huskies/server/src/agents/pool/pipeline/completion/tests.rs

//! Tests for agent pipeline completion handling.
use super::super::super::AgentPool;
use super::*;
use crate::agents::{AgentEvent, AgentStatus, CompletionReport};
use std::path::PathBuf;
use std::process::Command;

fn init_git_repo(repo: &std::path::Path) {
    Command::new("git")
        .args(["init"])
        .current_dir(repo)
        .output()
        .unwrap();
    Command::new("git")
        .args(["config", "user.email", "test@test.com"])
        .current_dir(repo)
        .output()
        .unwrap();
    Command::new("git")
        .args(["config", "user.name", "Test"])
        .current_dir(repo)
        .output()
        .unwrap();
    Command::new("git")
        .args(["commit", "--allow-empty", "-m", "init"])
        .current_dir(repo)
        .output()
        .unwrap();
}

// ── report_completion tests ────────────────────────────────────

#[tokio::test]
async fn report_completion_rejects_nonexistent_agent() {
    let pool = AgentPool::new_test(3001);
    let result = pool.report_completion("no_story", "no_bot", "done").await;
    assert!(result.is_err());
    let msg = result.unwrap_err();
    assert!(msg.contains("No agent"), "unexpected: {msg}");
}

#[tokio::test]
async fn report_completion_rejects_non_running_agent() {
    let pool = AgentPool::new_test(3001);
    pool.inject_test_agent("s6", "bot", AgentStatus::Completed);

    let result = pool.report_completion("s6", "bot", "done").await;
    assert!(result.is_err());
    let msg = result.unwrap_err();
    assert!(
        msg.contains("not running"),
        "expected 'not running' in: {msg}"
    );
}

#[tokio::test]
async fn report_completion_rejects_dirty_worktree() {
    use std::fs;
    use tempfile::tempdir;

    let tmp = tempdir().unwrap();
    let repo = tmp.path();

    // Init a real git repo and make an initial commit
    Command::new("git")
        .args(["init"])
        .current_dir(repo)
        .output()
        .unwrap();
    Command::new("git")
        .args(["commit", "--allow-empty", "-m", "init"])
        .current_dir(repo)
        .output()
        .unwrap();

    // Write an uncommitted file
    fs::write(repo.join("dirty.txt"), "not committed").unwrap();

    let pool = AgentPool::new_test(3001);
    pool.inject_test_agent_with_path("s7", "bot", AgentStatus::Running, repo.to_path_buf());

    let result = pool.report_completion("s7", "bot", "done").await;
    assert!(result.is_err());
    let msg = result.unwrap_err();
    assert!(
        msg.contains("uncommitted"),
        "expected 'uncommitted' in: {msg}"
    );
}

// ── server-owned completion tests ───────────────────────────────────────────

#[tokio::test]
async fn server_owned_completion_skips_when_already_completed() {
    let pool = AgentPool::new_test(3001);
    let report = CompletionReport {
        summary: "Already done".to_string(),
        gates_passed: true,
        gate_output: String::new(),
    };
    pool.inject_test_agent_with_completion(
        "s10",
        "coder-1",
        AgentStatus::Completed,
        PathBuf::from("/tmp/nonexistent"),
        report,
    );

    // Subscribe before calling so we can check if Done event was emitted.
    let mut rx = pool.subscribe("s10", "coder-1").unwrap();

    run_server_owned_completion(
        &pool.agents,
        pool.port,
        "s10",
        "coder-1",
        Some("sess-1".to_string()),
        pool.watcher_tx.clone(),
    )
    .await;

    // Status should remain Completed (unchanged) — no gate re-run.
    let agents = pool.agents.lock().unwrap();
    let key = super::super::super::composite_key("s10", "coder-1");
    let agent = agents.get(&key).unwrap();
    assert_eq!(agent.status, AgentStatus::Completed);
    // Summary should still be the original, not overwritten.
    assert_eq!(agent.completion.as_ref().unwrap().summary, "Already done");
    drop(agents);

    // No Done event should have been emitted.
    assert!(
        rx.try_recv().is_err(),
        "should not emit Done when completion already exists"
    );
}

#[tokio::test]
async fn server_owned_completion_runs_gates_on_clean_worktree() {
    use tempfile::tempdir;

    let tmp = tempdir().unwrap();
    let repo = tmp.path();
    init_git_repo(repo);

    let pool = AgentPool::new_test(3001);
    pool.inject_test_agent_with_path("s11", "coder-1", AgentStatus::Running, repo.to_path_buf());

    let mut rx = pool.subscribe("s11", "coder-1").unwrap();

    run_server_owned_completion(
        &pool.agents,
        pool.port,
        "s11",
        "coder-1",
        Some("sess-2".to_string()),
        pool.watcher_tx.clone(),
    )
    .await;

