rust-lang/cargo · error · AlreadyPrintedError

see above

Error message

see above

What it means

This is NOT a real error message — it is the `AlreadyPrintedError` sentinel returned by `GlobalDiagnosticStats::ok` (src/diagnostics/mod.rs:108) when `error_count > 0`. Cargo has already printed the substantive diagnostics to the shell; the top-level `CargoResult` still needs an `Err` variant, so it wraps the literal string `"see above"` to tell the user the real cause was already displayed.

Source

Thrown at src/diagnostics/mod.rs:110

    pub fn scope(&mut self) -> ScopedDiagnosticStats<'_> {
        ScopedDiagnosticStats {
            warning_count: 0,
            error_count: 0,
            global: self,
        }
    }

    pub fn error_count(&self) -> usize {
        self.error_count
    }

    pub fn lint_warning_count(&self) -> usize {
        self.lint_warning_count
    }

    pub fn ok(&self) -> CargoResult<PassOutput> {
        if 0 < self.error_count {
            Err(crate::Error::new(crate::AlreadyPrintedError::new(
                anyhow::format_err!("see above"),
            )))
        } else {
            Ok(PassOutput {
                lint_warning_count: self.lint_warning_count,
            })
        }
    }
}

pub struct ScopedDiagnosticStats<'g> {
    warning_count: usize,
    error_count: usize,
    global: &'g mut GlobalDiagnosticStats,
}

impl ScopedDiagnosticStats<'_> {
    pub fn warning_count(&self) -> usize {

View on GitHub (pinned to 0e07a15537)

Solutions

1. Scroll up or re-run and capture full output — the actual error was printed before this sentinel.
2. Pipe output to a file (`cargo build 2>&1 | tee build.log`) and search for the first error.
3. Use `--message-format=json` and parse all messages, not just the final exit error.

Example fix

# capture full output instead of relying on the last line
cargo build 2>&1 | tee build.log
# then inspect the first diagnostic, not the 'see above' sentinel

// Cargo commands return CargoResult; on Err, do not treat 'see above' as the cause —
// capture full stderr and surface the first real diagnostic.
let output = std::process::Command::new("cargo").args(&["build"]).output()?;
if !output.status.success() {
    let stderr = String::from_utf8_lossy(&output.stderr);
    let first_err = stderr.lines().find(|l| l.contains("error["));
    // report first_err, not the trailing 'see above'
}

When it happens

Trigger: Any cargo command that records errors via the diagnostics system and then calls `gds.ok()` to convert to a final result. The 'see above' error appears as the process exit reason whenever one or more earlier errors were already reported.

Common situations: Build/test failures where multiple compile errors were shown; the terminal scrolled past the real error; JSON message format showing `"see above"` as the final error because prior structured messages held the detail; scripts capturing only the last error line.

Related errors

• not implemented
• argument for --color must be auto, always, or never, but fou
• no executable for `{}` found in PATH
• subcommand is required, add a subcommand to the command alia
• alias {} has unresolvable recursive definition: {} -> {}

AI-assisted analysis of rust-lang/cargo@0e07a15537 (2026-08-06). Data as JSON: /data/errors/404fe4d57c47239f.json. Report an issue: GitHub.