Stream: t-compiler

Topic: pre-meeting triage 2019-10-03 #54818


pnkfelix (Oct 03 2019 at 12:02, on Zulip):

I will be doing pre-triage in this channel.

pnkfelix (Oct 03 2019 at 12:05, on Zulip):

There are nine nominated unprioritized issues assigned to T-compiler

pnkfelix (Oct 03 2019 at 12:05, on Zulip):

nom unpri: "UI Tests that check ICEing thread names break with parallel-compiler = true" #65047

pnkfelix (Oct 03 2019 at 12:06, on Zulip):

this seems like something that @WG-parallel-rustc would definitely care about

pnkfelix (Oct 03 2019 at 12:06, on Zulip):

but it does not need to be something we discuss amongst the whole team, I think

simulacrum (Oct 03 2019 at 12:06, on Zulip):

Agreed :) I think this is in practice just a case of "we need an ignore or separate compare mode or something for parallel"

centril (Oct 03 2019 at 12:06, on Zulip):

sgtm

pnkfelix (Oct 03 2019 at 12:07, on Zulip):

triage: P-high. Removing I-nominated label.

pnkfelix (Oct 03 2019 at 12:11, on Zulip):

nom unpri 2/9: "Add an unstable 'extra randomization' compile-time flag for hashers" #65042

centril (Oct 03 2019 at 12:12, on Zulip):

maybe design meeting?

centril (Oct 03 2019 at 12:12, on Zulip):

possibly wg-diagnostics or wg-parallel-rustc

centril (Oct 03 2019 at 12:12, on Zulip):

(I like the idea!)

pnkfelix (Oct 03 2019 at 12:13, on Zulip):

lets leave this nominated for now

pnkfelix (Oct 03 2019 at 12:13, on Zulip):

my initial reaction is that the idea as posted is likely to be untenable, but its worth at least raising amongst the team

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

(and if there is desire to investigate further, then we can post a design meeting proposal.)

centril (Oct 03 2019 at 12:14, on Zulip):

(cc @Aaron Hill )

pnkfelix (Oct 03 2019 at 12:16, on Zulip):

hey @centril , do you understand the popup text on the label C-feature-request ?

centril (Oct 03 2019 at 12:16, on Zulip):

I probably wrote it?

pnkfelix (Oct 03 2019 at 12:17, on Zulip):

"Category: a feature request, i.e. not implemented / a PR"

pnkfelix (Oct 03 2019 at 12:17, on Zulip):

its the "/ a PR" that I don't understand.

centril (Oct 03 2019 at 12:17, on Zulip):

"or"

pnkfelix (Oct 03 2019 at 12:17, on Zulip):

but what does that mean here

pnkfelix (Oct 03 2019 at 12:17, on Zulip):

like, its asking for a PR ?

pnkfelix (Oct 03 2019 at 12:18, on Zulip):

"not implemented or a PR" is combining two things in a weird way, IMO

centril (Oct 03 2019 at 12:18, on Zulip):

it's either an issue requesting a feature or a PR which requests a feature to be implemented

centril (Oct 03 2019 at 12:18, on Zulip):

I think that's what I meant by it when I wrote it

centril (Oct 03 2019 at 12:18, on Zulip):

but... please do improve it

centril (Oct 03 2019 at 12:18, on Zulip):

I wrote these hastily when I wrote all the other label descriptions

pnkfelix (Oct 03 2019 at 12:18, on Zulip):

hmm

pnkfelix (Oct 03 2019 at 12:18, on Zulip):

like a PR that puts a TODO into a comment in the code, you mean ?

pnkfelix (Oct 03 2019 at 12:19, on Zulip):

If I remember to improve it later, I'll have a go at it. Need to keep this triage moving right now though

centril (Oct 03 2019 at 12:19, on Zulip):

Yes you do :slight_smile:

pnkfelix (Oct 03 2019 at 12:19, on Zulip):

unpri nom 3/9: "ICE: src/librustc_lint/types.rs:858: unexpected type in foreign function: T" #65035

centril (Oct 03 2019 at 12:20, on Zulip):

This was a different ICE I found while minimizing the issue in the second comment

centril (Oct 03 2019 at 12:21, on Zulip):

(The two issues should probably be fixed together)

pnkfelix (Oct 03 2019 at 12:21, on Zulip):

Based on my intuition as to what the semantics of this construct would be, if it were accepted, this seems like something we would want to resolve.

