settled-reach/server/src/main.rs

// The Settled Reach - Simulation Server
// Entry point for standalone simulation binary
//
// Supports --test-mode for automated testing:
//   --test-mode       Enable test mode (fixed seed, LISTENING signal, quieter logs)
//   --port <PORT>     Bind to specific port (0 = OS-assigned). Overrides positional addr.
//   --seed <SEED>     RNG seed (default: 0, test-mode default: 42)
//   --dump-schedule   Print bevy_ecs schedule graph and exit (no TCP required)

use bevy_app::prelude::*;
use tracing_subscriber::{layer::SubscriberExt, util::SubscriberInitExt};

use settled_reach_server::bridge::tcp::TcpBridge;
use settled_reach_server::bridge::{BridgePlugin, BridgeResource, HandshakeState, ServerRunning};
use settled_reach_server::simulation::SimulationPlugin;

fn main() {
    let args: Vec<String> = std::env::args().collect();
    let test_mode = args.iter().any(|a| a == "--test-mode");
    let dump_schedule = args.iter().any(|a| a == "--dump-schedule");

    let port_flag = args
        .iter()
        .position(|a| a == "--port")
        .and_then(|i| args.get(i + 1))
        .and_then(|s| s.parse::<u16>().ok());

    let seed_flag = args
        .iter()
        .position(|a| a == "--seed")
        .and_then(|i| args.get(i + 1))
        .and_then(|s| s.parse::<u64>().ok());

    // Tracing: quieter in test mode, always to stderr so stdout stays clean
    // for the LISTENING:{port} handshake signal.
    // CI=true → JSON format for structured log ingestion.
    // RUST_LOG_FORMAT=json → same effect for local debugging.
    let default_filter = if test_mode {
        "settled_reach_server=warn"
    } else {
        "settled_reach_server=debug"
    };
    let env_filter = tracing_subscriber::EnvFilter::try_from_default_env()
        .unwrap_or_else(|_| default_filter.into());
    let use_json =
        std::env::var("CI").is_ok() || std::env::var("RUST_LOG_FORMAT").as_deref() == Ok("json");
    if use_json {
        tracing_subscriber::registry()
            .with(env_filter)
            .with(
                tracing_subscriber::fmt::layer()
                    .json()
                    .with_writer(std::io::stderr),
            )
            .init();
    } else {
        tracing_subscriber::registry()
            .with(env_filter)
            .with(tracing_subscriber::fmt::layer().with_writer(std::io::stderr))
            .init();
    }

    // Resolve bind address.
    // --port flag overrides everything (most common in test mode).
    // Otherwise: positional arg > SR_ADDR env > default.
    let addr = if let Some(port) = port_flag {
        format!("127.0.0.1:{}", port)
    } else {
        // Find positional address arg, skipping flags and their values.
        let positional = {
            let mut skip_next = false;
            let mut found = None;
            for arg in args.iter().skip(1) {
                if skip_next {
                    skip_next = false;
                    continue;
                }
                if arg == "--port" || arg == "--seed" {
                    skip_next = true;
                    continue;
                }
                if arg.starts_with("--") {
                    continue;
                }
                found = Some(arg.clone());
                break;
            }
            found
        };
        positional
            .or_else(|| std::env::var("SR_ADDR").ok())
            .unwrap_or_else(|| "127.0.0.1:9876".to_string())
    };

    // --dump-schedule: print bevy_ecs schedule graph and exit (no TCP required).
    // Useful for PR artifacts and detecting unintended system reordering (#346).
    if dump_schedule {
        dump_schedule_graph();
        return;
    }

    // Bind FIRST, print port, THEN accept.
    // Critical for --port 0: the OS assigns a random port at bind time.
    // The LISTENING:{port} line is the handshake signal for the test client.
    let listener = std::net::TcpListener::bind(&addr).unwrap_or_else(|e| {
        eprintln!("Failed to bind {}: {}", addr, e);
        std::process::exit(1);
    });
    let actual_port = listener.local_addr().unwrap().port();

