ICE with trait inheritance/generics

[brendanzab]

numeric.rs

#[link(name = "numeric",
       vers = "0.1")];
#[crate_type = "lib"];

pub trait Trig<T> {
    pure fn sin(&self) -> T;
}

pub pure fn sin<T:Trig<R>, R>(theta: &T) -> R { theta.sin() }

pub trait Angle<T>: Trig<T> {}

main.rs

extern mod numeric;
use numeric::*;

fn foo<T, A:Angle<T>>(theta: A) -> T { sin(&theta) }

fn main() {}

$ rustc numeric.rs                           
warning: no debug symbols in executable (-arch x86_64)
$ RUST_LOG=rustc=1,::rt::backtrace rustc -L . main.rs
rust: task failed at 'Key not found in table: 28', /usr/local/src/rust/src/librustc/rustc.rc:1
error: internal compiler error: unexpected failure
note: the compiler hit an unexpected failure path. this is a bug
note: try running with RUST_LOG=rustc=1,::rt::backtrace to get further details and report the results to github.com/mozilla/rust/issues
rust: task failed at 'explicit failure', /usr/local/src/rust/src/librustc/rustc.rc:481
rust: domain main @0x7fa62380bc10 root task failed
rust: task failed at 'killed', /usr/local/src/rust/src/libcore/task/mod.rs:570

[brson]

There are at least two problems here. The first is that the order in which type parameters are specified appears to be significant. This test fails:

pub trait Trig<T> {
    pure fn sin(&self) -> T;
}

pub trait Angle<T>: Trig<T> {}

pub pure fn sin<A:Angle<R>, R>(theta: &A) -> R { theta.sin() }

fn foo<R, A:Angle<R>>(theta: A) -> R { sin(&theta) }

impl int: Trig<int> {
    pure fn sin(&self) -> int { 10 }
}

impl int: Angle<int> {
}

fn main() {
    assert foo(0) == 10;
}

but if you switch the type parameter order on sin to <R, A:Angle<R> to match foo it works.

The second is some cross-crate issue with inheritance.

[brson]

My above example is a different problem, since it makes sin have a bound of <T:Angle> instead of <T:Trig>. I'm coming up with a lot of variations that fail in different ways. I'll open another issue.

[brendanzab]

@catamorphism (nag nag)

[catamorphism]

This ICEs with a different error now, but still ICEs.

[catamorphism]

Unfortunately, I got derailed and this won't be done for 0.6. De-milestoning.

[catamorphism]

This compiles now. I'll check in the test case.

[msullivan]

So is this fixed?

[catamorphism]

(bug triage) Huh, I wonder whether I checked in the test case. I'll look at that tomorrow.

[catamorphism]

#9233 will close this.