centril (Oct 03 2019 at 12:22, on Zulip):

To be clear, both issues should result in "depending on type param from outer scope" errors

centril (Oct 03 2019 at 12:22, on Zulip):

since we don't have https://www.microsoft.com/en-us/research/publication/lexically-scoped-type-variables/

pnkfelix (Oct 03 2019 at 12:23, on Zulip):

okay; I wasn't sure if we potentially were willing to make duplicate static definitions for each monomorphized instance

centril (Oct 03 2019 at 12:23, on Zulip):

we don't have generic statics

centril (Oct 03 2019 at 12:23, on Zulip):

but this wouldn't be accepted for non-statics either

pnkfelix (Oct 03 2019 at 12:24, on Zulip):

in any case it doesn't strike me as an ultra high priority item to resolve

centril (Oct 03 2019 at 12:24, on Zulip):

yep

pnkfelix (Oct 03 2019 at 12:24, on Zulip):

definitely something to fix, but not something we need a status update on everyweek

pnkfelix (Oct 03 2019 at 12:24, on Zulip):

triage: P-medium, removing nomination label

centril (Oct 03 2019 at 12:24, on Zulip):

(maybe @davidtwco wants a peek at it since they are looking at improper_ctypes anyways)

pnkfelix (Oct 03 2019 at 12:25, on Zulip):

unpri nom 4/9: "type parameter A/#0 (A/0) out of range when substituting (root type=Some(*const A)) substs=[]" #65025 -- this is what @centril just mentioned.

pnkfelix (Oct 03 2019 at 12:25, on Zulip):

triage: P-medium, removing nomination label

pnkfelix (Oct 03 2019 at 12:26, on Zulip):

unpri nom 5/9: "linking of libtest failed" #64872

pnkfelix (Oct 03 2019 at 12:29, on Zulip):

I cannot quickly tell how bad this regression is

centril (Oct 03 2019 at 12:29, on Zulip):

let's cc @Alex Crichton and @mw

pnkfelix (Oct 03 2019 at 12:29, on Zulip):

I'm going to make this P-high priority and assign it to myself. And yes, cc'ing @Alex Crichton and @mw seems like a good idea.

pnkfelix (Oct 03 2019 at 12:31, on Zulip):

unpri nom 6/9: "clippy-driver no longer builds after rust-lang/rust#64864" #64867

pnkfelix (Oct 03 2019 at 12:31, on Zulip):

triage: P-medium. Has PR (I hope).

pnkfelix (Oct 03 2019 at 12:32, on Zulip):

wait no the PR is closed...

centril (Oct 03 2019 at 12:32, on Zulip):

being worked on; nothing to do beyond that

pnkfelix (Oct 03 2019 at 12:32, on Zulip):

okay P-medium. Can I assign to @centril at least?

centril (Oct 03 2019 at 12:32, on Zulip):

no please :slight_smile:

centril (Oct 03 2019 at 12:32, on Zulip):

assign to someone in clippy

pnkfelix (Oct 03 2019 at 12:32, on Zulip):

its currently assigned to 7 people

centril (Oct 03 2019 at 12:32, on Zulip):

e.g. Manish

pnkfelix (Oct 03 2019 at 12:33, on Zulip):

that's what I call spreading responsbility around to the point where I cannot follow up

centril (Oct 03 2019 at 12:33, on Zulip):

agree -- cc @Manish Goregaokar

pnkfelix (Oct 03 2019 at 12:33, on Zulip):

I don't know what to do about that assignment system; seems like something to fix in highfive?

centril (Oct 03 2019 at 12:34, on Zulip):

@pnkfelix I think it means "don't triage clippy issues and assume it will get fixed"

centril (Oct 03 2019 at 12:34, on Zulip):

that's the benefit of so many people assigned

pnkfelix (Oct 03 2019 at 12:34, on Zulip):

I don't see how assigning 7 people lets me assume much of anything

pnkfelix (Oct 03 2019 at 12:35, on Zulip):

or maybe that's your point, that I cannot assume anything

centril (Oct 03 2019 at 12:35, on Zulip):

@pnkfelix I think it means that this is not t-compilers responsibility; it's wg-clippy's responsibility

centril (Oct 03 2019 at 12:35, on Zulip):

and 7 people getting pinged means one of them will work on it

centril (Oct 03 2019 at 12:35, on Zulip):