    // LISTENING signal to stdout. The test client parses this to discover the port.
    // All tracing goes to stderr (see .with_writer above), so stdout is clean.
    println!("LISTENING:{}", actual_port);
    {
        use std::io::Write;
        std::io::stdout().flush().ok();
    }

    tracing::info!("Waiting for client connection on port {}", actual_port);
    let bridge = TcpBridge::accept_on(listener).unwrap_or_else(|e| {
        tracing::error!("Failed to accept: {}", e);
        std::process::exit(1);
    });
    tracing::info!("Client connected, sending protocol handshake");

    // Protocol handshake: first framed message on the wire (#555).
    // Client reads this and validates protocol_version before sending any input.
    use settled_reach_server::bridge::SimBridge;
    bridge.send_handshake().unwrap_or_else(|e| {
        tracing::error!("Failed to send handshake: {}", e);
        std::process::exit(1);
    });

    // Read client's startup message containing world_seed (#175).
    // Client sends this immediately after validating the handshake.
    let startup = bridge.receive_startup().unwrap_or_else(|e| {
        tracing::error!("Failed to receive startup message: {}", e);
        std::process::exit(1);
    });

    tracing::info!("Handshake complete, initializing simulation");

    // RNG seed: --seed flag overrides client's world_seed (useful for testing).
    // Production: client sends world_seed via StartupMessage (#175).
    // Test mode default: 42 for deterministic replay.
    let seed = seed_flag.unwrap_or(if test_mode { 42 } else { startup.world_seed });

    let mut app = App::new();
    settled_reach_server::tick_phases::TickPhase::configure(&mut app);
    app.add_plugins(SimulationPlugin { seed });
    app.add_plugins(BridgePlugin);
    app.add_plugins(settled_reach_server::knowledge::KnowledgePlugin);
    app.add_plugins(settled_reach_server::npc::NpcPlugin);
    app.add_plugins(settled_reach_server::storyteller::StorytellerPlugin);
    app.add_plugins(settled_reach_server::settings::SettingsPlugin);
    app.add_plugins(settled_reach_server::bookmark::BookmarkPlugin::default());
    app.add_plugins(settled_reach_server::atlas::GenerationPlugin);

    // Initialize culture resolver (#679, D-128).
    // systems.db is shipped read-only alongside the binary.
    let systems_db_path = std::path::PathBuf::from("data/systems.db");
    match settled_reach_server::knowledge::CultureResolver::open(&systems_db_path) {
        Ok(resolver) => {
            tracing::info!("Culture resolver opened: {:?}", systems_db_path);
            app.insert_resource(settled_reach_server::knowledge::CultureResolverResource(
                resolver,
            ));
        }
        Err(e) => {
            tracing::warn!(
                "Culture resolver unavailable ({}). Culture lookups will not work.",
                e
            );
        }
    }

    // Mod-first body source resolver for the atlas layer proxy (#969, D-225).
    // terrain_reference is repo-root-relative; the repo root is systems.db's
    // 3rd ancestor (<repo>/server/data/systems.db).
    let world_root = systems_db_path
        .canonicalize()
        .ok()
        .and_then(|p| p.ancestors().nth(3).map(std::path::Path::to_path_buf))
        .unwrap_or_else(|| std::path::PathBuf::from(".."));
    match settled_reach_server::atlas::source_resolver::BodySourceResolver::open(
        &systems_db_path,
        vec![world_root.clone()],
    ) {
        Ok(resolver) => {
            tracing::info!("Body source resolver opened (root: {:?})", world_root);
            app.insert_resource(
                settled_reach_server::atlas::source_resolver::BodySourceResolverResource(resolver),
            );
        }
        Err(e) => tracing::warn!(
            "Body source resolver unavailable ({}). Atlas layer requests will error.",
            e
        ),
    }

