feat: add evaluateAll benchmark with tinybench

- Add tinybench devDependency and 'bench' script
- Create 20-flag config exercising all operator types, segments,
  variants, and disabled flags (3 true, 6 false, 11 variant results)
- Benchmark runner for evaluateAll: ~418K ops/s (~2.6μs per call)

Run with: npm run bench (from packages/core)

Co-Authored-By: Mastra Code (anthropic/claude-opus-4-6) <noreply@mastra.ai>

Author: mfrachet

packages/core/package.json

@@ -20,6 +20,7 @@
     "test": "vitest",
     "coverage": "vitest run --coverage",
     "lint": "eslint .",
+    "bench": "tsx src/__benchmarks__/evaluateAll.bench.ts",
     "bundlesize": "bundlesize"
   },
   "keywords": [],
@@ -36,6 +37,7 @@
     "bundlesize": "^0.18.2",
     "eslint": "^9.29.0",
     "rollup": "^4.44.0",
+    "tinybench": "^6.0.0",
     "tslib": "^2.8.1",
     "tsx": "^4.20.3",
     "typescript": "^5.8.3",

packages/core/src/__benchmarks__/evaluateAll.bench.ts

@@ -0,0 +1,31 @@
+import { Bench } from "tinybench";
+import { createFlagEngine } from "../index";
+import { benchmarkFlags } from "./flagsConfig";
+
+const bench = new Bench({ warmupIterations: 100 });
+
+const engine = createFlagEngine(benchmarkFlags);
+
+// User context that exercises every operator type in the config
+const userCtx = engine.createUserContext({
+  __id: "bench-user-42",
+  plan: "premium",
+  country: "FR",
+  email: "bench@company.com",
+  age: 28,
+  status: "active",
+  totalOrders: 150,
+  lifetimeValue: 1200,
+  browser: "Chrome",
+  role: "admin",
+  accountAgeDays: 400,
+  onboardingComplete: true,
+  language: "fr",
+});
+
+bench.add("evaluateAll (20 flags)", () => {
+  userCtx.evaluateAll();
+});
+
+await bench.run();
+console.table(bench.table());