Internal Compiler Error on stable (Broken MIR)

[conradludgate]

Code

There's only so much that I can share. The code that seems to be erroring is similar to

enum Enum {
    Variant(Variant)
}

    tokio::spawn(async move {
        let Enum::Variant(value) = enum;
        dispatch(value).await.unwrap();
    });

Meta

rustc --version --verbose:

rustc 1.60.0 (7737e0b5c 2022-04-04)
binary: rustc
commit-hash: 7737e0b5c4103216d6fd8cf941b7ab9bdbaace7c
commit-date: 2022-04-04
host: aarch64-apple-darwin
release: 1.60.0
LLVM version: 14.0.0

rustc 1.62.0-nightly (69a5d2481 2022-04-27)
binary: rustc
commit-hash: 69a5d2481e856a5a18885390b8cf6950b9ff8dd3
commit-date: 2022-04-27
host: aarch64-apple-darwin
release: 1.62.0-nightly
LLVM version: 14.0.1

Error output

warning: Error finalizing incremental compilation session directory `.../target/debug/incremental/app-k3kpcf5j5syk/s-g9a4ygf568-1sdh6k2-working`: No such file or directory (os error 2)

error: internal compiler error: no errors encountered even though `delay_span_bug` issued

error: internal compiler error: broken MIR in DefId(0:2156 ~ app[24b6]::{closure#0}::{closure#0}) ((_33.0: Variant)): can't project out of PlaceTy { ty: Enum, variant_index: None }
  --> src/app/src/handler.rs:90:29
   |
90 |       tokio::spawn(async move {
   |  _____________________________^
91 | |         let Enum::Variant(value) = enum;
92 | |         dispatch(value).await.unwrap();
93 | |     });
   | |_____^
   |
   = note: delayed at compiler/rustc_borrowck/src/type_check/mod.rs:860:31

error: internal compiler error: TyKind::Error constructed but no error reported
  |
  = note: delayed at compiler/rustc_borrowck/src/type_check/mod.rs:795:20

error: internal compiler error: TyKind::Error constructed but no error reported
  |
  = note: delayed at /rustc/69a5d2481e856a5a18885390b8cf6950b9ff8dd3/compiler/rustc_middle/src/ty/relate.rs:419:59

error: internal compiler error: broken MIR in DefId(0:2159 ~ app[24b6]::{closure#0}::{closure#0}::{closure#2}) (((_1.0: Variant).0: Variant)): bad field access (uuid::Uuid: Variant): NoSolution
  --> src/app/src/handler.rs:91:26
   |
91 |         let Enum::Variant(value) = enum;
   |                           ^^^^^
   |
   = note: delayed at compiler/rustc_borrowck/src/type_check/mod.rs:770:29