    // Initialize SQLite settings store (#627).
    // Path: alongside save files in the server's working directory.
    let settings_path = std::path::PathBuf::from("settings.db");
    match settled_reach_server::settings::SettingsStore::open(&settings_path) {
        Ok(store) => {
            tracing::info!("Settings store opened: {:?}", settings_path);
            app.insert_resource(settled_reach_server::settings::SettingsStoreResource::new(
                store,
                "default".to_string(),
            ));
        }
        Err(e) => {
            tracing::error!(
                "Failed to open settings store: {}. Settings will not persist.",
                e
            );
        }
    }

    app.insert_resource(BridgeResource::new(bridge));
    app.insert_resource(HandshakeState::Complete);

    // SimulationPlugin { seed } already inserts SimRng with the correct seed
    // during plugin build. We re-insert here as a defensive override for one
    // specific ordering risk: any future plugin that registers *before*
    // SimulationPlugin in `App::add_plugins` order (e.g. a pre-simulation
    // observability plugin) and consumes SimRng at plugin build time would
    // see a stale resource that was never seeded from StartupMessage. This
    // `insert_resource` call happens AFTER all plugins have built, so it
    // always overwrites whatever SimRng is currently in the world with the
    // authoritative value from the StartupMessage. If you remove this line,
    // also audit every `app.add_plugins(...)` call in this file and in
    // `SimulationPlugin::build` for plugins that touch SimRng, and verify
    // none of them run before SimulationPlugin's seeding logic. #826 thread.
    app.insert_resource(settled_reach_server::simulation::rng::SimRng::new(seed));

    // Initialize empty line pool index (populated by generator pipeline in v0.2).
    app.insert_resource(
        settled_reach_server::simulation::line_pool::LinePoolIndexResource(
            settled_reach_server::simulation::line_pool::LinePoolIndex::default(),
        ),
    );

    // Gauntlet test world for --test-mode, proof room for normal mode.
    if test_mode {
        #[cfg(feature = "gauntlet")]
        settled_reach_server::test_world::setup_gauntlet(&mut app);
        #[cfg(not(feature = "gauntlet"))]
        {
            eprintln!("--test-mode requires the 'gauntlet' feature");
            std::process::exit(1);
        }
    } else {
        setup_proof_room(&mut app, seed);
    }

    tracing::info!(
        "Simulation initialized (seed={}, test_mode={})",
        seed,
        test_mode
    );

    // Game loop: run until client disconnects.
    // Targets ~20 ticks/sec (2 game-minutes/sec). The TCP bridge uses
    // non-blocking reads, so without throttling this loop would spin.
    // Remaining frame budget is available for NPC AI and pathfinding.
    //
    // Panic supervision (#85): each tick is wrapped in catch_unwind.
    // On panic, the server sends a structured SimError to the client
    // before shutting down, rather than an abrupt disconnect.
    let target_frame_time = std::time::Duration::from_millis(50);
    loop {
        let frame_start = std::time::Instant::now();

        // Wrap app.update() in catch_unwind to handle system panics (#85).
        // AssertUnwindSafe is required because App is not UnwindSafe.
        let tick_result = std::panic::catch_unwind(std::panic::AssertUnwindSafe(|| {
            app.update();
        }));

        match tick_result {
            Ok(()) => {}
            Err(panic_payload) => {
                // Extract panic message for error reporting
                let panic_msg = if let Some(s) = panic_payload.downcast_ref::<&str>() {
                    s.to_string()
                } else if let Some(s) = panic_payload.downcast_ref::<String>() {
                    s.clone()
                } else {
                    "unknown panic".to_string()
                };

                tracing::error!("Simulation panic caught: {}", panic_msg);

