ref pattern nested inside let LHS causes LLVM assertion failure

[nejucomo]

The following code (which may not be correct rust) causes an llvm abort on my build.

$ cat ./experiment.rs

enum PureCounter { PureCounter(uint) }

impl PureCounter : iter::BaseIter<uint> {
    pure fn each(blk: fn(v: &uint) -> bool) {
        let PureCounter(ref x) = self;
        match blk(x) {
          true => PureCounter( (*x) + 1 ).each(blk),
          false => ()
        }
    }

    pure fn size_hint() -> Option<uint> { None }
}

fn main() {
    do PureCounter(0).each |nbox| {
        io::println( (*nbox).to_str() );
        (*nbox) < 5
    }
}

$ rustc ./experiment.rs
rustc: /home/n/sandbox/3p/rust/src/llvm/lib/VMCore/Instructions.cpp:1062: void llvm::StoreInst::AssertOK(): Assertion `getOperand(0)->getType() == cast<PointerType>(getOperand(1)->getType())->getElementType() && "Ptr must be a pointer to Val type!"' failed.
Aborted

$ rustc --version
rustc 0.5 (07edf90 2012-10-13 05:57:13 -0700)
host: x86_64-unknown-linux-gnu

[catamorphism]

Minimized version:

enum PureCounter { PureCounter(uint) }

pure fn each(self: PureCounter, blk: fn(v: &uint)) {
    let PureCounter(ref x) = self;
    blk(x);
}

fn main() {}

I suspect ref patterns in fields in let patterns don't work right. This will be addressed when I finish fixing #3235.

Changing the title to something more informative.

[catamorphism]

Bumping to 0.6 since #3235 is not done.

[martica]

I'm not sure if this is the same bug, but I'm getting the same assert with this code (I've minimized it as much as I can):

macro_rules! a_macro {
    ($zero:expr) => {
        pub fn foo() -> int {
            let x:int = $zero;
            x + $zero.add(&1)
        }
    }
}

a_macro!(0i)

fn main() { }

The code works when it has been manually expanded, so it appears to related to the use of a macro.

[catamorphism]

Reproduced as of 373504, but not critical for 0.6; de-milestoning

[catamorphism]

Nominating for milestone 5 (production-ready)

[nikomatsakis]

I dusted off my branch recently and am looking at this again.

[alexcrichton]

Closed by #7262