Stream: t-compiler

Topic: pre-meeting triage 2019-09-19 #54818


pnkfelix (Sep 19 2019 at 12:12, on Zulip):

I will be doing pre-triage in this channel.

pnkfelix (Sep 19 2019 at 12:16, on Zulip):

first up: we have 2 unprioritized I-nominated T-compiler issues.

pnkfelix (Sep 19 2019 at 12:24, on Zulip):

unpri nom (1): "ICE: borrowck/check_loans: assertion failed: self.bccx.region_scope_tree.scopes_intersect(...)" #64453

pnkfelix (Sep 19 2019 at 12:29, on Zulip):

okay looks like this is a stable-to-stable regression that is fixed in beta, if I understand correctly.

pnkfelix (Sep 19 2019 at 12:29, on Zulip):

going to quickly confirm that via rustup

pnkfelix (Sep 19 2019 at 12:37, on Zulip):

marking as E-needstest, P-medium. Removing nomination label. Assigning to self to deal with adding the test.

pnkfelix (Sep 19 2019 at 12:37, on Zulip):

unpri nom 2/2: "Duplicate error code usage lint removed" #64426

pnkfelix (Sep 19 2019 at 12:42, on Zulip):

Seems like this is indeed something we need to talk about. Lets leave it nominated.

pnkfelix (Sep 19 2019 at 12:46, on Zulip):

there are zero unprioritized nominated issues with no team assigned; that's good.

pnkfelix (Sep 19 2019 at 12:46, on Zulip):

there are zero unprioritized issues marked as stable-to-beta regressions

pnkfelix (Sep 19 2019 at 12:46, on Zulip):

there are zero unprioritized issues marked as stable-to-nightly regressions

pnkfelix (Sep 19 2019 at 12:47, on Zulip):

there are (deep breath) 34 P-high T-compiler issues

pnkfelix (Sep 19 2019 at 12:48, on Zulip):

11 of which are unassigned

pnkfelix (Sep 19 2019 at 12:51, on Zulip):

before I start going through those lists

pnkfelix (Sep 19 2019 at 12:51, on Zulip):

I'll just make sure I mention up front

pnkfelix (Sep 19 2019 at 12:51, on Zulip):

there are 5 nominated T-compiler issues

pnkfelix (Sep 19 2019 at 12:51, on Zulip):

(two of which are pull requests)

pnkfelix (Sep 19 2019 at 12:52, on Zulip):

there are two stable-to-beta regressions for T-compiler (both P-high, both assigned)

pnkfelix (Sep 19 2019 at 12:54, on Zulip):

and one stable-to-nightly regression (P-high), #63967, with fix in PR #64387

pnkfelix (Sep 19 2019 at 12:56, on Zulip):

ah I almost overlooked a beta nom I made last week

pnkfelix (Sep 19 2019 at 12:56, on Zulip):

(forgot to add a T-label to it. Maybe we need rustbot to warn people about I-nominated and beta-nominated with no T-label?)

pnkfelix (Sep 19 2019 at 12:58, on Zulip):

so, okay. I think those are the main things I wanted to summarize. Now I'll start going through the P-highs (after I go get coffee)

pnkfelix (Sep 19 2019 at 13:06, on Zulip):

so, as previously noted, we have 34 P-high issues.

pnkfelix (Sep 19 2019 at 13:07, on Zulip):

I'm going to spend the next 50 or so minutes attempting to traverse them, sorted by "least-recently updated"

pnkfelix (Sep 19 2019 at 13:07, on Zulip):

P-high: "LNK1189 "library limit of 65535 obj exceeded" building rustc" #53014

pnkfelix (Sep 19 2019 at 13:08, on Zulip):

I was last person assigned to this bug. I went through steps of setting up a VM, but could not reproduce the problem.

pnkfelix (Sep 19 2019 at 13:08, on Zulip):

I did ask (in May) for one of the participants to provide their reproducing test case, but they did not respond.

pnkfelix (Sep 19 2019 at 13:09, on Zulip):

So I'm going to close this.

pnkfelix (Sep 19 2019 at 13:12, on Zulip):

P high 2/34: "Recursive type with infinite size leads to internal compiler error (try_mark_green() - Forcing the DepNode should have set its color)" #57373

pnkfelix (Sep 19 2019 at 13:12, on Zulip):

assigned to niko (for delegation).

pnkfelix (Sep 19 2019 at 13:12, on Zulip):

reassigning to self.

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

P-high 3/34: "tracking issue for lifetime inference error work (E0495)" #42516

pnkfelix (Sep 19 2019 at 13:15, on Zulip):

I made a comment about this back in June, and then went on leave. :)

pnkfelix (Sep 19 2019 at 13:15, on Zulip):

I'm not really sure we should call this P-high at this point, thoguh

pnkfelix (Sep 19 2019 at 13:15, on Zulip):

realistically we are not (and cannot) prioritize this effort.

pnkfelix (Sep 19 2019 at 13:15, on Zulip):

(beyond what we have already done that is)

pnkfelix (Sep 19 2019 at 13:16, on Zulip):

I'm going to retriage this to P-medium, but leave it assigned to self.

pnkfelix (Sep 19 2019 at 13:18, on Zulip):

P-high 4/34: "Stable rustc always panics on arm/musl" #60297

pnkfelix (Sep 19 2019 at 13:18, on Zulip):

this is still nominated and marked P-high. But the platform is tier 2.