    // Agent entry should be removed from the map after completion.
    let agents = pool.agents.lock().unwrap();
    let key = super::super::super::composite_key("s11", "coder-1");
    assert!(
        agents.get(&key).is_none(),
        "agent should be removed from map after completion"
    );
    drop(agents);

    // A Done event should have been emitted with the session_id.
    let event = rx.try_recv().expect("should emit Done event");
    match &event {
        AgentEvent::Done { session_id, .. } => {
            assert_eq!(*session_id, Some("sess-2".to_string()));
        }
        other => panic!("expected Done event, got: {other:?}"),
    }
}

#[tokio::test]
async fn server_owned_completion_fails_on_dirty_worktree() {
    use std::fs;
    use tempfile::tempdir;

    let tmp = tempdir().unwrap();
    let repo = tmp.path();
    init_git_repo(repo);
    // Create an uncommitted file.
    fs::write(repo.join("dirty.txt"), "not committed").unwrap();

    let pool = AgentPool::new_test(3001);
    pool.inject_test_agent_with_path("s12", "coder-1", AgentStatus::Running, repo.to_path_buf());

    let mut rx = pool.subscribe("s12", "coder-1").unwrap();

    run_server_owned_completion(
        &pool.agents,
        pool.port,
        "s12",
        "coder-1",
        None,
        pool.watcher_tx.clone(),
    )
    .await;

    // Agent entry should be removed from the map after completion (even on failure).
    let agents = pool.agents.lock().unwrap();
    let key = super::super::super::composite_key("s12", "coder-1");
    assert!(
        agents.get(&key).is_none(),
        "agent should be removed from map after failed completion"
    );
    drop(agents);

    // A Done event should have been emitted.
    let event = rx.try_recv().expect("should emit Done event");
    assert!(
        matches!(event, AgentEvent::Done { .. }),
        "expected Done event, got: {event:?}"
    );
}

#[tokio::test]
async fn server_owned_completion_nonexistent_agent_is_noop() {
    let pool = AgentPool::new_test(3001);
    // Should not panic or error — just silently return.
    run_server_owned_completion(
        &pool.agents,
        pool.port,
        "nonexistent",
        "bot",
        None,
        pool.watcher_tx.clone(),
    )
    .await;
}

/// Regression test for bug 445: a rate-limited mergemaster exits before
/// calling start_merge_agent_work.  run_server_owned_completion must be a
/// no-op for mergemaster agents — it must not run acceptance gates and must
/// not advance the story to 5_done/ even when a passing script/test exists.
///
/// Before the fix: run_server_owned_completion would call run_pipeline_advance
/// for the Mergemaster stage, which ran post-merge tests on master (they pass
/// because nothing changed), then called move_story_to_done — advancing the
/// story without any squash merge having occurred.
#[cfg(unix)]
#[tokio::test]
async fn server_owned_completion_is_noop_for_mergemaster() {
    use std::fs;
    use std::os::unix::fs::PermissionsExt;
    use tempfile::tempdir;

    let tmp = tempdir().unwrap();
    let root = tmp.path();
    init_git_repo(root);

    // Create a passing script/test so post-merge tests would succeed if
    // run_pipeline_advance were incorrectly called for this mergemaster.
    let script_dir = root.join("script");
    fs::create_dir_all(&script_dir).unwrap();
    let script_test = script_dir.join("test");
    fs::write(&script_test, "#!/usr/bin/env sh\nexit 0\n").unwrap();
    let mut perms = fs::metadata(&script_test).unwrap().permissions();
    perms.set_mode(0o755);
    fs::set_permissions(&script_test, perms).unwrap();

    // Story in 4_merge/ — must NOT be moved to 5_done/.
    let merge_dir = root.join(".huskies/work/4_merge");
    fs::create_dir_all(&merge_dir).unwrap();
    let story_path = merge_dir.join("99_story_merge445.md");
    fs::write(&story_path, "---\nname: Merge 445 Test\n---\n").unwrap();

    let pool = AgentPool::new_test(3001);
    pool.inject_test_agent_with_path(
        "99_story_merge445",
        "mergemaster",
        AgentStatus::Running,
        root.to_path_buf(),
    );

    run_server_owned_completion(
        &pool.agents,
        pool.port,
        "99_story_merge445",
        "mergemaster",
        None,
        pool.watcher_tx.clone(),
    )
    .await;