(and several are trying afaik)

pnkfelix (Oct 03 2019 at 12:35, on Zulip):

In that scenario I would prefer to have highfive assign such a bug to the leads of WG-clippy

pnkfelix (Oct 03 2019 at 12:36, on Zulip):

that would be one or two points of contact, easier to deal with

centril (Oct 03 2019 at 12:36, on Zulip):

I would personally just opt to remove the t-compiler label and possibly add t-release instead

pnkfelix (Oct 03 2019 at 12:37, on Zulip):

we definitely should have some team assigned

pnkfelix (Oct 03 2019 at 12:37, on Zulip):

we apparently don't have a WG-clippy label. Is there actually a WG-clippy?

pnkfelix (Oct 03 2019 at 12:37, on Zulip):

or is that WG-diagnostics ?

centril (Oct 03 2019 at 12:37, on Zulip):

wg-diagnostics is for diagnostics work for rustc

centril (Oct 03 2019 at 12:37, on Zulip):

(not lints)

pnkfelix (Oct 03 2019 at 12:37, on Zulip):

right, okay, that's what I thought.

centril (Oct 03 2019 at 12:38, on Zulip):

should probably be t-dev-tools

pnkfelix (Oct 03 2019 at 12:38, on Zulip):

unpri nom 7/9: 'ICE "unexpected region in outlives inference: ReEmpty"' #64855

centril (Oct 03 2019 at 12:38, on Zulip):

( https://www.rust-lang.org/governance/teams/dev-tools )

centril (Oct 03 2019 at 12:38, on Zulip):

@Esteban Küber has r+ed PR

pnkfelix (Oct 03 2019 at 12:39, on Zulip):

triage: P-high, has PR, removing nomination label.

pnkfelix (Oct 03 2019 at 12:40, on Zulip):

unpri nom 8/9: "ICE: Associated constant and impl trait" #64848

centril (Oct 03 2019 at 12:41, on Zulip):

feels like const-eval / wg-traits

pnkfelix (Oct 03 2019 at 12:41, on Zulip):

yeah

pnkfelix (Oct 03 2019 at 12:41, on Zulip):

just debating about severity

centril (Oct 03 2019 at 12:41, on Zulip):

probably mostly traits

centril (Oct 03 2019 at 12:41, on Zulip):

p-medium? (not working on stable)

pnkfelix (Oct 03 2019 at 12:42, on Zulip):

yeah I think so

pnkfelix (Oct 03 2019 at 12:42, on Zulip):

triage: P-medium. Removing nomination label.

pnkfelix (Oct 03 2019 at 12:43, on Zulip):

unpri nom 9/9: "Internal compiler error but only when running under rls" #64821

centril (Oct 03 2019 at 12:43, on Zulip):

yay... save-analysis :frown:

pnkfelix (Oct 03 2019 at 12:44, on Zulip):

triage: P-high. Already CC'ed @Igor Matuszewski . Remvoing nomination label.

pnkfelix (Oct 03 2019 at 12:45, on Zulip):

next up: the unprioritized regressions

pnkfelix (Oct 03 2019 at 12:45, on Zulip):

we have zero unprioritized beta regressions

centril (Oct 03 2019 at 12:45, on Zulip):

nice

pnkfelix (Oct 03 2019 at 12:45, on Zulip):

and zero unprioritized nightly regressions

pnkfelix (Oct 03 2019 at 12:45, on Zulip):

(at least, none tagged with T-compiler for both of those)

centril (Oct 03 2019 at 12:46, on Zulip):

the world is not on fire?! what a strange feeling :D

pnkfelix (Oct 03 2019 at 12:47, on Zulip):

next I'm supposed to pre-pass over the nominated issues and make sure that they are all still current

pnkfelix (Oct 03 2019 at 12:47, on Zulip):

there's just two. They were obth opened in the past week: https://github.com/rust-lang/rust/issues?q=is%3Aopen+is%3Aissue+label%3AI-nominated+label%3AT-compiler

pnkfelix (Oct 03 2019 at 12:48, on Zulip):

but nonetheless

pnkfelix (Oct 03 2019 at 12:48, on Zulip):

nom: "ICE building Apache Arrow PR in librustc_traits/normalize_erasing_regions.rs:42" #64908

pnkfelix (Oct 03 2019 at 12:48, on Zulip):

