Stream: t-compiler

Topic: weekly meeting 2019-08-01 #54818


nagisa (Aug 01 2019 at 13:17, on Zulip):

@T-compiler/meeting ; the triage meeting will be starting in slightly more than 40 minutes

nagisa (Aug 01 2019 at 13:18, on Zulip):

Pre-triage is here.

nagisa (Aug 01 2019 at 14:01, on Zulip):

@T-compiler/meeting lets begin, shall we? Starting with 5 mins for announcements.

Santiago Pastorino (Aug 01 2019 at 14:04, on Zulip):

we are still looking to fill the experts map https://github.com/rust-lang/compiler-team/issues/87, with experts, knowledgeable people and directories of the different parts of the compiler, even commenting on the issue if easier would work :)

nagisa (Aug 01 2019 at 14:04, on Zulip):

I will note that there will not be a checkin this week either, but there appears to be progress towards making schedule available.

nagisa (Aug 01 2019 at 14:05, on Zulip):

Okay, so without further ado, issues. I have pre-triaged some I-nominated issues, but two are remaining

nagisa (Aug 01 2019 at 14:05, on Zulip):

Bug running cargo check #63150

nagisa (Aug 01 2019 at 14:06, on Zulip):

P-high, probably. Is there any guesses as to whom would be the best person to look into this?

nagisa (Aug 01 2019 at 14:07, on Zulip):

Appears to be related to dependency tracking stuff, so perhaps @mw? It has been a while since I’ve last seen them

nagisa (Aug 01 2019 at 14:08, on Zulip):

Okay, marked P-high, without assignee.

nagisa (Aug 01 2019 at 14:08, on Zulip):

Next is [MIR] [ICE] Associated types, impl traits and closures; oh my, an ICE. #63154

davidtwco (Aug 01 2019 at 14:09, on Zulip):

Appears to be related to dependency tracking stuff, so perhaps mw? It has been a while since I’ve last seen them

I believe mw is still on leave until some time in August.

nagisa (Aug 01 2019 at 14:09, on Zulip):

I see.

nagisa (Aug 01 2019 at 14:10, on Zulip):

For the MIR issue, it seems like a typical insufficient normalization issue.

nagisa (Aug 01 2019 at 14:11, on Zulip):

The fact that miri check does work suggests that all that’s needed here will be a normalize() call somewhere.

nagisa (Aug 01 2019 at 14:13, on Zulip):

Marked as P-high, wrote a comment. Moving on.

nagisa (Aug 01 2019 at 14:13, on Zulip):

Our list of P-high issues is longer than we could hope to handle, so I picked only a few that appear to have had something interesting happen.

nagisa (Aug 01 2019 at 14:14, on Zulip):

internal error: entered unreachable code #63164

nagisa (Aug 01 2019 at 14:14, on Zulip):

I wrote a comment there about how it would be great to have a reproducer that did not depend on plugin codegen, but I also suspect that it may not be possible to write such an example.

nagisa (Aug 01 2019 at 14:15, on Zulip):

Not sure what else is there to discuss.

nagisa (Aug 01 2019 at 14:16, on Zulip):

@Esteban Küber also noted a possible area in code where this is likely coming out from.

nagisa (Aug 01 2019 at 14:17, on Zulip):

FWIW using all sorts of procedural macros did lead to all weird error conditions in my experience, but never an ICE.

nagisa (Aug 01 2019 at 14:18, on Zulip):

/me removes I-nominated and moves on seeing as there are few opinions on the topic.

nagisa (Aug 01 2019 at 14:18, on Zulip):

libcore fails to compile for thumbv6m-none-eabi #62932 has had nice progress, LLVM bug reported, so skipping this one.

nagisa (Aug 01 2019 at 14:19, on Zulip):

Segfault compiling libc on nightly-armv7-unknown-linux-gnueabihf #62896 related to LLVM upgrade as well, but the root cause has not yet been pinned down.

nagisa (Aug 01 2019 at 14:20, on Zulip):

For this issue the backtraces are incorrect as they are for the SIGUSR1 signal, not sigsegv one.

nagisa (Aug 01 2019 at 14:20, on Zulip):

but people did narrow it down to LLVM.

nagisa (Aug 01 2019 at 14:21, on Zulip):

I suspect that this is gonna be a LLVM assertion

nagisa (Aug 01 2019 at 14:21, on Zulip):

would be good for somebody to get it.

nagisa (Aug 01 2019 at 14:24, on Zulip):

Wrote a comment to that effect.

nagisa (Aug 01 2019 at 14:25, on Zulip):

Undefined symbol _fltused when compiling to x86_64-unknown-uefi #62785 No update since last week, skip.

nagisa (Aug 01 2019 at 14:25, on Zulip):

Recent nightly versions break thumb* targets #62781

nagisa (Aug 01 2019 at 14:25, on Zulip):

Now this is a good issue to discuss our policy on

nagisa (Aug 01 2019 at 14:25, on Zulip):

namely, does adding or removing sections emitted by the compiler constitute a breaking change?

nagisa (Aug 01 2019 at 14:26, on Zulip):

I’m strongly of the opinion that it does not, but it appears that other people think otherwise.

nagisa (Aug 01 2019 at 14:26, on Zulip):

If we do consider that to be an acceptable change, then this issue is technically not a bug.

nagisa (Aug 01 2019 at 14:28, on Zulip):

It may make sense to ask T-lang to have their say on the topic.

nagisa (Aug 01 2019 at 14:31, on Zulip):

Okay, we have no backports to investigate

nagisa (Aug 01 2019 at 14:31, on Zulip):

and I just got a page to go and put out some fires.

nagisa (Aug 01 2019 at 14:31, on Zulip):

So I’ll wrap this meeting early.

oli (Aug 01 2019 at 14:34, on Zulip):

Thanks for doing pre-triage and driving @nagisa

nagisa (Aug 01 2019 at 14:43, on Zulip):

Thanks everybody for participating!

centril (Aug 01 2019 at 18:17, on Zulip):

@nagisa Did you want to nominate #62781 for the lang team?

nagisa (Aug 01 2019 at 18:18, on Zulip):

nagisa Did you want to nominate #62781 for the lang team?

Maybe. I don’t mind if it is not discussed ASAP as long as y'all ever get to it to make a decision :slight_smile:

centril (Aug 01 2019 at 18:18, on Zulip):

@nagisa I'll take that as a "yes, if you have time" :P

Last update: Nov 16 2019 at 01:05UTC