Unexpected dead_code warning when returning Result type from main

[zschreur]

I tried this code:

#[derive(Debug)]
enum MyError {
    IOErr(std::io::Error),
}

fn main() -> Result<(), MyError> {
    let io_err = std::io::Error::new(std::io::ErrorKind::Other, "Some IO Error");
    Err(MyError::IOErr(io_err))
}

I expect this to behave like so:

$ rustc ./main.rs
$ ./main
Error: IOErr(Custom { kind: Other, error: "Some IO Error" })

What I get is:

$ rustc ./main.rs
warning: field `0` is never read
 --> ./main.rs:3:11
  |
3 |     IOErr(std::io::Error),
  |     ----- ^^^^^^^^^^^^^^
  |     |
  |     field in this variant
  |
  = note: `#[warn(dead_code)]` on by default
help: consider changing the field to be of unit type to suppress this warning while preserving the field numbering, or remove the field
  |
3 |     IOErr(()),
  |           ~~

warning: 1 warning emitted

$ ./main
Error: IOErr(Custom { kind: Other, error: "Some IO Error" })

Meta

rustc --version --verbose:

rustc 1.77.1 (7cf61ebde 2024-03-27)
binary: rustc
commit-hash: 7cf61ebde7b22796c69757901dd346d0fe70bd97
commit-date: 2024-03-27
host: x86_64-apple-darwin
release: 1.77.1
LLVM version: 17.0.6

Backtrace

<backtrace>

[long-long-float]

@rustbot claim

[krtab]

This is the same diagnostic issue as #123068 to which I am assigned I believe.

@long-long-float I haven't started working on it yet, so you can take both if you want.

[long-long-float]

@krtab Thank you for informing me. I want to work on the issue #123068 in one PR or divied two PRs.

[long-long-float]

As discussed in #88900, this behavior is intentional. However there is a bug not to show an explanation about this behavior. I fixed the bug at #124460 and the message will be shown.

warning: field `0` is never read
 --> ../rust/test.rs:3:11
  |
3 |     IOErr(std::io::Error),
  |     ----- ^^^^^^^^^^^^^^
  |     |
  |     field in this variant
  |
  = note: `MyError` has a derived impl for the trait `Debug`, but this is intentionally ignored during dead code analysis
  = note: `#[warn(dead_code)]` on by default

[jieyouxu]

Triage: long-long-float added a remark (thank you!), but the behavior here is intentional as per Ignore derived Clone and Debug implementations during dead code analysis #85200, which was FCP'd by T-lang in #85200 (comment). This is also a duplicate of #88900.

In particular, there is a remark in the diagnostics now after long-long-float's contribution:

note: `Error` has a derived impl for the trait `Debug`, but this is intentionally ignored during dead code analysis

Hence, I am going to close this as a duplicate of #88900.

Thank you for the report and long-long-float for the diagnostics improvements.