Backtrace

thread 'rustc' panicked at 'Box<dyn Any>', compiler/rustc_errors/src/lib.rs:1347:13
stack backtrace:
   0:        0x104dc3e30 - <std::sys_common::backtrace::_print::DisplayBacktrace as core::fmt::Display>::fmt::h5766446765602b2f
   1:        0x104e15730 - core::fmt::write::h1454cf4bd27645af
   2:        0x104db6e88 - std::io::Write::write_fmt::h6d4bd601dc063fa4
   3:        0x104dc6b30 - std::panicking::default_hook::{{closure}}::he3278aaa6dd4ffe3
   4:        0x104dc680c - std::panicking::default_hook::h83712f1c80d5cc94
   5:        0x10c0cf59c - rustc_driver[ab005e2210e4776d]::DEFAULT_HOOK::{closure#0}::{closure#0}
   6:        0x104dc720c - std::panicking::rust_panic_with_hook::hb3ff747d28bd5b62
   7:        0x10fcb30ec - std[288a9308ea2c3692]::panicking::begin_panic::<rustc_errors[300f3d566adbb64d]::ExplicitBug>::{closure#0}
   8:        0x10fcb30a4 - std[288a9308ea2c3692]::sys_common::backtrace::__rust_end_short_backtrace::<std[288a9308ea2c3692]::panicking::begin_panic<rustc_errors[300f3d566adbb64d]::ExplicitBug>::{closure#0}, !>
   9:        0x10ff847f4 - std[288a9308ea2c3692]::panicking::begin_panic::<rustc_errors[300f3d566adbb64d]::ExplicitBug>
  10:        0x10fcbe0b8 - std[288a9308ea2c3692]::panic::panic_any::<rustc_errors[300f3d566adbb64d]::ExplicitBug>
  11:        0x10fcc21d0 - <rustc_errors[300f3d566adbb64d]::HandlerInner as core[30328a0cdeb81e9a]::ops::drop::Drop>::drop
  12:        0x10c0e364c - core[30328a0cdeb81e9a]::ptr::drop_in_place::<rustc_session[cb45daf2430bcebe]::parse::ParseSess>
  13:        0x10c0e6c50 - <alloc[ba9ec2f9f2fb29e6]::rc::Rc<rustc_session[cb45daf2430bcebe]::session::Session> as core[30328a0cdeb81e9a]::ops::drop::Drop>::drop
  14:        0x10c0d6498 - core[30328a0cdeb81e9a]::ptr::drop_in_place::<rustc_interface[1aaaf20b10225d55]::interface::Compiler>
  15:        0x10c0d454c - rustc_span[cfdc8e2acaf4a144]::with_source_map::<core[30328a0cdeb81e9a]::result::Result<(), rustc_errors[300f3d566adbb64d]::ErrorGuaranteed>, rustc_interface[1aaaf20b10225d55]::interface::create_compiler_and_run<core[30328a0cdeb81e9a]::result::Result<(), rustc_errors[300f3d566adbb64d]::ErrorGuaranteed>, rustc_driver[ab005e2210e4776d]::run_compiler::{closure#1}>::{closure#1}>
  16:        0x10c07c720 - rustc_interface[1aaaf20b10225d55]::interface::create_compiler_and_run::<core[30328a0cdeb81e9a]::result::Result<(), rustc_errors[300f3d566adbb64d]::ErrorGuaranteed>, rustc_driver[ab005e2210e4776d]::run_compiler::{closure#1}>
  17:        0x10c0758f0 - <scoped_tls[726aaecfe4f649cb]::ScopedKey<rustc_span[cfdc8e2acaf4a144]::SessionGlobals>>::set::<rustc_interface[1aaaf20b10225d55]::interface::run_compiler<core[30328a0cdeb81e9a]::result::Result<(), rustc_errors[300f3d566adbb64d]::ErrorGuaranteed>, rustc_driver[ab005e2210e4776d]::run_compiler::{closure#1}>::{closure#0}, core[30328a0cdeb81e9a]::result::Result<(), rustc_errors[300f3d566adbb64d]::ErrorGuaranteed>>
  18:        0x10c0c286c - std[288a9308ea2c3692]::sys_common::backtrace::__rust_begin_short_backtrace::<rustc_interface[1aaaf20b10225d55]::util::run_in_thread_pool_with_globals<rustc_interface[1aaaf20b10225d55]::interface::run_compiler<core[30328a0cdeb81e9a]::result::Result<(), rustc_errors[300f3d566adbb64d]::ErrorGuaranteed>, rustc_driver[ab005e2210e4776d]::run_compiler::{closure#1}>::{closure#0}, core[30328a0cdeb81e9a]::result::Result<(), rustc_errors[300f3d566adbb64d]::ErrorGuaranteed>>::{closure#0}, core[30328a0cdeb81e9a]::result::Result<(), rustc_errors[300f3d566adbb64d]::ErrorGuaranteed>>
  19:        0x10c09c520 - <<std[288a9308ea2c3692]::thread::Builder>::spawn_unchecked_<rustc_interface[1aaaf20b10225d55]::util::run_in_thread_pool_with_globals<rustc_interface[1aaaf20b10225d55]::interface::run_compiler<core[30328a0cdeb81e9a]::result::Result<(), rustc_errors[300f3d566adbb64d]::ErrorGuaranteed>, rustc_driver[ab005e2210e4776d]::run_compiler::{closure#1}>::{closure#0}, core[30328a0cdeb81e9a]::result::Result<(), rustc_errors[300f3d566adbb64d]::ErrorGuaranteed>>::{closure#0}, core[30328a0cdeb81e9a]::result::Result<(), rustc_errors[300f3d566adbb64d]::ErrorGuaranteed>>::{closure#1} as core[30328a0cdeb81e9a]::ops::function::FnOnce<()>>::call_once::{shim:vtable#0}
  20:        0x104dcfadc - std::sys::unix::thread::Thread::new::thread_start::ha367c009d033f6c1
  21:        0x1908e626c - __pthread_deallocate

[conradludgate]

I've discovered that destructuring the enum variant outside of the async block fixes the issue

[conradludgate]

Minimal repro: https://play.rust-lang.org/?version=stable&mode=debug&edition=2021&gist=9ffdec06b38e6986b37aa92e9d2955f3

[hellow554]