                // Attempt to send a final SimError snapshot to the client.
                // Best-effort: if the bridge is unavailable, we just log and exit.
                send_panic_error(&app, &panic_msg);

                tracing::error!("Server shutting down after panic");
                break;
            }
        }

        if !app.world().resource::<ServerRunning>().0 {
            break;
        }

        let elapsed = frame_start.elapsed();
        tracing::debug!(
            tick_ms = elapsed.as_millis(),
            budget_ms = target_frame_time.as_millis(),
            over_budget = elapsed > target_frame_time,
            "tick"
        );
        if elapsed < target_frame_time {
            std::thread::sleep(target_frame_time - elapsed);
        }
    }

    tracing::info!("Simulation server shutting down");
}

/// Best-effort: send a final SimError snapshot to the client on panic (#85).
///
/// Builds a minimal ObserverSnapshot with the panic error and sends it
/// through the bridge. If the bridge is unavailable or sending fails,
/// the error is logged but not fatal (we're already crashing).
fn send_panic_error(app: &App, panic_msg: &str) {
    use settled_reach_server::bridge::types::*;
    use settled_reach_server::simulation::time::{DayPhase, TickRate};

    let world = app.world();

    // Try to read current tick from SimulationTime
    let tick = world
        .get_resource::<settled_reach_server::simulation::time::SimulationTime>()
        .map(|t| t.tick)
        .unwrap_or(0);

    let bridge = match world.get_resource::<BridgeResource>() {
        Some(b) => b,
        None => {
            tracing::error!("Cannot send panic error: no BridgeResource");
            return;
        }
    };

    // Build a minimal snapshot carrying the panic error
    let snapshot = ObserverSnapshot {
        tick,
        game_time: GameTime {
            day: 0,
            time_of_day: 0,
            day_phase: DayPhase::Morning,
            tick_rate: TickRate::Paused,
        },
        player_facing: FacingDirection::North,
        player_stance: MovementStance::default(),
        player_inventory: vec![],
        entities: vec![],
        visible_tiles: vec![],
        nearby_interactions: vec![],
        current_monologue: None,
        pending_recognitions: vec![],
        dialogue_response: None,
        blocked_entities: vec![],
        scan_events: vec![],
        sound_events: vec![],
        follow_state: None,
        character_pressure: None,
        rng_seed: None,
        poi_list: vec![],
        examine_result: None,
        player_knowledge: None,
        save_result: None,
        triangle_crisis_events: vec![],
        state_hash: None,
        debug_response: None,
        current_ticker: None,
        settings_response: None,
        economy_snapshot: None,
        bookmark_catalog: None,
        sim_errors: vec![SimError {
            kind: SimErrorKind::Panic,
            message: format!("Simulation panic: {}", panic_msg),
            tick,
        }],
    };

    if let Err(e) = bridge.send_snapshot(&snapshot) {
        tracing::error!("Failed to send panic error to client: {}", e);
    } else {
        tracing::info!("Sent panic SimError to client at tick {}", tick);
    }
}

/// Print bevy_ecs schedule graph and exit.
/// Invoked by --dump-schedule CLI flag (#346).
///
/// Builds the full app with all plugins (no TCP bridge or world entities),
/// then prints each registered schedule and its system count to stdout.
/// Systems are counted from the registered (pre-initialization) graph, so
/// counts reflect what was registered by plugins.
///
/// CI integration: run on each PR via `make debug-schedule`, diff output
/// against a committed baseline to catch unintended system reordering.
fn dump_schedule_graph() {
    use bevy_ecs::schedule::Schedules;

    let mut app = App::new();
    settled_reach_server::tick_phases::TickPhase::configure(&mut app);
    app.add_plugins(SimulationPlugin { seed: 0 });
    app.add_plugins(BridgePlugin);
    app.add_plugins(settled_reach_server::knowledge::KnowledgePlugin);
    app.add_plugins(settled_reach_server::npc::NpcPlugin);
    app.add_plugins(settled_reach_server::storyteller::StorytellerPlugin);
    app.add_plugins(settled_reach_server::settings::SettingsPlugin);
    app.add_plugins(settled_reach_server::bookmark::BookmarkPlugin::default());
    app.add_plugins(settled_reach_server::atlas::GenerationPlugin);
    app.insert_resource(settled_reach_server::simulation::rng::SimRng::new(0));