@Wesley Wiser has been looking at it, came up with mcve

pnkfelix (Oct 03 2019 at 12:49, on Zulip):

hmm

pnkfelix (Oct 03 2019 at 12:50, on Zulip):

we do indeed need to be able to handle cases like this without ICE'ing. :)

pnkfelix (Oct 03 2019 at 12:50, on Zulip):

("this" being pub fn new<T: ArrowPrimitiveType>() { assert_eq!(0, std::mem::size_of::<T::Native>()); } )

pnkfelix (Oct 03 2019 at 12:50, on Zulip):

going to CC myself on this one.

pnkfelix (Oct 03 2019 at 12:51, on Zulip):

(do I dare add myself to assignee list? Not this week.)

centril (Oct 03 2019 at 12:51, on Zulip):

it somehow feels like due to mir optimizations it got bitten by the standard lack-of-lazy-norm problem

pnkfelix (Oct 03 2019 at 12:51, on Zulip):

so, next up: the P-high issues

pnkfelix (Oct 03 2019 at 12:52, on Zulip):

there are 36 of them

centril (Oct 03 2019 at 12:52, on Zulip):

@pnkfelix you know there's a "subscribe" button, right? :P

pnkfelix (Oct 03 2019 at 12:53, on Zulip):

that doesn't have the desired effect here

pnkfelix (Oct 03 2019 at 12:54, on Zulip):

I want the issues's notifications to be categorized with "Participating"

pnkfelix (Oct 03 2019 at 12:54, on Zulip):

I'm already subscribed to all of rust-lang/rust

centril (Oct 03 2019 at 12:54, on Zulip):

(...but why? :fear:)

pnkfelix (Oct 03 2019 at 12:54, on Zulip):

(for better or for worse)

centril (Oct 03 2019 at 12:55, on Zulip):

anyways; more triage, less talking

pnkfelix (Oct 03 2019 at 12:55, on Zulip):

okay, so

pnkfelix (Oct 03 2019 at 12:55, on Zulip):

P-high 1/36: "UI Tests that check ICEing thread names break with parallel-compiler = true" #65047

pnkfelix (Oct 03 2019 at 12:56, on Zulip):

already discussed

pnkfelix (Oct 03 2019 at 12:56, on Zulip):

and someone has already volunteered to try to help fix it

pnkfelix (Oct 03 2019 at 12:56, on Zulip):

I'll leave it formally unassigned

pnkfelix (Oct 03 2019 at 12:56, on Zulip):

P-high 2/36: "ICE building Apache Arrow PR in librustc_traits/normalize_erasing_regions.rs:42" #64908

pnkfelix (Oct 03 2019 at 12:56, on Zulip):

already discussed; assigned to @Wesley Wiser

pnkfelix (Oct 03 2019 at 12:57, on Zulip):

not sure whether they will be able to resolve if its indeed one of our many "waiting for lazy norm" bugs, but it seems fine

pnkfelix (Oct 03 2019 at 12:58, on Zulip):

I guess we can safely skip over all the bugs that we just covered during triage of nominated issues: 3/36: #64872, 4/36: #64855, 5/36: #64821

centril (Oct 03 2019 at 12:58, on Zulip):

If not then istm the mir opt will have to be reverted?

pnkfelix (Oct 03 2019 at 12:59, on Zulip):

If not then istm the mir opt will have to be reverted?

I guess that would be one approach to resolving this. but I suspect we'd want to investigate other avenues first. We'll see.

centril (Oct 03 2019 at 12:59, on Zulip):

yep

pnkfelix (Oct 03 2019 at 12:59, on Zulip):

P-high 6/36: "error: internal compiler error: unexpected panic: inconsistent resolution for a macro" #64803

pnkfelix (Oct 03 2019 at 13:01, on Zulip):

@Vadim Petrochenkov if you get a chance, please post status update (preferably on #64803 itself )

pnkfelix (Oct 03 2019 at 13:01, on Zulip):

P-high 7/36: "Rustc panics while compiling gstreamer in RLS" #64659

pnkfelix (Oct 03 2019 at 13:02, on Zulip):

@Igor Matuszewski these RLS bugs do deserve P-high, right? The alternative would be P-medium, in which case we wouldn't visit them each week

pnkfelix (Oct 03 2019 at 13:02, on Zulip):

