Document the WASM build process for the core

[ianthetechie]

This task covers the bare minimum to get a wasm module. I think this is pretty straightforward, but we need to decide what looks like the best option + document our choices.

• Do developers doing WASM builds need to have a target installed? If so, consider adding the target tuple to rust-toolchain.toml in the project root.
• Do you need to have any other tools installed?
• Do we use a tool to make it easier? I'd look at cargo-wasi; seems to make things easier?
• Document findings in the guide (mdbook in the repository: https://github.com/stadiamaps/ferrostar/tree/main/guide)

[ianthetechie]

I think GitHub's anti-spam is kicking in so I can't assign you yet @CatMe0w, but this is here for your reference. Maybe if you leave a comment on the issue I can assign it?

[CatMe0w]

Looks like I can assign myself. But I'll have to check the tasks later.

[CatMe0w]

Will we use WASI? Ferrostar Rust core is more like a "pure computation" library in my opinion and doesn't interact with system interfaces by itself. A basic, minimal wasm32-unknown-unknown target should be enough.

[ianthetechie]

Yes, excellent point :) I think wasm32-unknown-unknown is probably enough, and WASI probably isn't needed.

[ianthetechie]

Just tried this out myself @CatMe0w and remembered why I suggested WASI ;) We would need to support no-std environments first. See #26 which tracks this. We can compile the core today with WASI though.

[ianthetechie]

I've been mulling this over a bit. Might actually need to go ahead and make the core not require WASI as that is possible but not easy to run in a web browser right now. I'll give that a bit of thought. Let me know if you've turned up anything as well @CatMe0w.

[ianthetechie]

I think this will be closed by #120. There shouldn't be much else needed to get going this summer; building is a single command, and if you use rustup, it should automatically install the needed toolchains.