Stream: t-compiler

Topic: pre-meeting triage 2019-12-12 #54818


pnkfelix (Dec 12 2019 at 12:48, on Zulip):

I will (or hope to) be doing pre-triage in this channel. (I have movers packing around in my apt, so its a little hectic here.)

pnkfelix (Dec 12 2019 at 12:51, on Zulip):

first up: unprioritized nominated issues; there are nine

pnkfelix (Dec 12 2019 at 12:52, on Zulip):

unpri nom 1/9: " ICE: no name for expr || " #67252

davidtwco (Dec 12 2019 at 12:55, on Zulip):

I'll aim to look into that as soon as, someone else can grab it if they have time before I do though.

pnkfelix (Dec 12 2019 at 12:56, on Zulip):

okay. As I'm sure you're aware, its been bisected to PR #65345. (just letting public know.)

pnkfelix (Dec 12 2019 at 12:56, on Zulip):

triage: P-high. Removing nomination tag.

pnkfelix (Dec 12 2019 at 12:57, on Zulip):

unpri nom 2/9: "Casting or adding type ascription to panic!() triggers unreachable_code" #67227

centril (Dec 12 2019 at 12:57, on Zulip):

assigned @davidtwco to the issue

pnkfelix (Dec 12 2019 at 12:58, on Zulip):

looks like #67227 is largely nominated as a lang-team concern.

pnkfelix (Dec 12 2019 at 12:58, on Zulip):

(and its not clear whether its a bug, which is why its a lang team issue)

centril (Dec 12 2019 at 12:58, on Zulip):

yep; for now

pnkfelix (Dec 12 2019 at 12:58, on Zulip):

Debating about whether this deserves a P-label

pnkfelix (Dec 12 2019 at 12:59, on Zulip):

If I were to give it any P-label, it would probably be P-medium

centril (Dec 12 2019 at 12:59, on Zulip):

nah; if t-lang decides a change is necessary we can triage next week

pnkfelix (Dec 12 2019 at 12:59, on Zulip):

sounds good

pnkfelix (Dec 12 2019 at 13:00, on Zulip):

unpri nom 3/9: "Replace our fragile safety scheme around erroneous constants" #67191

centril (Dec 12 2019 at 13:00, on Zulip):

also lang team, but possibly also for compiler for discussion of impl

pnkfelix (Dec 12 2019 at 13:05, on Zulip):

yeah okay. lets leave that for next week too

pnkfelix (Dec 12 2019 at 13:06, on Zulip):

unpri nom 4/9: "thread 'rustc' panicked at 'index out of bounds: the len is 168 but the index is 171'" #67165

pnkfelix (Dec 12 2019 at 13:09, on Zulip):

ugh sounds like a hard-to-reproduce incremental compilation issue.

pnkfelix (Dec 12 2019 at 13:13, on Zulip):

sometimes I think the compiler should check point a copy of the source in target/ on each non -ICE'ing incremental compile. And then when an ICE'ing one occurs, auto-file the bug.

centril (Dec 12 2019 at 13:14, on Zulip):

@pnkfelix that makes sense to me actually; sometimes when I get incremental ICEs or inconsistencies I just do ./x.py clean without filing a bug

pnkfelix (Dec 12 2019 at 13:14, on Zulip):

(or rather, provide an easy way for the user to confirm that the ICE can be replicated in two steps: a clean build of the checkpointed state, and the incremental build off of that cached state)

pnkfelix (Dec 12 2019 at 13:16, on Zulip):

triage: P-medium (unfortunately); until we can get a more readily reproducible scenario, we cannot readily act on this

pnkfelix (Dec 12 2019 at 13:16, on Zulip):

removing nomination label

pnkfelix (Dec 12 2019 at 13:18, on Zulip):

unpri nom 5/9: "cdylib fails to link with incremental compilation after panic -> no panic transition" #67118

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

this is also an incr comp issue, but this one sounds like it has a MCVE!

pnkfelix (Dec 12 2019 at 13:20, on Zulip):

hmm I wonder if this is actually fixed by my recent PR #67020

pnkfelix (Dec 12 2019 at 13:21, on Zulip):

anyway, I'm so excited by an incr comp bug with a reproducible test (hopefully)

pnkfelix (Dec 12 2019 at 13:22, on Zulip):

triage: P-high, removing nomination. Assigning to self to see if I can replicate, and if so, how it behaves with respect to PR #67020

pnkfelix (Dec 12 2019 at 13:22, on Zulip):

unpri nom 6/9: "ICE while compiling async/await code" #67087

pnkfelix (Dec 12 2019 at 13:23, on Zulip):

needs MCVE

pnkfelix (Dec 12 2019 at 13:23, on Zulip):

removing nomination until we get more info on how to reproduce.