P-high 8/36: "Drop implementations not being called on panic with Fat LTO." #64655

pnkfelix (Oct 03 2019 at 13:02, on Zulip):

has fix

pnkfelix (Oct 03 2019 at 13:03, on Zulip):

P-high 9/36: "debuginfo/pretty-uninitialized-vec fails with Cannot access memory at address 0x7fffff7fe000" #64343

pnkfelix (Oct 03 2019 at 13:03, on Zulip):

we talked about allocating a planning meeting for this, but I don't think we did ... /me goes to double-check

pnkfelix (Oct 03 2019 at 13:04, on Zulip):

oh no, we did

pnkfelix (Oct 03 2019 at 13:04, on Zulip):

posted note in issue

pnkfelix (Oct 03 2019 at 13:05, on Zulip):

I'm going to self-assign #64343 in the meantime

pnkfelix (Oct 03 2019 at 13:05, on Zulip):

since @nikomatsakis isn't going to be able to make the aforementioned planning meeting.

pnkfelix (Oct 03 2019 at 13:06, on Zulip):

P-high 10/36: "ThinLTO and -o compiler option can lead to duplicate object file inclusion in staticlib" #64153

pnkfelix (Oct 03 2019 at 13:07, on Zulip):

hmm. not sure if @mw has the bandwidth to address this...

pnkfelix (Oct 03 2019 at 13:07, on Zulip):

lets nominate this for discussion at today's T-compiler meeting, maybe we can find someone to take it on there.

pnkfelix (Oct 03 2019 at 13:08, on Zulip):

P-high 11/36: "STATUS_ACCESS_VIOLATION and STATUS_HEAP_CORRUPTION during compilation" #63959

pnkfelix (Oct 03 2019 at 13:10, on Zulip):

Mrrh, issue filer reports that they are again observing the bug here (as of 1st Oct)

pnkfelix (Oct 03 2019 at 13:12, on Zulip):

hmm the issue filer reported in the description that there was a Cranelift bug that was failing in a similar fashion prior the LLVM upgrade that people were theorizing to be the cause here ...

pnkfelix (Oct 03 2019 at 13:14, on Zulip):

anyway not clear what further action we can take here

pnkfelix (Oct 03 2019 at 13:15, on Zulip):

left a comment

pnkfelix (Oct 03 2019 at 13:15, on Zulip):

P-high 12/36: "Associated types, impl traits and closures; oh my, an ICE." #63154

pnkfelix (Oct 03 2019 at 13:16, on Zulip):

I self-assigned this two weeks ago but haven't had a chance to look at it

pnkfelix (Oct 03 2019 at 13:16, on Zulip):

P-high 13/36: "Bug running cargo check" #63150

pnkfelix (Oct 03 2019 at 13:19, on Zulip):

left a comment

pnkfelix (Oct 03 2019 at 13:19, on Zulip):

P-high 14/36: "Segfault compiling libc on armv7-unknown-linux-gnueabihf" #62896

pnkfelix (Oct 03 2019 at 13:21, on Zulip):

hmm got some unhappy campers in that comment thread

pnkfelix (Oct 03 2019 at 13:24, on Zulip):

hmm @msizanoen says it reproduces in QEMU. I had not seen that before.

nagisa (Oct 03 2019 at 13:26, on Zulip):

There is a backtrace with a debug compiler somewhere in there that does quite literally put smoking gun at heap corruption

nagisa (Oct 03 2019 at 13:27, on Zulip):

so it should reproduce on anything pretty much. This also ~100% an LLVM bug, most likely a double-free

pnkfelix (Oct 03 2019 at 13:27, on Zulip):

Its already assigned to me, I guess I'll just need to try allocate time to go over the comment thread

pnkfelix (Oct 03 2019 at 13:27, on Zulip):

to figure out what I can do next

nagisa (Oct 03 2019 at 13:27, on Zulip):

I can recommend https://cfarm.tetaneutral.net/machines/list/ for "hardware"

pnkfelix (Oct 03 2019 at 13:28, on Zulip):

oh interesting, thanks @nagisa

nagisa (Oct 03 2019 at 13:28, on Zulip):

then I think the next step would be to run under address sanitizer.

pnkfelix (Oct 03 2019 at 13:28, on Zulip):

these are freely availalbe?

pnkfelix (Oct 03 2019 at 13:28, on Zulip):

what the heck is this

pnkfelix (Oct 03 2019 at 13:28, on Zulip):

wait I cannot take time looking into that now

pnkfelix (Oct 03 2019 at 13:28, on Zulip):

