Allow dead_code in tests

[Grinkers]

Just less noise.

[naomijub]

I'm OK with the change but my suggestion is to create a tests/lib.rs and add a single #![allow(dead_code)] there

[Grinkers]

I'm OK with the change but my suggestion is to create a tests/lib.rs and add a single #![allow(dead_code)] there

I believe tests/ is a special case. Each mod inside ends up being compiled as a completely separate crate, so this doesn't work. I actually ran into this issue with edn-rs trying to add a helper function for all mods https://github.com/edn-rs/edn-rs/blob/3965bc3221d79b883a7059d7ce72c4c5c2eee366/tests/deserialize.rs#L10. The workarounds involve directly importing rust files, which is just way more ugly than a couple repeated snippets.

Reference: https://stackoverflow.com/questions/44539729/what-is-an-idiomatic-way-to-have-shared-utility-functions-for-integration-tests

It's further weird with the crate trybuild being used, which ends up in a completely separate compile process for each file.
https://github.com/edn-rs/edn-derive/blob/main/tests/progress.rs

[Grinkers]

Edited my above link. Also found what I was originally thinking

https://doc.rust-lang.org/book/ch11-03-test-organization.html#the-tests-directory

Each file in the tests directory is a separate crate

[Grinkers]

I was able to clean up edn-rs's lib.rs. I don't think it's really any more clean, but better for long-term usage. I don't think this would work with attributes here, especially with trybuild. It's really quite weird how tests/ gets treated so differently.

naomijub/edn-rs@7efaab5

[Grinkers]

@naomijub @evaporei
merging unless either of you have any other ideas. Adding a lib.rs is a no-op and I don't think trybuild can inject attributes. Each test is actually its own binary.