Automatic adaption to 64bit architectures in guest code

[cpetig]

No description provided.

[pavelsavara]

cc @jsturtevant for C#

[cpetig]

Sorry for the extensive patch, basically this changes the Rust and C code generation (in this case passing a list) to

#[cfg_attr(target_pointer_width = "64", repr(align(8)))]
#[cfg_attr(target_pointer_width = "32", repr(align(4)))]
struct RetArea(
    [::core::mem::MaybeUninit<u8>; 2 * core::mem::size_of::<*const u8>()],
);
let mut ret_area = RetArea(
    [::core::mem::MaybeUninit::uninit(); 2 * core::mem::size_of::<*const u8>()],
);

*base.add(core::mem::size_of::<*const u8>()).cast::<usize>() = len0;
*base.add(0).cast::<*mut u8>() = ptr0.cast_mut();

[cpetig]

cc @jsturtevant for C#

I wasn't able to output the proper code for C# because I don't know enough of it, but Rust or C should give a good template.

[cpetig]

The publish test will be fixed by the next wasm-tools release (blocked there)

[cpetig]

Also part of #1098

[alexcrichton]

This all looks great to me, thanks! I'll work on getting a wasm-tools release out next week in order to merge this.

[alexcrichton]

The wasm-tools update is at bytecodealliance/wasm-tools#2054 which is coupled with some extra testing in Wasmtime. I also want to do some more testing in this repository before publishing to ensure that not too much is broken here.

[alexcrichton]

Ok the wasm-tools update should be going through in #1167

[alexcrichton]

Ok this should be mergable after a rebase I believe.