resolving bounds after type-checking: predict trait for std::ops::FnMut

[clouds56]

Code

playgound

pub trait Trait1 {
  type C;
}

struct T1;
impl Trait1 for T1 {
  type C = usize;
}
pub trait Callback<T: Trait1>: FnMut(<T as Trait1>::C) {}
impl<T: Trait1, F: FnMut(<T as Trait1>::C)> Callback<T> for F { }


pub struct State<T: Trait1> {
  callback: Option<Box<dyn Callback<T>>>,
}
impl<T: Trait1> State<T> {
  fn new() -> Self { Self { callback: None } }
  fn test_cb(&mut self, d: <T as Trait1>::C) {
    (self.callback.as_mut().unwrap())(d)
  }
}

fn main() {
  let mut s = State::<T1>::new();
  s.test_cb(1);
}

Meta

rustc --version --verbose:

rustc 1.42.0 (b8cedc004 2020-03-09)
binary: rustc
commit-hash: b8cedc00407a4c56a3bda1ed605c6fc166655447
commit-date: 2020-03-09
host: x86_64-apple-darwin
release: 1.42.0
LLVM version: 9.0

Error output

error: internal compiler error: src/librustc_trait_selection/traits/codegen/mod.rs:107: Encountered errors `[FulfillmentError(Obligation(predicate=Binder(TraitPredicate(<dyn Callback<T1, Output = ()> as std::ops::FnMut<(usize,)>>)), depth=1),Unimplemented)]` resolving bounds after type-checking

thread 'rustc' panicked at 'Box<Any>', src/librustc_errors/lib.rs:880:9
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace

note: the compiler unexpectedly panicked. this is a bug.

note: we would appreciate a bug report: https://github.com/rust-lang/rust/blob/master/CONTRIBUTING.md#bug-reports

note: rustc 1.44.0-nightly (537ccdf3a 2020-04-02) running on x86_64-apple-darwin

note: compiler flags: -C debuginfo=2 -C incremental --crate-type bin

note: some of the compiler flags provided by cargo are hidden

Backtrace

thread 'rustc' panicked at 'Box<Any>', src/librustc_errors/lib.rs:873:9
stack backtrace:
   0: <unknown>
   1: <unknown>
   2: <unknown>
   3: <unknown>
   4: <unknown>
   5: <unknown>
   6: <unknown>
   7: <unknown>
   8: <unknown>
   9: <unknown>
  10: <unknown>
...

query stack during panic:
#0 [codegen_fulfill_obligation] checking if `std::ops::FnMut` fulfills its obligations
#1 [collect_and_partition_mono_items] collect_and_partition_mono_items
end of query stack

the backtrace itself is all <unknown>

[clouds56]

It's strange, I tried these code in main

fn main() {
  // let mut s = State::<T1>::new();
  // s.test_cb(1);
  let cb: Option<Box<dyn Callback<T1>>> = None;
  (cb.unwrap())(1);
}

and it complains about the trait `std::ops::FnMut<(usize,)>` is not implemented for `dyn Callback<T1, Output = ()>` (Compile Error but not ICE)
I think the trait Callback<T1, Output = ()> is depends on std::ops::FnMut<(usize,)> and dyn could resolve this.

[spastorino]

Assigning P-medium and removing nomination. This was discussed as part of pre-triage meeting in Zulip.

This one also sounds a bit similar to #69892 and #27675 which is very old. Would be nice to find if there are duplicates too.

[spastorino]

Pinging cleanup to find duplicates ...

@rustbot ping cleanup

[rustbot]

Hey Cleanup Crew ICE-breakers! This bug has been identified as a good
"Cleanup ICE-breaking candidate". In case it's useful, here are some
instructions for tackling these sorts of bugs. Maybe take a look?
Thanks! <3

cc @AminArria @chrissimpkins @contrun @DutchGhost @elshize @ethanboxx @h-michael @HallerPatrick @hdhoang @hellow554 @imtsuki @jakevossen5 @kanru @KarlK90 @LeSeulArtichaut @MAdrianMattocks @matheus-consoli @mental32 @nmccarty @Noah-Kennedy @pard68 @PeytonT @pierreN @Redblueflame @RobbieClarken @RobertoSnap @robjtede @SarthakSingh31 @senden9 @shekohex @sinato @spastorino @turboladen @woshilapin @yerke

[JohnTitor]

Triage: It's no longer ICE with the latest nightly, I think it's fixed by #77720.