkfelix (Mar 14 2019 at 14:00, on Zulip):

I'll post request-for-volunteer in main meeting topic

pnkfelix (Mar 14 2019 at 14:01, on Zulip):

hup its about time for the meeting itself.

nikomatsakis (Mar 14 2019 at 14:31, on Zulip):

hmm do we have a team dedicated to rustfix maintenance?

not really, but @Zack M. Davis may have some thoughts? (maybe we should ping them on #59065)

Zack M. Davis (Mar 14 2019 at 15:40, on Zulip):

@nikomatsakis @Pascal sent out an email the other week querying for interest in a Rustfix working group, which got interested responses (including from me), but I don't think the group has done anything yet

Pascal (Mar 14 2019 at 15:44, on Zulip):

@Zack M. Davis yep! I wanted to sent out a doodle this week probably today

pnkfelix (Mar 14 2019 at 20:09, on Zulip):

okay that is good to know

pnkfelix (Mar 14 2019 at 20:10, on Zulip):

we probably should have someone look into #59065 in a shorter time frame than it would take to wait for a new working group to spin up. I'll assign it to myself for now for more investigation

Last update: Nov 16 2019 at 01:15UTC