Lots of warnings when compiling

[porglezomp]

There are 82 warnings when compiling.

[porglezomp]

It looks like almost every single one can be resolved by setting the structs to #[repr(C)], is there any reason not to do this?

It looks like what it primarily does is tell the compiler that it's going to be used by foreign functions, do you know if there are any downsides?

[kornelski]

Good idea. Done.

[porglezomp]

You missed the it on

pub struct Error(pub c_uint);

When I add it there, there are no more warnings at all.

[kornelski]

Odd, I'm not getting any warnings in rustc 0.12.0-pre-nightly (85fd37f87). The error "struct" is Rust-private, so it shouldn't have repr(C).

[porglezomp]

I'm using rustc 0.13.0-nightly (b6e0d3a5b 2014-10-16 09:12:17 +0000), so perhaps that's the problem. It does seem odd, perhaps we found a regression?

Nonetheless, I'm getting

/Users/caleb/Development/lodepng-rust/src/lib.rs:196:41: 196:46 warning: found type without foreign-function-safe
                                      representation annotation in foreign module, consider adding a #[repr(...)] attribute to the type, #[warn(improper_ctypes)] on by default
/Users/caleb/Development/lodepng-rust/src/lib.rs:196         pub fn lodepng_error_text(code: Error) -> &'static i8;
                                                                                             ^~~~~

In addition to a lot more ones, all pointing at the Error struct.

[kornelski]

Ah, I forgot it's passed to lodepng_error_text. I guess the compiler is smarter than me ;)

[porglezomp]

It is also complaining about all the return values, but it might be complaining only because of that one.

[kornelski]

I've upgraded rustc, and with one more tweak dda3c78 it builds cleanly for me.