    // Wait briefly in case any background task fires.
    tokio::time::sleep(std::time::Duration::from_millis(150)).await;

    // Story must remain in 4_merge/ — not moved to 5_done/.
    let done_path = root.join(".huskies/work/5_done/99_story_merge445.md");
    assert!(
        !done_path.exists(),
        "Story must NOT be moved to 5_done/ when run_server_owned_completion \
         is (incorrectly) called for a mergemaster agent"
    );
    assert!(
        story_path.exists(),
        "Story must remain in 4_merge/ when mergemaster completion is a no-op"
    );

    // The agent entry should remain in the pool (lifecycle cleanup is the
    // caller's responsibility, not run_server_owned_completion's).
    let agents = pool.agents.lock().unwrap();
    let key = super::super::super::composite_key("99_story_merge445", "mergemaster");
    assert!(
        agents.get(&key).is_some(),
        "Agent must remain in pool — run_server_owned_completion is a no-op for mergemaster"
    );
}

/// Bug 645 + 651: when an agent crashes leaving dirty files but committed
/// work, server-owned completion should stash the dirty files during gates
/// and restore them afterward.  Uncommitted work is never junk.
#[tokio::test]
async fn server_owned_completion_preserves_dirty_worktree_with_committed_work() {
    use std::fs;
    use tempfile::tempdir;

    let tmp = tempdir().unwrap();
    let project_root = tmp.path().join("project");
    fs::create_dir_all(&project_root).unwrap();
    init_git_repo(&project_root);

    // Create a worktree on a feature branch with committed code.
    let wt_path = tmp.path().join("wt");
    Command::new("git")
        .args([
            "worktree",
            "add",
            &wt_path.to_string_lossy(),
            "-b",
            "feature/story-645_test",
        ])
        .current_dir(&project_root)
        .output()
        .unwrap();

    // Commit a valid file.
    fs::write(wt_path.join("work.txt"), "done").unwrap();
    Command::new("git")
        .args(["add", "."])
        .current_dir(&wt_path)
        .output()
        .unwrap();
    Command::new("git")
        .args(["commit", "-m", "coder: add work"])
        .current_dir(&wt_path)
        .output()
        .unwrap();

    // Now simulate crash leaving dirty files.
    fs::write(wt_path.join("dirty.txt"), "crash residue").unwrap();

    let pool = AgentPool::new_test(3001);
    pool.inject_test_agent_with_path("645_test", "coder-1", AgentStatus::Running, wt_path.clone());

    let mut rx = pool.subscribe("645_test", "coder-1").unwrap();

    run_server_owned_completion(
        &pool.agents,
        pool.port,
        "645_test",
        "coder-1",
        Some("sess-645".to_string()),
        pool.watcher_tx.clone(),
    )
    .await;

    // Bug 651: The dirty file must be PRESERVED — uncommitted work is never junk.
    assert!(
        wt_path.join("dirty.txt").exists(),
        "dirty file should be preserved after server-owned completion (bug 651)"
    );
    assert_eq!(
        fs::read_to_string(wt_path.join("dirty.txt")).unwrap(),
        "crash residue",
        "dirty file contents should be unchanged"
    );

    // A Done event should have been emitted (completion ran, didn't fail
    // on dirty worktree).
    let event = rx.try_recv().expect("should emit Done event");
    assert!(
        matches!(event, AgentEvent::Done { .. }),
        "expected Done event, got: {event:?}"
    );
}

/// AC3 (bug 651): simulate an agent killed by the watchdog with a
/// substantive uncommitted diff.  After the orchestrator's full
/// post-termination handling (gates, completion, advance check),
/// `git status --short` must still show the same modified files.
#[tokio::test]
async fn watchdog_kill_preserves_uncommitted_diff() {
    use std::fs;
    use tempfile::tempdir;

    let tmp = tempdir().unwrap();
    let project_root = tmp.path().join("project");
    fs::create_dir_all(&project_root).unwrap();
    init_git_repo(&project_root);

    // Create a worktree on a feature branch with committed code.
    let wt_path = tmp.path().join("wt");
    Command::new("git")
        .args([
            "worktree",
            "add",
            &wt_path.to_string_lossy(),
            "-b",
            "feature/story-651_watchdog",
        ])
        .current_dir(&project_root)
        .output()
        .unwrap();

