needless_doctest_main vs. question mark operator

[niklasf]

I am currently using main with return type Result<(), Box<Error>> in order to be able to use the question mark operator in examples.

This is recommended in the API guidelines: https://rust-lang.github.io/api-guidelines/documentation.html#examples-use--not-try-not-unwrap-c-question-mark

Clippy should not suggest removing main() in this case.

/// # Example
///
/// ```
/// # use std::error::Error;
/// #
/// # fn try_main() -> Result<(), Box<Error>> {
/// use shakmaty::fen::Fen;
///
/// let fen = Fen::from_ascii(b"rnbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/RNBQKBNR w KQkq - 0 1")?;
/// assert_eq!(fen, Fen::default());
/// #
/// #     Ok(())
/// # }
/// #
/// # fn main() {
/// #     try_main().unwrap();
/// # }
/// ```

[niklasf]

Sorry, nevermind. This was only happening because I was still using the old try_main pattern.

[llogiq]

Yeah, we should detect this. You can just do the following:

/// # Example
///
/// ```
/// # use std::error::Error;
/// use shakmaty::fen::Fen;
///
/// let fen = Fen::from_ascii(b"rnbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/RNBQKBNR w KQkq - 0 1")?;
/// assert_eq!(fen, Fen::default());
/// # Ok::<(), Box<Error>(())
/// ```

[niklasf]

Oh, suggesting that would be nice. Reopening.

[xiongmao86]

I am a little bit confused, what does this issue suggest we do, to detect and lint for appearance of main(), or not to detect and lint for appearance of main() in this case?

[xiongmao86]

@llogiq, do you suggest detecting main() -> Result<(), Box<Error>> { ?

[llogiq]

We could certainly do that. I have a branch open that should be easier to extend in that regard, but it failed CI for some reason.

[xiongmao86]

Oh, thanks for replying.

[pickfire]

@llogiq are you working on this?