adds anyhow where result used to be

[eureka-cpu]

Intended to close #851

I've gone through and replaced any instance I could find of Result<(), String> with anyhow, along with error messages.

[JoshuaBatty]

Nice, looks good thanks. Just left some comments to help you get the tests working again.

There does still remain the question if we should be considering using thiserror instead of anyhow if we are exposing forc as a library. But this is still an upgrade from the current error handling and can be addressed in a separate PR if we decide to move to thiserror.

[mitchmindtree]

Ahh was hoping to have this addressed after landing #825 but I should have mentioned so in the issue 🥲 if this happens to land first I can rebase on top no worries 👍

---

I would like to see thiserror with more granular types eventually, but I think separating the parts of forc that we wish to expose as a library more clearly first might help to go about this, I'll open an issue with some thoughts on this. For now though I think anyhow and this PR makes a lot of sense!

[JoshuaBatty]

I think it probably looked tidier the way you first had it, eg:

assert_eq!(
    contains_invalid_char("test#proj", "package name").map_err(|e| e.to_string()),
    std::result::Result::Err(
        "invalid character `#` in package name: `test#proj`, \
            characters must be Unicode XID characters \
            (numbers, `-`, `_`, or most letters)"
            .into()
    )
);

[JoshuaBatty]

LGTM, thanks for addressing those changes.