@conradludgate amazing that you did reduce your testcase like that.
I removed the external dependecy to tokio and silenced the unused warning:

struct Variant;

enum Enum {
    Variant(Variant),
}

fn spawn<T>(_: T)
where
    T: std::future::Future + Send + 'static,
    T::Output: Send + 'static,
{
}

async fn dispatch(_: Variant) {}

fn main() {
    let eh = Enum::Variant(Variant);

    spawn(async move {
        let Enum::Variant(value) = eh;
        dispatch(value).await;
    });
}

Regressed in rollup #91555 probably in #91450

---

searched nightlies: from nightly-2021-12-01 to nightly-2022-04-24
regressed nightly: nightly-2021-12-06
searched commits: from efec545 to e2116ac
regressed commit: 772d51f

bisected with cargo-bisect-rustc v0.6.0

Host triple: x86_64-unknown-linux-gnu
Reproduce with:

cargo bisect-rustc --end=2022-04-24 --without-cargo --access=github --regress=ice

[tmiasko]

Reduced a bit more (requires an edition with precise captures, I think):

struct Struct;
enum Enum { Variant(Struct) }
fn main() {
    let _enum = Enum::Variant(Struct);
    || {
        let Enum::Variant(_value) = _enum;
    };
}

[matthiaskrgr]

Looks like a duplicate of #96299 and #95271

[lqd]

These started ICEing in nightly-2020-12-13 (GH range), with #79553 looking very likely.

[conradludgate]

If no one is yet investigating this, I'd like to pick it up

[conradludgate]

So far, I've found 2 issues that cause this behaviour.

The first is that it's having trouble in this context to get the field type of the single variant enum:

rust/compiler/rustc_borrowck/src/type_check/mod.rs

Line 822 in 61469b6

ty::Adt(adt_def, substs) if !adt_def.is_enum() => {

Fix is easy,

- ty::Adt(adt_def, substs) if !adt_def.is_enum() => {
+ ty::Adt(adt_def, substs) if adt_def.variants().len() == 1 => {

The second issue is that the MIR is being badly formed. Looking at the traces, I see

│ │ │ │ ├─┐rustc_borrowck::type_check::field_ty parent=((_1.0: Struct).0: Struct), base_ty=PlaceTy { ty: Struct, variant_index: None }, field=field[0], location=bb0[1]
│ │ │ │ │ ├─0ms DEBUG rustc_borrowck::type_check kind=Adt(Struct, [])
│ │ │ │ │ ├─0ms DEBUG rustc_borrowck::type_check adt_def=Struct, substs=[], variants=[VariantDef { def_id: DefId(0:3 ~ ice_96512[22c4]::Struct), ctor_def_id: Some(DefId(0:4 ~ ice_96512[22c4]::Struct::{constructor#0})), name: "Struct", discr: Relative(0), fields: [], ctor_kind: Const, flags: NO_VARIANT_FLAGS }]
│ │ │ │ │ ├─0ms DEBUG rustc_borrowck::type_check substs=[], variant=VariantDef { def_id: DefId(0:3 ~ ice_96512[22c4]::Struct), ctor_def_id: Some(DefId(0:4 ~ ice_96512[22c4]::Struct::{constructor#0})), name: "Struct", discr: Relative(0), fields: [], ctor_kind: Const, flags: NO_VARIANT_FLAGS }
│ │ │ │ ├─┘

Which hints to me that the code is not being read as

let Enum::Variant(_value) = _enum;

but instead

let Enum::Variant(Struct(_value)) = _enum;

[tmiasko]

The second issue is that the MIR is being badly formed. Looking at the traces, I see

Yeah, it looks like constructed MIR is incorrect.

One issue seems to be in to_upvars_resolved_place_builder, towards the end it has to rebase projections on top of a capture. First it has to remove the captured prefix, and the implementation makes the assumption that the number of projections in HIR and MIR will coincide. In the HIR downcasting is part of field projection, in the MIR those are two separate elements. Thus we are left with an extra incorrect projection.

There seems to be one more issue when constructing the capture - the downcasting operation is completely missing.

While the documentation for MIR ProjectionElem::Downcast claims that

Currently, we only introduce this for ADTs with more than one variant. It may be better to just introduce it always, or always for enums.

I think that is an outdated information.

[conradludgate]

I also just noticed the Downcast issue. In my traces, I do see the variant projection using this Downcast for just a single variant, and the suspect PR changes does seem to ignore all Downcast projections

(there's a little bit of discussion taking place on zulip https://rust-lang.zulipchat.com/#narrow/stream/182449-t-compiler.2Fhelp/topic/reading.20MIR.20.28.2396512.29/near/280855883)

[conradludgate]

Should the fix be

1. Keeping Downcast and truncation the capture correctly
2. Making the single variant capture not use Downcast