    // Commit some work.
    fs::write(wt_path.join("committed.txt"), "committed work").unwrap();
    Command::new("git")
        .args(["add", "."])
        .current_dir(&wt_path)
        .output()
        .unwrap();
    Command::new("git")
        .args(["commit", "-m", "coder: add committed work"])
        .current_dir(&wt_path)
        .output()
        .unwrap();

    // Simulate substantive uncommitted diff left by watchdog kill.
    fs::write(wt_path.join("in_progress.rs"), "fn wip() {}").unwrap();
    fs::write(wt_path.join("committed.txt"), "modified after commit").unwrap();

    // Snapshot git status before completion.
    let status_before = Command::new("git")
        .args(["status", "--short"])
        .current_dir(&wt_path)
        .output()
        .unwrap();
    let status_before = String::from_utf8_lossy(&status_before.stdout)
        .trim()
        .to_string();
    assert!(
        !status_before.is_empty(),
        "pre-condition: worktree should be dirty"
    );

    let pool = AgentPool::new_test(3001);
    pool.inject_test_agent_with_path(
        "651_watchdog",
        "coder-1",
        AgentStatus::Running,
        wt_path.clone(),
    );

    run_server_owned_completion(
        &pool.agents,
        pool.port,
        "651_watchdog",
        "coder-1",
        Some("sess-651".to_string()),
        pool.watcher_tx.clone(),
    )
    .await;

    // After full post-termination handling, git status must be unchanged.
    let status_after = Command::new("git")
        .args(["status", "--short"])
        .current_dir(&wt_path)
        .output()
        .unwrap();
    let status_after = String::from_utf8_lossy(&status_after.stdout)
        .trim()
        .to_string();

    assert_eq!(
        status_before, status_after,
        "Bug 651: uncommitted diff must survive post-termination handling.\n\
         Before: {status_before}\nAfter: {status_after}"
    );

    // Verify file contents are intact.
    assert_eq!(
        fs::read_to_string(wt_path.join("in_progress.rs")).unwrap(),
        "fn wip() {}",
    );
    assert_eq!(
        fs::read_to_string(wt_path.join("committed.txt")).unwrap(),
        "modified after commit",
    );
}

/// Story 910 regression: a coder that exits with zero commits on the feature
/// branch must NOT be promoted to Merge.  The server-owned completion path
/// detects `git rev-list master..HEAD == 0`, records `gates_passed=false`,
/// and the pipeline advance retries (or blocks at the cap) instead of
/// advancing the story.
#[tokio::test]
async fn zero_commit_coder_exit_stays_in_coding_not_promoted_to_merge() {
    use std::fs;
    use std::process::Command;

    let tmp = tempfile::tempdir().unwrap();
    let project_root = tmp.path().join("project");
    fs::create_dir_all(&project_root).unwrap();

    // Init a git repo with an initial master commit.
    Command::new("git")
        .args(["init"])
        .current_dir(&project_root)
        .output()
        .unwrap();
    Command::new("git")
        .args(["config", "user.email", "test@test.com"])
        .current_dir(&project_root)
        .output()
        .unwrap();
    Command::new("git")
        .args(["config", "user.name", "Test"])
        .current_dir(&project_root)
        .output()
        .unwrap();
    Command::new("git")
        .args(["commit", "--allow-empty", "-m", "init"])
        .current_dir(&project_root)
        .output()
        .unwrap();

    // Create a feature-branch worktree with ZERO commits ahead of master.
    let wt_path = tmp.path().join("wt");
    Command::new("git")
        .args([
            "worktree",
            "add",
            &wt_path.to_string_lossy(),
            "-b",
            "feature/story-9910_zero_exit",
        ])
        .current_dir(&project_root)
        .output()
        .unwrap();

    // Set up the story with max_retries=1 so it blocks on the first failure.
    crate::crdt_state::init_for_test();
    crate::db::ensure_content_store();
    crate::db::write_item_with_content(
        "9910_zero_exit",
        "2_current",
        "---\nname: Zero Exit Test\n---\n",
        crate::db::ItemMeta::from_yaml("---\nname: Zero Exit Test\n---\n"),
    );

    fs::create_dir_all(project_root.join(".huskies")).unwrap();
    fs::write(
        project_root.join(".huskies/project.toml"),
        "max_retries = 1\n\n[[agent]]\nname = \"coder-1\"\nstage = \"coder\"\n",
    )
    .unwrap();

    let pool = AgentPool::new_test(3001);
    pool.inject_test_agent_with_root_and_path(
        "9910_zero_exit",
        "coder-1",
        AgentStatus::Running,
        project_root.clone(),
        wt_path.clone(),
    );

    let mut rx = pool.watcher_tx.subscribe();

    run_server_owned_completion(
        &pool.agents,
        pool.port,
        "9910_zero_exit",
        "coder-1",
        None,
        pool.watcher_tx.clone(),
    )
    .await;