pnkfelix (Dec 12 2019 at 13:25, on Zulip):

unpri nom 7/9: "Usage of errorneous constant can be omitted on nightly and beta" #67083

pnkfelix (Dec 12 2019 at 13:26, on Zulip):

this is a duplicate of #67191, .... right?

pnkfelix (Dec 12 2019 at 13:27, on Zulip):

(or at least strongly related)

pnkfelix (Dec 12 2019 at 13:28, on Zulip):

but this is also a downright regression

centril (Dec 12 2019 at 13:28, on Zulip):

yeah strongly related

pnkfelix (Dec 12 2019 at 13:28, on Zulip):

oh but we have a PR to fix it (and it just needs to land, and be backported to beta)

pnkfelix (Dec 12 2019 at 13:28, on Zulip):

namely #67134

pnkfelix (Dec 12 2019 at 13:29, on Zulip):

okay then

pnkfelix (Dec 12 2019 at 13:29, on Zulip):

triage: P-high, removing nomination, assigning to @oli (well, already was)

pnkfelix (Dec 12 2019 at 13:31, on Zulip):

unpri 8/9: " Compiler panic playing with "Err" and lifetimes" #67072

pnkfelix (Dec 12 2019 at 13:31, on Zulip):

minification shows this is actually about a mismatch between enum types

pnkfelix (Dec 12 2019 at 13:32, on Zulip):

triage: P-high, removing nomination

pnkfelix (Dec 12 2019 at 13:33, on Zulip):

unpri 9/9: "./x.py check failed if incremental builds enabled" #58633

pnkfelix (Dec 12 2019 at 13:45, on Zulip):

hmm

pnkfelix (Dec 12 2019 at 13:47, on Zulip):

@centril merged PR #67101, which says it papers over this problem

pnkfelix (Dec 12 2019 at 13:47, on Zulip):

oh, yes, that is quite a big hammer: #[allow(unused_attributes)] // FIXME(#58633): do a principled fix instead. (in src/libcore/convert/mod.rs)

pnkfelix (Dec 12 2019 at 13:48, on Zulip):

okay, leaving nominated. Maybe we'll talk about it this week, though I personally know I won't be much use at today's meeting

pnkfelix (Dec 12 2019 at 13:50, on Zulip):

marking P-high

centril (Dec 12 2019 at 13:53, on Zulip):

@pnkfelix what do you think of that branch name? ;)

pnkfelix (Dec 12 2019 at 13:54, on Zulip):

heh. are ostriches the only birds that put their heads in the sand?

centril (Dec 12 2019 at 13:55, on Zulip):

no idea, but they are one of them ^^

pnkfelix (Dec 12 2019 at 13:56, on Zulip):

there are no unprioritized beta-regressions

pnkfelix (Dec 12 2019 at 13:57, on Zulip):

nor any unpriroritized nightly regressions

pnkfelix (Dec 12 2019 at 13:58, on Zulip):

looking over the nominations now to see if anything was already discussed last week and should be unnominated. Also, I guess I'll try to build up a queue to discuss for today, using this hackmd

pnkfelix (Dec 12 2019 at 13:59, on Zulip):

this sounds familiar: "[WIP] [DO NOT MERGE] combine #66020 and #66821." #66838

pnkfelix (Dec 12 2019 at 14:00, on Zulip):

last week's discussion, from zulip archive

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

and I think it concluded here

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

removing nomination

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

this also sounds familiar: "Necromancing (putting back some removed error codes explanations)" #66836. I'll write a note

pnkfelix (Dec 12 2019 at 14:11, on Zulip):

this also sounds familiar: "Regression in Error conversion from Infallible" #66757

pnkfelix (Dec 12 2019 at 14:12, on Zulip):

ah right, T-compiler and T-lang discussed. Revert for short-term seemed prudent.

pnkfelix (Dec 12 2019 at 14:13, on Zulip):

does this need to still be nominated? I don't think it doees.

centril (Dec 12 2019 at 14:15, on Zulip):

I think you're right

pnkfelix (Dec 12 2019 at 14:16, on Zulip):

removed nomination

pnkfelix (Dec 12 2019 at 14:17, on Zulip):

okay so lets see if I can put together the agenda of nominated issues for today

Last update: May 26 2020 at 11:05UTC