test/e2e/graceful-shutdown.spec.ts (3)

29-48: Consider closing proxyServer in afterEach to avoid leaks on test failure.

proxyServer is started inside each it and only closed at the end of the test body. If an assertion fails (or closeServer is never reached), the listening server leaks across tests, which can cascade into port conflicts or hung Vitest workers. Moving the cleanup into afterEach makes the suite more robust.

♻️ Proposed cleanup

   afterEach(async () => {
+    if (proxyServer?.listening) {
+      await closeServer(proxyServer);
+    }
     await mockTargetServer.stop();
   });

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@test/e2e/graceful-shutdown.spec.ts` around lines 29 - 48, The tests can leak
the listening proxyServer if a test fails before calling closeServer; update the
cleanup to close any started proxyServer in afterEach by invoking
closeServer(proxyServer) when proxyServer is defined (and await it), and swallow
or rethrow errors as appropriate; locate the proxyServer variable and the
existing afterEach/mockTargetServer.stop sequence and add the proxyServer
shutdown there (and reset proxyServer to undefined after closing) so each test
always attempts to close the server even on failure.

---

61-61: Minor: prefer 127.0.0.1 for consistency.

Other E2E suites in this PR (e.g., test/e2e/hono.spec.ts) target http://127.0.0.1:${port}. Using localhost here can cause flakiness on dual-stack hosts where DNS resolves to ::1 while the server listens on IPv4 only. Aligning on 127.0.0.1 is safer and consistent with the rest of the suite.

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@test/e2e/graceful-shutdown.spec.ts` at line 61, Replace the hardcoded
"localhost" host in the E2E request with the IPv4 loopback to avoid DNS/IPv6
flakiness: update the call that constructs the URL using SERVER_PORT (the
request(`http://localhost:${SERVER_PORT}`).get('/api').expect(200) invocation)
to use "127.0.0.1" instead of "localhost" so the test targets
http://127.0.0.1:${SERVER_PORT} and matches the other E2E suites.

---

79-79: Redundant mock setup.

mockTargetServer.forGet('/api').thenReply(200, 'OK') is already configured in beforeEach (line 42). This duplicate registration adds another handler for the same path; it's harmless here but obscures intent. Safe to remove.

♻️ Proposed fix

-    await mockTargetServer.forGet('/api').thenReply(200, 'OK');
-
     // Multiple requests should not cause multiple close subscriptions

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@test/e2e/graceful-shutdown.spec.ts` at line 79, Remove the redundant mock
registration: delete the duplicate call to
mockTargetServer.forGet('/api').thenReply(200, 'OK') in the test body since the
same route is already registered in the beforeEach setup; look for the
mockTargetServer.forGet invocation in graceful-shutdown.spec.ts and keep only
the beforeEach-registered handler to avoid adding duplicate handlers for '/api'.