ICE due to missing associated type in trait implementation

[japaric]

STR

Didn't have time to write a shorter snippet

#![crate_type = "lib"]
#![feature(associated_types, lang_items, unboxed_closures)]
#![no_std]

use Option::{None, Some};

trait Iterator {
    type Item;

    fn next(&mut self) -> Option<Self::Item>;
}

trait DoubleEndedIterator: Iterator {
    fn next_back(&mut self) -> Option< <Self as Iterator>::Item>;
}

struct Rev<I>(I);

impl<I> Iterator for Rev<I> where I: DoubleEndedIterator {
    // forgot this!
    //type Item = <I as Iterator>::Item;

    fn next(&mut self) -> Option< <I as Iterator>::Item> {
        self.0.next_back()
    }
}

#[lang = "copy"] trait Copy {}
#[lang = "sized"] trait Sized {}

enum Option<T> {
    None,
    Some(T),
}

#[lang = "fn_mut"]
trait FnMut<Args, Result> {
    extern "rust-call" fn call_mut(&mut self, Args) -> Result;
}

Output

iter.rs:23:5: 25:6 error: internal compiler error: impl `VtableImpl(impl_def_id=DefId { krate: 0, node: 38 }:Rev<I>.Iterator, substs=Substs[types=[[_];[];[]], regions=[[];[];[]]], nested=[[Obligation(predicate=Binder(TraitPredicate(TraitRef(I, Sized))),depth=1), Obligation(predicate=Binder(TraitPredicate(TraitRef(I, DoubleEndedIterator))),depth=1)];[];[]])` did not contain projection for `Obligation(predicate=<Rev<I> as TraitRef(Rev<I>, Iterator)>::Item,depth=0)`
iter.rs:23     fn next(&mut self) -> Option< <I as Iterator>::Item> {
iter.rs:24         self.0.next_back()
iter.rs:25     }
note: the compiler unexpectedly panicked. this is a bug.
note: we would appreciate a bug report: http://doc.rust-lang.org/complement-bugreport.html
note: run with `RUST_BACKTRACE=1` for a backtrace
thread 'rustc' panicked at 'Box<Any>', /root/rust/src/libsyntax/diagnostic.rs:123

stack backtrace:
   1:     0x7fddb1e70120 - sys::backtrace::write::h3fe61c98bf398fe8P8s
   2:     0x7fddb1e8fef0 - failure::on_fail::h55da319a91ee267cEpz
   3:     0x7fddb1e05b40 - rt::unwind::begin_unwind_inner::h2a51826b093fcd9eJ3y
   4:     0x7fddad24b3d0 - rt::unwind::begin_unwind::h3298842058425057001
   5:     0x7fddad24b360 - diagnostic::SpanHandler::span_bug::heeaf2edd6daa003fNQF
   6:     0x7fddb022acb0 - middle::traits::project::project_type::hcf50d7eec736df9c35P
   7:     0x7fddb0222800 - middle::traits::fulfill::process_predicate::h3703aabc133e79cayGP
   8:     0x7fddb0221ab0 - middle::traits::fulfill::FulfillmentContext<$u{27}tcx$GT$::select::h827e4d042486e566SzP
   9:     0x7fddb00ee7c0 - middle::traits::fulfill::FulfillmentContext<$u{27}tcx$GT$::select_all_or_error::heebbe09eddbd26dfDwP
  10:     0x7fddb1510cd0 - check::compare_impl_method::h359a8a44e8f8def4Zfk
  11:     0x7fddb14ff0b0 - check::check_item::hfedb86714512d4d8XRj
  12:     0x7fddb169d450 - check_crate::unboxed_closure.39783
  13:     0x7fddb16981f0 - check_crate::h7446c5344d10b3c1cGx
  14:     0x7fddb23a98d0 - driver::phase_3_run_analysis_passes::hb59d2d67157ec124gva
  15:     0x7fddb2397fb0 - driver::compile_input::hff9e8d16e3108315vba
  16:     0x7fddb24e4900 - thunk::F.Invoke<A,$u{20}R$GT$::invoke::h10803646332669730367
  17:     0x7fddb24e36c0 - rt::unwind::try::try_fn::h3820096971255462672
  18:     0x7fddb1ef4eb0 - rust_try_inner
  19:     0x7fddb1ef4ea0 - rust_try
  20:     0x7fddb24e3a10 - thunk::F.Invoke<A,$u{20}R$GT$::invoke::h15250775183210022334
  21:     0x7fddb1e7f8d0 - sys::thread::thread_start::h7b82ef93cab3e580K1v
  22:     0x7fddaca690c0 - start_thread
  23:     0x7fddb1aab2d9 - __clone
  24:                0x0 - <unknown>

Version

84f5ad8

cc @nikomatsakis

[nikomatsakis]

The fix here is just to remove that assert and instead generate a ty::ty_err, I think. We can rely on other phases of the compiler to generate a suitable error.

[nikomatsakis]

Adding to 1.0 alpha milestone. This is something people hit a lot.

[nrc]

This seems to be fixed, when running the above snippet, I get "not all trait items implemented, missing: Item [E0046]"

[nikomatsakis]

And yet people keep hitting ICEs? Maybe some of the other examples cause trouble?

[nikomatsakis]

But I can't reproduce it either. Maybe they're just using older compilers?

[nikomatsakis]

Ah, no, reading the code, I see I landed a patch for this. Argh. Sorry, so much going on I just forgot.