pnkfelix (Sep 19 2019 at 13:19, on Zulip):

Leaving nominated, but downgrading to P-medium.

pnkfelix (Sep 19 2019 at 13:21, on Zulip):

P-high 5/34: "Compiler panic with generic-typed nested closures" #59494

pnkfelix (Sep 19 2019 at 13:22, on Zulip):

I'll reassign myself here, see if I can make any headway on it.

pnkfelix (Sep 19 2019 at 13:22, on Zulip):

P-high 6/34: "Better error message when attempting to instantiate tuple structs with private fields" #58017

pnkfelix (Sep 19 2019 at 13:23, on Zulip):

I'll post in the meeting topic to see if we can get someone to look at this.

pnkfelix (Sep 19 2019 at 13:27, on Zulip):

P-high 7/34: "regression: rustdoc resolution changes" #61560

pnkfelix (Sep 19 2019 at 13:28, on Zulip):

assigned to nagisa who was going ot bisect it. I will ping them on the issue.

pnkfelix (Sep 19 2019 at 13:29, on Zulip):

(also, I'm not sure it deserves P-high either at this point...)

pnkfelix (Sep 19 2019 at 13:29, on Zulip):

(but I guess the bisection will tell us the true story there)

pnkfelix (Sep 19 2019 at 13:29, on Zulip):

P-high 8/34: "1.30 -> 1.31 dylib late-binding regression with GNU binutils 2.28 or older." #61539

pnkfelix (Sep 19 2019 at 13:31, on Zulip):

hmm. This was nominated for discussion. No conclusion was reached, but some of the points made in the discussion are probably worth transcribing to the issue, since it influences how we would fix this, if at all

pnkfelix (Sep 19 2019 at 13:34, on Zulip):

anyway this remains assigned to myself and nagisa.

pnkfelix (Sep 19 2019 at 13:34, on Zulip):

I would renominate it for discussion, but I'm not sure whether that's worthwhile;

pnkfelix (Sep 19 2019 at 13:34, on Zulip):

yes, we were short on time during the last discussion

pnkfelix (Sep 19 2019 at 13:35, on Zulip):

but a discussion would probably be more fruitful if we had concrete data in hand

pnkfelix (Sep 19 2019 at 13:36, on Zulip):

I wonder e.g. if I could prototype something that detects when we have a call to dlopen, and only then do the dynamically determine the linker version...

pnkfelix (Sep 19 2019 at 13:36, on Zulip):

I guess it can stay as it is for now; hopefully I'll find chance to play around with it.

pnkfelix (Sep 19 2019 at 13:37, on Zulip):

P-high 9/34: " ICE when trying to match on non-PartialEq slice." #61188

pnkfelix (Sep 19 2019 at 13:37, on Zulip):

oh yeah, this one is a bummer

pnkfelix (Sep 19 2019 at 13:37, on Zulip):

but I think @Matthew Jasper has the right idea in terms of how we might approach a solution.

pnkfelix (Sep 19 2019 at 13:38, on Zulip):

this remains assigned to me

pnkfelix (Sep 19 2019 at 13:38, on Zulip):

and I was musing about trying to fix it recently.... but I guess what I should really do is write up mentorship instructions

pnkfelix (Sep 19 2019 at 13:49, on Zulip):

P-high 10/34: "ICE on higher-trait bounds" #60218

pnkfelix (Sep 19 2019 at 13:49, on Zulip):

pinging assignee for progress report

pnkfelix (Sep 19 2019 at 13:50, on Zulip):

@nikomatsakis maybe you should re-delegate the above?

pnkfelix (Sep 19 2019 at 13:50, on Zulip):

(not sure who is the best person since its bisected to the (in)famous Universes PR ...)

pnkfelix (Sep 19 2019 at 13:51, on Zulip):

P-high 11/34: " Associated types, impl traits and closures; oh my, an ICE." #63154

pnkfelix (Sep 19 2019 at 13:52, on Zulip):

assigning to self for investigation

pnkfelix (Sep 19 2019 at 13:53, on Zulip):

P-high 12/34: "Compiler crash on associated type violating its bounds in a blanket impl" #54108

pnkfelix (Sep 19 2019 at 13:56, on Zulip):

I wrote a comment asking a question about an observation made by @Jonas Schievink during triage in August

nikomatsakis (Sep 19 2019 at 13:57, on Zulip):

P-high 4/34: "Stable rustc always panics on arm/musl" #60297

also, I tweeted about it recently and some folks seemed to think that it's working ok

nikomatsakis (Sep 19 2019 at 13:58, on Zulip):

P-high 10/34: "ICE on higher-trait bounds" #60218

I'm going to schedule some time on my calendar to investigate.

pnkfelix (Sep 19 2019 at 13:58, on Zulip):

P-high 13/34: "ICE when using anonymization in impl Trait with HRTB" #54895

pnkfelix (Sep 19 2019 at 13:59, on Zulip):

not sure why this is P-high either, apart from our general policy to mark ICE's as P-high

pnkfelix (Sep 19 2019 at 14:00, on Zulip):

i do wonder if I could find time to look at this

nikomatsakis (Sep 19 2019 at 14:07, on Zulip):

Hmm. I wonder if a good idea is to focus initially on closing out ICEs in #wg-traits.

Last update: Nov 16 2019 at 01:40UTC