Port "Fcdr()" and "cdr()" to Rust

[TheKK]

No description provided.

[Wilfred]

Apart from a teeny-tiny docstring issue, looks great! :)

[Wilfred]

Minor nit: could you remove the space before Return?

[TheKK]

Fixed!

[TheKK]

Sorry for bringing this up but I'd like to ask some questions about implementation details (hope they're not too trivial):

1. I saw how Lisp_Cons implemented in C and known there's an union inside it. My question is, how

#[repr(C)]
#[allow(unused_variables)]
struct LispCons {
    car: LispObject,
    cdr: LispObject,
}

#[repr(C)]
#[allow(dead_code)]
pub struct LispConsChain {
    chain: *const LispCons,
}

achieves the same behaviour of C union? Does this mean we'll use mem::transmute to cast LispCons.cdr into LispConsChain.chain?

2. Is there any concern exposing fn car(object: LispObject) -> LispObject in cons.rs to C code?

3. In practical, what's the strategy while choosing between

unsafe fn Foo() { ... }

and

fn Foo() { unsafe { } }

[Wilfred]

1: Yep, exactly. We'll use mem::transmute to cast from LispCons.cdr to LispConsChain.cdr. A struct is just a blob of bits, and since they're the same size this is fine.

I'm open to suggestions of cleaner approaches, but I think the current approach is reasonable.

2: car is currently not exposed to the C code, it's a function private to cons.rs. Note that it's not extern "C". We could rename it to CAR and replace the CAR inline function lisp.h. I don't have any concerns with us doing so.

3: See #63, but do let me know if it's unclear.