hive-forge: die by SIGPIPE like every other CLI in a pipeline

Rust's runtime sets SIGPIPE to SIG_IGN at startup, so writing to a pipe
whose reader has gone away returns EPIPE and println! panics. `hive-forge
<verb> | head` printed a panic and exited 101 where cat, grep and every
other pipeline member exit quietly.

101 is not cosmetic: to a caller running `set -o pipefail` it is a real
failure, so a shell script that pipes our output stops on a condition
that is not an error.

Restore SIG_DFL first thing in main, before any output. Measured against
the same command: piped to head, 141 (killed by the signal) with empty
stderr; unpiped, 0; the pre-change binary, 101 with a panic.

Closes #3972

Cargo.lock

@@ -1740,6 +1740,7 @@ dependencies = [
  "anyhow",
  "clap",
  "forgejo-api",
+ "libc",
  "reqwest",
  "serde",
  "serde_json",

hive-forge/Cargo.toml

@@ -38,6 +38,9 @@ reqwest = { workspace = true, features = [
 ] }
 serde = { workspace = true }
 serde_json = { workspace = true }
+# Only for restoring the default SIGPIPE disposition at startup — see
+# `restore_sigpipe` in src/main.rs. std has no safe API for it.
+libc = { workspace = true }
 
 [lints]
 workspace = true

hive-forge/src/main.rs

@@ -207,6 +207,7 @@ enum Verb {
 /// (alternate `Display`), which keeps the full `context` chain inline
 /// rather than dropping it the way plain `Display` would.
 fn main() -> ExitCode {
+    restore_sigpipe();
     if let Err(e) = run() {
         eprintln!("hive-forge: FAILED: {e:#}");
         return ExitCode::FAILURE;
@@ -214,6 +215,26 @@ fn main() -> ExitCode {
     ExitCode::SUCCESS
 }
 
+/// Give `SIGPIPE` back its default disposition before any output happens.
+///
+/// Rust's runtime sets `SIGPIPE` to `SIG_IGN` at startup, so writing to a
+/// pipe whose reader has gone away returns `EPIPE` and `println!` panics.
+/// `hive-forge <verb> | head` therefore dies with a panic message and exit
+/// **101** where every other CLI in a pipeline exits quietly — and 101 is
+/// a real failure to a caller running `set -o pipefail`, so a shell script
+/// that pipes our output stops on a condition that is not an error.
+///
+/// Restoring `SIG_DFL` makes the process die by signal instead, which is
+/// what the surrounding shell already knows how to interpret.
+fn restore_sigpipe() {
+    // SAFETY: called first thing in `main`, before any thread exists and
+    // before any output — so no concurrent handler mutation is possible,
+    // and no write can race the change.
+    unsafe {
+        libc::signal(libc::SIGPIPE, libc::SIG_DFL);
+    }
+}
+
 fn run() -> Result<()> {
     let cli = Cli::parse();
     let verb = cli.verb;