astral-sh/ruff · error

Git clone failed: {}

Error message

Git clone failed: {}

What it means

The benchmark harness clones real-world projects on first use (blob-filtered, no checkout); if `git clone` exits nonzero this error is raised with git's stderr.

Source

Thrown at crates/ruff_benchmark/src/real_world_projects.rs:218

fn clone_repository(repo_url: &str, target_dir: &Path, commit: &str) -> Result<()> {
    // Create parent directory if it doesn't exist
    if let Some(parent) = target_dir.parent() {
        std::fs::create_dir_all(parent).context("Failed to create parent directory for clone")?;
    }

    // Clone with minimal depth and fetch only the specific commit
    let output = Command::new("git")
        .args([
            "clone",
            "--filter=blob:none", // Don't download large files initially
            "--no-checkout",      // Don't checkout files yet
            repo_url,
            target_dir.to_str().unwrap(),
        ])
        .output()
        .context("Failed to execute git clone command")?;

    anyhow::ensure!(
        output.status.success(),
        "Git clone failed: {}",
        String::from_utf8_lossy(&output.stderr)
    );

    // Fetch the specific commit
    let output = Command::new("git")
        .args(["fetch", "origin", commit])
        .current_dir(target_dir)
        .output()
        .context("Failed to execute git fetch command")?;

    anyhow::ensure!(
        output.status.success(),
        "Git fetch of commit {} failed: {}",
        commit,
        String::from_utf8_lossy(&output.stderr)
    );

View on GitHub (pinned to 672bb4edf0)

Solutions

1. Read the embedded git stderr to identify network vs repository vs disk failure
2. Verify the repo URL is reachable: `git ls-remote <repo_url> HEAD`
3. Free disk space under target/ if the clone ran out of room, then rerun
4. Retry after transient network failures; the clone is retried on the next run

Example fix

# before
 bench run fails with: Git clone failed: fatal: unable to access ...

# after
git ls-remote <repo_url> HEAD && cargo bench --bench <real-world-bench>

git ls-remote "$repo_url" HEAD >/dev/null 2>&1 || {
  echo "cannot reach $repo_url — check network/proxy" >&2
  exit 1
}
df -h target | tail -1 # ensure headroom before a multi-repo clone

for i in 1 2 3; do
  cargo bench --bench "$bench" && break
  echo "clone attempt $i failed; retrying" >&2
  rm -rf "target/benchmark_cache/$project"
  sleep 5
done

When it happens

Trigger: First `cargo bench` run of a real-world project when cloning fails: network unreachable, proxy blocking https git, repository moved or made private, or insufficient disk space.

Common situations: Air-gapped or proxied CI environments; disk pressure under target/; upstream repos relocated since the benchmark table was written.

Related errors

• Git fetch of commit {} failed: {}
• Git checkout of commit {} failed: {}
• Dependency installation failed: {}
• Failed to fetch playground ${id}: ${response.status}
• Failed to save playground: ${response.status}

AI-assisted analysis of astral-sh/ruff@672bb4edf0 (2026-08-16). Data as JSON: /api/errors/c6a948af1adc0166. Report an issue: GitHub.