Typoing dot instead of comma in println! gives a warning at the wrong place

[vivia]

use std::time::SystemTime;

fn main() {
    println!("Hello, {:?}!". SystemTime::now());
}

says:

error: expected token: `,`
 --> src/main.rs:4:40
  |
4 |     println!("Hello, {:?}!". SystemTime::now());
  |                                        ^^ expected `,`

... which isn't exactly where the , was expected.

(Playground)

Another variation is this code:

fn random_xkcd_number () -> i32 {
    4
}

fn main() {
    println!("Hello, {}!". random_xkcd_number());
}

which gives this result:

error: format argument must be a string literal
 --> src/main.rs:6:14
  |
6 |     println!("Hello, {}!". random_xkcd_number());
  |              ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  |
help: you might be missing a string literal to format with
  |
6 |     println!("{}", "Hello, {}!". random_xkcd_number());
  |              ^^^^^

(Playground)

[scrabsha]

Hello,
I am a quite new contributor, and I'd like to tackle this issue. Do you think that is possible?

I have looked a bit at the code, and I think this mostly comes from the parse_args function, more precisely, this condition can be tweaked. Is that correct?

[scrabsha]

I have looked a bit at the code, and I think this mostly comes from the parse_args function, more precisely, this condition can be tweaked. Is that correct?

Turns out I was terribly wrong. The problem is that the following snippet:

println!("A number: {}". std::iter::once(42).next().unwrap());

Can be rewritten as:

println!("A number: {}".std ::iter::once(42).next().unwrap());

And that the parse_args function looks for an expression as the first argument (corresponding code, code example). Consequently, "A number: {}".std is the said expression, which explain why the error is raised on the first ::.

Now that I know where the error comes from. I think the best way to fix this issue is to check if this expression can be a format string, perhaps with the expr_to_spanned_string function. I am still unsure about the quality of the emitted error, but at least it would have the correct span.