test: migrate unit tests from bun to jest

Standardize test runner to jest across CI and local development:
- Update CI workflow to use jest for coverage reporting
- Update Makefile test targets to use jest
- Update tokenizer worker to use node:assert (jest-compatible)
- Remove async/await from tokenizer.loadTokenizerModules
  (Promise.allSettled already returns Promise)
- Change test model to gpt-5 for future-proofing

This ensures consistent test execution between local and CI
environments and prepares for better coverage reporting.

Author: ThomasK33

.github/workflows/ci.yml

@@ -75,7 +75,7 @@ jobs:
       - uses: ./.github/actions/setup-cmux
 
       - name: Run tests with coverage
-        run: bun test --coverage --coverage-reporter=lcov ${{ github.event.inputs.test_filter || 'src' }}
+        run: bun x jest --coverage --coverageReporters=lcov ${{ github.event.inputs.test_filter || 'src' }}
 
       - name: Upload coverage to Codecov
         uses: codecov/codecov-action@v5

Makefile

@@ -204,11 +204,11 @@ check-deadcode: node_modules/.installed ## Check for potential dead code (manual
 
 ## Testing
 test-integration: node_modules/.installed ## Run all tests (unit + integration)
-	@bun test src
+	@bun x jest src
 	@TEST_INTEGRATION=1 bun x jest tests
 
-test-unit: node_modules/.installed ## Run unit tests
-	@bun test src
+test-unit: node_modules/.installed build-main ## Run unit tests (requires build-main for worker compilation)
+	@bun x jest src
 
 test: test-unit ## Alias for test-unit
 

src/utils/main/tokenizer.test.ts

@@ -1,4 +1,4 @@
-import { beforeEach, describe, expect, jest, test } from "@jest/globals";
+import { beforeAll, beforeEach, describe, expect, jest, test } from "@jest/globals";
 
 import {
   __resetTokenizerForTests,
@@ -10,22 +10,24 @@ import {
 
 jest.setTimeout(20000);
 
+const model = "openai:gpt-5";
+beforeAll(async () => {
+  // warm up the worker_thread and tokenizer before running tests
+  await expect(loadTokenizerModules([model])).resolves.toHaveLength(1);
+});
+
 beforeEach(() => {
   __resetTokenizerForTests();
 });
 
 describe("tokenizer", () => {
-  const model = "openai:gpt-4-turbo";
-
   test("loadTokenizerModules warms known encodings", async () => {
-    await expect(loadTokenizerModules([model])).resolves.toBeUndefined();
     const tokenizer = await getTokenizerForModel(model);
     expect(typeof tokenizer.encoding).toBe("string");
     expect(tokenizer.encoding.length).toBeGreaterThan(0);
   });
 
   test("countTokens returns stable values", async () => {
-    await loadTokenizerModules([model]);
     const text = "cmux-tokenizer-smoke-test";
     const first = await countTokens(model, text);
     const second = await countTokens(model, text);
@@ -34,7 +36,6 @@ describe("tokenizer", () => {
   });
 
   test("countTokensBatch matches individual calls", async () => {
-    await loadTokenizerModules([model]);
     const texts = ["alpha", "beta", "gamma"];
     const batch = await countTokensBatch(model, texts);
     expect(batch).toHaveLength(texts.length);

src/utils/main/tokenizer.ts

@@ -56,7 +56,7 @@ function normalizeModelKey(modelName: string): ModelName {
   return normalized as ModelName;
 }
 
-async function resolveEncoding(modelName: ModelName): Promise<string> {
+function resolveEncoding(modelName: ModelName): Promise<string> {
   let promise = encodingPromises.get(modelName);
   if (!promise) {
     promise = tokenizerPool
@@ -117,17 +117,13 @@ async function countTokensInternal(modelName: ModelName, text: string): Promise<
   return pending;
 }
 
-export async function loadTokenizerModules(
+export function loadTokenizerModules(
   modelsToWarm: string[] = Array.from(DEFAULT_WARM_MODELS)
-): Promise<void> {
-  await Promise.allSettled(
-    modelsToWarm.map(async (modelString) => {
-      try {
-        const modelName = normalizeModelKey(modelString);
-        await resolveEncoding(modelName);
-      } catch (error) {
-        console.warn(`[tokenizer] Failed to warm encoding for '${modelString}':`, error);
-      }
+): Promise<Array<PromiseSettledResult<string>>> {
+  return Promise.allSettled(
+    modelsToWarm.map((modelString) => {
+      const modelName = normalizeModelKey(modelString);
+      return resolveEncoding(modelName);
     })
   );
 }

src/utils/main/tokenizer.worker.ts

@@ -1,4 +1,4 @@
-import assert from "@/utils/assert";
+import assert from "node:assert";
 import { Tokenizer, models } from "ai-tokenizer";
 import type { ModelName } from "ai-tokenizer";
 import * as encoding from "ai-tokenizer/encoding";

tests/ipcMain/forkWorkspace.test.ts

@@ -17,7 +17,6 @@ import {
 import { detectDefaultTrunkBranch } from "../../src/git";
 import { HistoryService } from "../../src/services/historyService";
 import { createCmuxMessage } from "../../src/types/message";
-import * as path from "path";
 
 // Skip all tests if TEST_INTEGRATION is not set
 const describeIntegration = shouldRunIntegrationTests() ? describe : describe.skip;

tests/ipcMain/openai-web-search.test.ts

@@ -1,9 +1,4 @@
-import {
-  setupWorkspace,
-  shouldRunIntegrationTests,
-  validateApiKeys,
-  type TestEnvironment,
-} from "./setup";
+import { setupWorkspace, shouldRunIntegrationTests, validateApiKeys } from "./setup";
 import { sendMessageWithModel, createEventCollector, assertStreamSuccess } from "./helpers";
 
 // Skip all tests if TEST_INTEGRATION is not set

tests/ipcMain/resumeStream.test.ts

@@ -1,9 +1,4 @@
-import {
-  setupWorkspace,
-  shouldRunIntegrationTests,
-  validateApiKeys,
-  type TestEnvironment,
-} from "./setup";
+import { setupWorkspace, shouldRunIntegrationTests, validateApiKeys } from "./setup";
 import { sendMessageWithModel, createEventCollector, waitFor } from "./helpers";
 import { IPC_CHANNELS } from "../../src/constants/ipc-constants";
 import type { Result } from "../../src/types/result";

tests/ipcMain/sendMessage.test.ts

@@ -5,7 +5,6 @@ import {
   setupWorkspaceWithoutProvider,
   shouldRunIntegrationTests,
   validateApiKeys,
-  type TestEnvironment,
 } from "./setup";
 import {
   sendMessageWithModel,