proc_metrics::sample_tree supports Linux and macOS (this is

Error message

proc_metrics::sample_tree supports Linux and macOS (this is a {} build)

What it means

proc_metrics::sample_tree samples the process plus its children (children via /proc on Linux and proc_listchildpids on macOS, skipping children whose rusage call fails) and assembles a TreeSample; every platform other than Linux and macOS compiles a stub that bails with the build's OS name instead of fabricating readings. Same intentional-honesty design as sample_self, one level up the process tree.

Source

Thrown at src/openhuman/platform/proc_metrics/tree.rs:202

                pid,
                name: macos::proc_name(pid),
                rss_kib,
            }),
            None => {
                eprintln!(
                    "[proc_metrics] tree: skipping child pid={pid}: proc_pid_rusage unavailable (exited or permission denied)"
                );
            }
        }
    }

    Ok(TreeSample::assemble(self_sample, children))
}

/// Unsupported-platform stub — fails loudly rather than fabricating a reading.
#[cfg(not(any(target_os = "linux", target_os = "macos")))]
pub fn sample_tree() -> anyhow::Result<TreeSample> {
    anyhow::bail!(
        "proc_metrics::sample_tree supports Linux and macOS (this is a {} build)",
        std::env::consts::OS
    )
}

#[cfg(target_os = "macos")]
mod macos {
    use std::collections::HashSet;
    use std::mem::{size_of, MaybeUninit};

    /// Direct children of `ppid` via `proc_listchildpids`. Empty on any error.
    ///
    /// Note the Darwin ABI quirk: `proc_listchildpids` returns the **number of
    /// pids** written (not a byte count, unlike `proc_listpids`), and fills the
    /// buffer with that many `pid_t`. It also doesn't reliably support the NULL
    /// size-probe form, so we allocate a real buffer up front and grow it if the
    /// kernel filled it completely (the list may have been truncated).
    fn child_pids(ppid: i32) -> Vec<i32> {

View on GitHub (pinned to 7491200858)

Solutions

Gate the caller: skip tree metrics on unsupported platforms.
Contribute a Windows implementation of the child-enumeration path upstream if needed.
Design dashboards to tolerate an absent tree sample rather than erroring.

Example fix

// before
let tree = proc_metrics::sample_tree()?;

// after — degrade to "no tree sample" on unsupported platforms
let tree = match proc_metrics::sample_tree() {
    Ok(t) => Some(t),
    Err(e) if e.to_string().contains("supports Linux and macOS") => None,
    Err(e) => return Err(e),
};

Defensive patterns

Strategy: fallback

Validate before calling

fn proc_metrics_supported() -> bool {
    matches!(std::env::consts::OS, "linux" | "macos")
}

if proc_metrics_supported() {
    let tree = proc_metrics::sample_tree()?;
    // ...
}

Type guard

fn can_sample_tree() -> bool {
    matches!(std::env::consts::OS, "linux" | "macos")
}

Try / catch

Match the 'supports Linux and macOS' message and degrade to a self-only or absent tree sample; distinguish it from per-child skip warnings (proc_pid_rusage unavailable), which the Linux/macOS path already tolerates.

Prevention

cfg-gate tree-metric wiring per target.
Design health panels to render without child-process data.
Contribute platform implementations rather than stubbing silently.

When it happens

Trigger: Calling proc_metrics::sample_tree() in a windows (or otherwise non-linux/macos) build — e.g. a health monitor that includes child-process CPU/memory on the Windows desktop target.

Common situations: Windows desktop builds with tree-aware health panels; cross-platform test matrices; ports to BSD/Android.

Understand the failure class

Background: "unsupported platform" / "not supported on this platform" errors: what they mean and how to fix them — this error's family across 47 libraries.

Related errors

AI-assisted analysis of tinyhumansai/openhuman@7491200858 (2026-08-17). Data as JSON: /api/errors/872e0913d8922b39. Report an issue: GitHub.