    // The pipeline advance spawns asynchronously — poll with a timeout.
    let mut got_blocked = false;
    let deadline = tokio::time::Instant::now() + std::time::Duration::from_secs(5);
    while tokio::time::Instant::now() < deadline {
        while let Ok(evt) = rx.try_recv() {
            if let crate::io::watcher::WatcherEvent::StoryBlocked { story_id, .. } = &evt
                && story_id == "9910_zero_exit"
            {
                got_blocked = true;
            }
        }
        if got_blocked {
            break;
        }
        tokio::time::sleep(std::time::Duration::from_millis(50)).await;
    }

    assert!(
        got_blocked,
        "Story 910 regression: a zero-commit coder exit must block/retry \
         the story rather than advancing it to Merge"
    );

    // The story must NOT be in 4_merge.
    if let Ok(Some(item)) = crate::pipeline_state::read_typed("9910_zero_exit") {
        assert_ne!(
            item.stage.dir_name(),
            "4_merge",
            "Story must NOT be in Merge after a zero-commit coder exit"
        );
    }
}

/// AC4 (bug 651 regression for 645): when an agent crashes with committed
/// work AND uncommitted noise, the auto-advance still picks up the
/// committed work.  The committed-state check is authoritative; the
/// uncommitted state is just preserved on disk for the next agent.
#[tokio::test]
async fn committed_work_advances_despite_uncommitted_noise() {
    use std::fs;
    use tempfile::tempdir;

    let tmp = tempdir().unwrap();
    let project_root = tmp.path().join("project");
    fs::create_dir_all(&project_root).unwrap();
    init_git_repo(&project_root);

    // Create a minimal Cargo project so cargo check works.
    fs::write(
        project_root.join("Cargo.toml"),
        "[package]\nname = \"test_proj\"\nversion = \"0.1.0\"\nedition = \"2021\"\n",
    )
    .unwrap();
    fs::create_dir_all(project_root.join("src")).unwrap();
    fs::write(project_root.join("src/lib.rs"), "// empty\n").unwrap();
    Command::new("git")
        .args(["add", "."])
        .current_dir(&project_root)
        .output()
        .unwrap();
    Command::new("git")
        .args(["commit", "-m", "add cargo project"])
        .current_dir(&project_root)
        .output()
        .unwrap();

    // Create a worktree on a feature branch.
    let wt_path = tmp.path().join("wt");
    Command::new("git")
        .args([
            "worktree",
            "add",
            &wt_path.to_string_lossy(),
            "-b",
            "feature/story-651_regression",
        ])
        .current_dir(&project_root)
        .output()
        .unwrap();

    // Commit valid code on the feature branch.
    fs::write(wt_path.join("src/lib.rs"), "pub fn good() {}\n").unwrap();
    Command::new("git")
        .args(["add", "."])
        .current_dir(&wt_path)
        .output()
        .unwrap();
    Command::new("git")
        .args(["commit", "-m", "add good fn"])
        .current_dir(&wt_path)
        .output()
        .unwrap();

    // Simulate crash leaving uncommitted noise (broken syntax in tracked file).
    fs::write(wt_path.join("src/lib.rs"), "THIS IS BROKEN SYNTAX!!!\n").unwrap();
    fs::write(wt_path.join("crash_junk.tmp"), "untracked noise").unwrap();

    // The "work survived" check should detect committed work and pass cargo check
    // despite the dirty worktree, WITHOUT destroying the dirty files.
    assert!(
        crate::agents::gates::worktree_has_committed_work(&wt_path),
        "committed work should be detected"
    );
    assert!(
        crate::agents::gates::cargo_check_in_worktree(&wt_path),
        "cargo check should pass on committed code (stash/pop, not reset)"
    );

    // Dirty files must still exist after the check.
    assert_eq!(
        fs::read_to_string(wt_path.join("src/lib.rs")).unwrap(),
        "THIS IS BROKEN SYNTAX!!!\n",
        "uncommitted noise in tracked file must be preserved"
    );
    assert!(
        wt_path.join("crash_junk.tmp").exists(),
        "untracked noise file must be preserved"
    );
}