:)

nagisa (Oct 03 2019 at 13:28, on Zulip):

@pnkfelix you need to request access, but it is pretty easy process – just a mail, and then you get ssh access to any and all of those machines.

pnkfelix (Oct 03 2019 at 13:29, on Zulip):

(I wonder if gcc farm has an AMD zen box too)

pnkfelix (Oct 03 2019 at 13:30, on Zulip):

P-high 15/36: "Undefined symbol _fltused when compiling to x86_64-unknown-uefi" #62785

pnkfelix (Oct 03 2019 at 13:31, on Zulip):

oli says PR #64990 was a fix and then that got moved over to https://github.com/rust-lang/compiler-builtins/pull/317 ...

pnkfelix (Oct 03 2019 at 13:31, on Zulip):

so I guess this is as under control as it can be

pnkfelix (Oct 03 2019 at 13:32, on Zulip):

P-high 16/36: " ICE: Generic type alias to invalid type crashes during type check on latest stable" #62742

pnkfelix (Oct 03 2019 at 13:32, on Zulip):

no word from @nikomatsakis on it. I'll nominate for reassignment.

pnkfelix (Oct 03 2019 at 13:33, on Zulip):

P-high 17/36: "internal compiler error: src/librustc/dep_graph/graph.rs:688: DepNode Hir(...) should have been pre-allocated but wasn't." #62649

pnkfelix (Oct 03 2019 at 13:33, on Zulip):

I'm going to reassign this to myself

pnkfelix (Oct 03 2019 at 13:34, on Zulip):

P-high 18/36: "Self as default type isnt typechecked" #61631

pnkfelix (Oct 03 2019 at 13:35, on Zulip):

ah I was working on updating my PR for this before I started triage. Hopefully I'll finish that tonight.

pnkfelix (Oct 03 2019 at 13:35, on Zulip):

P-high 19/36: "couldn't load codegen backend" #61561

pnkfelix (Oct 03 2019 at 13:36, on Zulip):

I self-assigned last week but haven't had a chance to really do anything yet

pnkfelix (Oct 03 2019 at 13:36, on Zulip):

(I do have a windows laptop but have not even replicated bug there yet)

pnkfelix (Oct 03 2019 at 13:36, on Zulip):

P-high 20/36: "regression: rustdoc resolution changes" #61560

pnkfelix (Oct 03 2019 at 13:37, on Zulip):

I self-assigned last week but ... ibid

pnkfelix (Oct 03 2019 at 13:37, on Zulip):

nominating for reassignment; I have a lot on my plate, and the bisection here seems like it could be delegated.

pnkfelix (Oct 03 2019 at 13:38, on Zulip):

P-high 21/36: "1.30 -> 1.31 dylib late-binding regression with GNU binutils 2.28 or older." #61539

pnkfelix (Oct 03 2019 at 13:39, on Zulip):

hmm

pnkfelix (Oct 03 2019 at 13:39, on Zulip):

I'm going to downgrade this I think

pnkfelix (Oct 03 2019 at 13:39, on Zulip):

or ... hmm.

pnkfelix (Oct 03 2019 at 13:39, on Zulip):

no, we need to resolv it

pnkfelix (Oct 03 2019 at 13:40, on Zulip):

(that or say "oh no you can't use dlopen with rust")

pnkfelix (Oct 03 2019 at 13:41, on Zulip):

or wait, bug says you can change crate type to cdylib instead of dylib

pnkfelix (Oct 03 2019 at 13:41, on Zulip):

@mw weren't you saying a goal for 2020 should be to get rid of dylibs? Am I misremembering that?

pnkfelix (Oct 03 2019 at 13:42, on Zulip):

the problem is fixed if you use a recent enough linker

pnkfelix (Oct 03 2019 at 13:42, on Zulip):

which makes me think that maybe we can downgrade, or something. Not sure.

pnkfelix (Oct 03 2019 at 13:42, on Zulip):

I'll try to dedicate some time to this over the next week.

pnkfelix (Oct 03 2019 at 13:44, on Zulip):

P-high 22/36: "Creating a recursive type with infinite size leads to ICE" #61323

pnkfelix (Oct 03 2019 at 13:46, on Zulip):

reprioritizing as P-medium

pnkfelix (Oct 03 2019 at 13:46, on Zulip):

