push updated state to clients over SSE (ISS-129)

CHANGELOG.md

@@ -1,5 +1,10 @@
 # Changelog
 
+## v0.57.0 — 2026-09-22 19:19
+- **Push updated state to clients over SSE (ISS-129).** Clients no longer rely on pure polling for live updates: new `GET /api/events` streams full game-state snapshots as Server-Sent Events — the current snapshot immediately on connect, then one after every sim tick (~1 s), so the UI updates push-driven even while idle. New module `lib/sse.ts` is a tiny in-process fan-out registry (`subscribe(fn)` → unsubscribe, `broadcast(text)` with per-subscriber try/catch) — the three process-wide singletons (this subscriber set, the sql.js DB promise (`getGameDb`), and the tick lock (`withLock`)) live on `globalThis`: Next.js can load a module more than once per process (instrumentation.ts and route handlers get separate module instances), so plain module-level state would give each layer its own registry/DB/lock, and background-tick pushes would never reach route-handler subscribers. The broadcast hook lives in `tickNow()` in `lib/sim/engine.ts`: after each locked tick on the singleton path it serializes via the SAME `getStateSnapshot()` that `GET /api/state` returns (no duplicated serialization) and pushes to all subscribers; failures are logged, never fatal. Action mutations reach other clients within one background tick (~1 s); the acting client still gets its immediate post-action snapshot from the POST response as before. The SSE route (`app/api/events/route.ts`) sends `Content-Type: text/event-stream`, `Cache-Control: no-cache`, `Connection: keep-alive` and `X-Accel-Buffering: no` (prod sits behind Traefik), plus a `: ping` heartbeat comment every 15 s so idle proxies don't drop the connection; snapshot-first-then-subscribe ordering guarantees events are strictly ordered by sim clock, and stream cancel unsubscribes + clears the heartbeat; broadcast pushes are wrapped in `data: <json>\n\n` SSE frames at the route (matching the initial snapshot). Client (`lib/useGameState.ts`): one `EventSource('/api/events')` per page — onmessage parses the full snapshot into the shared state (same shape as today); polling stays as an automatic fallback: if no event arrives within 5 s of connect, or on a stream error, the old poll loop starts (immediate catch-up fetch + interval), and it stops again when fresh events flow. EventSource's native reconnect is relied on — no custom reconnect framework. `/api/state` behavior is unchanged (still ticks-then-snapshots; polling remains fully functional).
+- **API.** OpenAPI spec gains `GET /api/events` (200 → `text/event-stream`) in `lib/openapi.ts`, so the machine-readable spec at `/api/openapi` stays accurate.
+- **Tests.** Unit: new `lib/sse.test.ts` — subscribe/broadcast/unsubscribe semantics, a throwing subscriber doesn't break others, idempotent unsubscribe; new `app/api/events/route.test.ts` (throwaway-DB pattern from the state route test) — SSE headers + first chunk parses as a snapshot with `simState.clockMs`. E2E: new `e2e/events.spec.ts` (+1) — node-side fetch of `/api/events` against the app under test asserts content-type `text/event-stream`, the first event parses as a valid snapshot (`simState.clockMs`), and within ~5 s at least one MORE event arrives with an advanced clock (background tick pushes while idle since v0.56.0 — no scale hook); stream closed in `finally`. The exact-endpoint-list assertions in `lib/openapi.test.ts` and `e2e/api.spec.ts` move twelve → thirteen endpoints (`/api/events` added).
+
 ## v0.56.0 — 2026-09-22 18:28
 - **Idle fleet slows the world to real time instead of freezing it (ISS-128).** A fully grounded fleet no longer freezes the sim clock — `tickOnce` in `lib/sim/engine.ts` now advances by `min(elapsed, MAX_TICK_ELAPSED_MS) × IDLE_WORLD_TIME_SCALE` when nothing is flying or in repair, where new constant `IDLE_WORLD_TIME_SCALE = WORLD_TIME_SCALE / 24` (= exactly 1, i.e. real time) lives next to `WORLD_TIME_SCALE` in `lib/worldClock.ts`. Packages keep spawning/expiring and the market stays fresh while idle; an active fleet keeps its full scale including any FLIGHT_TEST_HOOKS override (grounded + hook scale 1000 → advances at ×1, not ×1000), and `lastRealTickMs` still always tracks real time so resuming activity never fast-forwards the idle duration. The client mirror (`advanceSimClock` in `lib/useSimClock.ts`) extrapolates at the same idle rate instead of freezing, so polls re-anchor without a visible jump.
 - **Tests.** Unit: `lib/sim/engine.test.ts` — the grounded-fleet tests now pin real-time advance clamped to one idle-rate tick (no freeze, no fast-forward), resume-without-catch-up after takeoff, and that a raised timeScale is ignored while grounded; new describe "idle slowdown" pins `IDLE_WORLD_TIME_SCALE = WORLD_TIME_SCALE / 24 = 1`. `lib/useSimClock.test.ts` — the frozen-while-idle test now asserts real-time advance at the idle rate. E2E: `e2e/sim.spec.ts`'s freeze test becomes a bounded real-time delta (advances > 0, ≥ half the wall span clamped to one tick, ≤ wall span + one clamp, clearly below ×24); `e2e/state.spec.ts`'s restart test replaces byte-identical clock equality with derived bounds — never backwards (not re-seeded) and ≤ `MAX_TICK_ELAPSED_MS` of clamped downtime + the live wall span; `e2e/right-pane.spec.ts`'s penalty preview no longer races a frozen clock — it verifies the idle rate, then allows the formula's slope over one UI poll interval (~1 s at ×1) instead of an exact string match. The `activateWorld`/repair wrappers in actions/panes/sim specs are unchanged (a repairing plane still means full scale); only their "freezes" comments were updated.