    // Access Schedules resource directly — schedules are populated by plugins
    // via add_systems() before any tick runs. No app.update() needed here:
    // running a tick would require full world setup (WalkabilityMap, etc.) that
    // isn't needed for schedule inspection.
    let world = app.world();
    let schedules = world.resource::<Schedules>();

    println!("=== Schedule Graph (settled-reach-server) ===");
    let mut entries: Vec<String> = schedules
        .iter()
        .map(|(label, schedule)| format!("  {:?}  [{} systems]", label, schedule.systems_len()))
        .collect();
    entries.sort(); // deterministic output for baseline diffs
    let schedule_count = entries.len();
    for entry in &entries {
        println!("{}", entry);
    }
    println!("=== {} schedules total ===", schedule_count);
    println!();
    println!("Note: use RUST_LOG=trace with the live server for per-tick timing.");
    println!("      system names visible with `cargo build --features bevy/debug`.");
}

/// Proof room: 32x32 map, wall at (16,14), player at (16,16), 3 NPCs.
/// Extracted from the original inline setup for reuse by both test-mode and normal mode.
fn setup_proof_room(app: &mut App, world_seed: u64) {
    use settled_reach_server::knowledge::registry::EntityRegistry;
    use settled_reach_server::knowledge::KnowledgeGraph;
    use settled_reach_server::npc::relationships::{RelationshipEdge, RelationshipGraph};
    use settled_reach_server::npc::{
        Contentment, DailyRoutine, Npc, RelationshipKind, RoutineEntry, ToleranceThreshold, Want,
        WantKind,
    };
    use settled_reach_server::perception::cognitive_delay::CognitiveDelay;
    use settled_reach_server::perception::vision_cone::Facing;
    use settled_reach_server::simulation::interaction::{Interactable, NearbyInteractionBuffer};
    use settled_reach_server::simulation::listening::ListeningFocus;
    use settled_reach_server::simulation::monologue::{
        MonologueBuffer, MonologueState, SprintAnomalyQueue,
    };
    use settled_reach_server::simulation::movement::{
        PlayerCharacter, TilePosition, WalkabilityMap,
    };
    use settled_reach_server::simulation::path_follow::MovementSpeed;
    use settled_reach_server::simulation::rng::EntityRng;
    use settled_reach_server::simulation::stance::{MovementProfile, PlayerMoveCooldown};
    use settled_reach_server::simulation::time::DayPhase;

    app.insert_resource(WalkabilityMap::new(32, 32, 1));

    // Wall at (16,14) between player and NPC 1
    {
        let mut wm = app.world_mut().resource_mut::<WalkabilityMap>();
        wm.set_walkable(&TilePosition::new(16, 14, 0), false);
    }

    let mut registry = EntityRegistry::new(0);

    // Player at (16,16). Archetype-specific spawn logic was removed in Sprint 37
    // (D-032 purge); per-culture/per-role voice is reintroduced in Phase 6.
    let profile = MovementProfile::default();
    let player = app
        .world_mut()
        .spawn((
            PlayerCharacter,
            TilePosition::new(16, 16, 0),
            Facing::default(),
            KnowledgeGraph::new(),
            NearbyInteractionBuffer::default(),
            MonologueState::default(),
            MonologueBuffer::default(),
            SprintAnomalyQueue::default(),
            CognitiveDelay::default(),
            ListeningFocus::new(TilePosition::new(16, 16, 0)),
            profile,
            profile.initial_stance(),
            PlayerMoveCooldown::default(),
        ))
        .id();
    let player_sid = registry.register(player);
    app.world_mut()
        .entity_mut(player)
        .insert(EntityRng::from_seed_and_id(world_seed, player_sid.0));

