For E0277 suggest adding Result return type for function when using QuestionMark ? in the body.

[surechen]

Adding suggestions for following function in E0277.

fn main() {
    let mut _file = File::create("foo.txt")?;
}

to

fn main() -> Result<(), Box<dyn std::error::Error>> {
    let mut _file = File::create("foo.txt")?;

    return Ok(());
}

According to the issue #125997, only the code examples in the issue are targeted, but the issue covers a wider range of situations.

[rustbot]

r? @oli-obk

rustbot has assigned @oli-obk.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

[oli-obk]

Please add this information as a doc comment to the span field.

[surechen]

Please add this information as a doc comment to the span field.

Thank you.
Sorry, I didn't understand, what should I do?

[oli-obk]

Add documentation to https://doc.rust-lang.org/nightly/nightly-rustc/rustc_hir/hir/struct.Block.html#structfield.span stating that this span includes the squirly brackets around the block

[oli-obk]

since we know there must be at least one statement, you can instead do b.stmts.last().unwrap().span.shrink_to_hi(), this way you'll get a span right after the last statement.

[surechen]

since we know there must be at least one statement, you can instead do b.stmts.last().unwrap().span.shrink_to_hi(), this way you'll get a span right after the last statement.

Thank you. I tried this way. But when the last stmt is a macro call like println!();, it's span will point to std
like: D:\source\rust\rust\library\std\src\macros.rs:85:6: 85:6 (#8).

[oli-obk]

ah indeed. hmm... Maybe we should change the value of Block::span to not include the brackets around it. We can already obtain the "with brackets" span by taking the span of the outer expression. That's a bit more involved though, as you'd probably need to touch the parser.

I don't think the - BytePos(1) will work in case the block itself is created by a macro or other expansion. Please add some tests for async functions, which should show us any problematic behaviour. And try to generate an entire function with a macro and see how your suggestion behaves on that.

[surechen]

ah indeed. hmm... Maybe we should change the value of Block::span to not include the brackets around it. We can already obtain the "with brackets" span by taking the span of the outer expression. That's a bit more involved though, as you'd probably need to touch the parser.

I don't think the - BytePos(1) will work in case the block itself is created by a macro or other expansion. Please add some tests for async functions, which should show us any problematic behaviour. And try to generate an entire function with a macro and see how your suggestion behaves on that.

Hello, thank you very much.
I tried async and declaring macros and it handles the case of declaring macros correctly.
For async functions, now the current logic will skip it. I added the corresponding conditions and check a async function like:

async fn test2() {
    let mut file = File::create("foo.txt")?;
    println!();
}

The fix result will be :

async fn test2() -> Result<(), Box<dyn std::error::Error>> {
    Ok(())
}

The span seems point to the body correct. But after fixed the expr let mut file = File::create("foo.txt")?; and println!(); in the body will be overwriten.

Because the current logic does not handle the closure situation, and I have not thought of an effective way to repair the closure, I think we can skip it in this PR.

[oli-obk]

Thanks for looking into the details!

@bors r+ rollup

[bors]

📌 Commit 0b3fec9 has been approved by oli-obk

It is now in the queue for this repository.