P-high 23/36: "ICE when trying to match on non-PartialEq slice." #61188

pnkfelix (Oct 03 2019 at 13:47, on Zulip):

this still deserves P-high status.

pnkfelix (Oct 03 2019 at 13:48, on Zulip):

Do we have a label that represents "this needs mentorship instructions" ? Heh.

centril (Oct 03 2019 at 13:48, on Zulip):

E-needs-mentor

pnkfelix (Oct 03 2019 at 13:49, on Zulip):

P-high 24/36: "Compiler panic at Box<Any>" #60363

pnkfelix (Oct 03 2019 at 13:49, on Zulip):

E-needs-mentor

I considered that ... I personally am still willing to mentor

pnkfelix (Oct 03 2019 at 13:49, on Zulip):

though the fact that I have not had time to write up instructions

pnkfelix (Oct 03 2019 at 13:49, on Zulip):

is probably sign that is the right label to us

pnkfelix (Oct 03 2019 at 13:49, on Zulip):

so okay, I'll switch back to that.

pnkfelix (Oct 03 2019 at 13:51, on Zulip):

regarding #60363, it sounds like people are still hitting it, in some form

pnkfelix (Oct 03 2019 at 13:51, on Zulip):

assigning to self

pnkfelix (Oct 03 2019 at 13:52, on Zulip):

P-high 25/36: "ICE on higher-trait bounds" #60218

pnkfelix (Oct 03 2019 at 13:52, on Zulip):

yay has PR

pnkfelix (Oct 03 2019 at 13:52, on Zulip):

P-high 26/36: "Compiler panic with generic-typed nested closures" #59494

pnkfelix (Oct 03 2019 at 13:53, on Zulip):

self-assigned two weeks ago. haven't had time to look at it

pnkfelix (Oct 03 2019 at 13:53, on Zulip):

(tempted to downgrade, but I guess that can wait until next week)

pnkfelix (Oct 03 2019 at 13:53, on Zulip):

P-high 27/36: "Better error message when attempting to instantiate tuple structs with private fields" #58017

nikomatsakis (Oct 03 2019 at 13:55, on Zulip):

P-high 16/36: " ICE: Generic type alias to invalid type crashes during type check on latest stable" #62742

I missed your ping -- I will try to look at it tomorrow (in fact, it was on my list for this week)

pnkfelix (Oct 03 2019 at 13:55, on Zulip):

I would like to downgrade, but maybe before going that far, perhaps we can find someone to mentor a fix here.

pnkfelix (Oct 03 2019 at 13:56, on Zulip):

left a note

pnkfelix (Oct 03 2019 at 13:56, on Zulip):

P-high 28/36: "Coherence can be bypassed by an indirect impl for a trait object" #57893

centril (Oct 03 2019 at 13:56, on Zulip):

@pnkfelix I think that is p-high for diagnostics, not t-compiler

pnkfelix (Oct 03 2019 at 13:57, on Zulip):

this use of the labelling system to drive process remains a problem

nikomatsakis (Oct 03 2019 at 13:57, on Zulip):

under active development (#57893)

pnkfelix (Oct 03 2019 at 13:57, on Zulip):

@nikomatsakis is blitzerr still on it?

nikomatsakis (Oct 03 2019 at 13:58, on Zulip):

yeah

nikomatsakis (Oct 03 2019 at 13:58, on Zulip):

and myself

pnkfelix (Oct 03 2019 at 13:58, on Zulip):

oka

pnkfelix (Oct 03 2019 at 13:58, on Zulip):

P-high: 29/36: "Recursive type with infinite size leads to internal compiler error (try_mark_green() - Forcing the DepNode should have set its color)" #57373

pnkfelix (Oct 03 2019 at 13:58, on Zulip):

downgrading to P-medium

pnkfelix (Oct 03 2019 at 13:58, on Zulip):

I just cannot justify revisiting this every week

pnkfelix (Oct 03 2019 at 13:59, on Zulip):

though I still want to look into it

pnkfelix (Oct 03 2019 at 14:00, on Zulip):

(basically same attitiude here as I took in #61323 )

pnkfelix (Oct 03 2019 at 14:00, on Zulip):

sigh, almost finished this week

pnkfelix (Oct 03 2019 at 14:00, on Zulip):

oh well

Last update: Nov 16 2019 at 01:10UTC