    // NPC 1: Dock worker at (16,13) — behind wall, full routine
    let npc1 = app
        .world_mut()
        .spawn((
            Npc,
            Interactable,
            TilePosition::new(16, 13, 0),
            Want {
                primary: WantKind::Wealth,
                intensity: 6,
                description: "Wants a bigger share of docking fees".into(),
            },
            DailyRoutine {
                entries: vec![
                    RoutineEntry {
                        phase: DayPhase::Morning,
                        location: TilePosition::new(16, 13, 0),
                        activity: "Prep cargo bay".into(),
                    },
                    RoutineEntry {
                        phase: DayPhase::Afternoon,
                        location: TilePosition::new(20, 10, 0),
                        activity: "Unload freight".into(),
                    },
                    RoutineEntry {
                        phase: DayPhase::Evening,
                        location: TilePosition::new(10, 20, 0),
                        activity: "Drink at canteen".into(),
                    },
                    RoutineEntry {
                        phase: DayPhase::Night,
                        location: TilePosition::new(16, 13, 0),
                        activity: "Sleep in bunk".into(),
                    },
                ],
                description: "Dock worker shift pattern".into(),
            },
            Contentment { level: 20 },
            ToleranceThreshold {
                current_stress: 30,
                threshold: 70,
            },
            MovementSpeed::new(2),
        ))
        .id();
    let npc1_sid = registry.register(npc1);
    app.world_mut()
        .entity_mut(npc1)
        .insert(EntityRng::from_seed_and_id(world_seed, npc1_sid.0));

    // NPC 2: Field tech at (14,18) — visible to player, has routine
    let npc2 = app
        .world_mut()
        .spawn((
            Npc,
            Interactable,
            TilePosition::new(14, 18, 0),
            Want {
                primary: WantKind::Knowledge,
                intensity: 8,
                description: "Obsessed with pre-Collapse sensor arrays".into(),
            },
            DailyRoutine {
                entries: vec![
                    RoutineEntry {
                        phase: DayPhase::Morning,
                        location: TilePosition::new(14, 18, 0),
                        activity: "Calibrate instruments".into(),
                    },
                    RoutineEntry {
                        phase: DayPhase::Afternoon,
                        location: TilePosition::new(22, 22, 0),
                        activity: "Field survey".into(),
                    },
                ],
                description: "Field tech survey pattern".into(),
            },
            Contentment { level: 45 },
            ToleranceThreshold {
                current_stress: 10,
                threshold: 60,
            },
            MovementSpeed::default(),
        ))
        .id();
    let npc2_sid = registry.register(npc2);
    app.world_mut()
        .entity_mut(npc2)
        .insert(EntityRng::from_seed_and_id(world_seed, npc2_sid.0));

    // NPC 3: Guard at (18,14) — stationary, no routine
    let npc3 = app
        .world_mut()
        .spawn((
            Npc,
            Interactable,
            TilePosition::new(18, 14, 0),
            Want {
                primary: WantKind::Safety,
                intensity: 4,
                description: "Wants a quiet shift".into(),
            },
            Contentment { level: -5 },
            ToleranceThreshold {
                current_stress: 45,
                threshold: 55,
            },
        ))
        .id();
    let npc3_sid = registry.register(npc3);
    app.world_mut()
        .entity_mut(npc3)
        .insert(EntityRng::from_seed_and_id(world_seed, npc3_sid.0));

    // Populate RelationshipGraph with a few edges
    {
        let mut rel_graph = app.world_mut().resource_mut::<RelationshipGraph>();
        // Dock worker and guard are colleagues with moderate trust
        rel_graph.set_relationship(
            npc1_sid,
            npc3_sid,
            RelationshipEdge {
                kind: RelationshipKind::Colleague,
                trust: 3,
                history: vec![],
                last_interaction_tick: 0,
            },
        );
        // Guard distrusts the field tech (rival for resources)
        rel_graph.set_relationship(
            npc3_sid,
            npc2_sid,
            RelationshipEdge {
                kind: RelationshipKind::Rival,
                trust: -4,
                history: vec![],
                last_interaction_tick: 0,
            },
        );
    }

    app.insert_resource(registry);
}