test(e2e): improve test reliability and fix silent failures (#730)

- Add 30s timeout to agent URL wait loop to prevent infinite hangs
- Propagate errors in SetupGPG instead of silently returning nil
- Use dedicated http.ServeMux in ServeAgent to avoid global state conflicts
- Add explicit timeouts/polling to all Eventually assertions missing them
- Replace bare defer cleanup with ginkgo.DeferCleanup for timeout safety
- Remove 3s sleep in SSH port forwarding test
- Consolidate duplicate setupDockerProvider into framework.SetupDockerProvider
- Delete commented-out GPG forwarding test code
- Fix "nachine" typos and apt-get install spacing in error messages

Author: skevetter

e2e/e2e_suite_test.go

@@ -33,8 +33,14 @@ func TestRunE2ETests(t *testing.T) {
 	if runtime.GOOS != "linux" {
 		go framework.ServeAgent()
 
-		// wait for http server to be up and running
+		// wait for http server to be up and running (max 30s)
+		deadline := time.After(30 * time.Second)
 		for {
+			select {
+			case <-deadline:
+				t.Fatal("timeout waiting for DEVPOD_AGENT_URL to be set after 30s")
+			default:
+			}
 			time.Sleep(time.Second)
 			if os.Getenv("DEVPOD_AGENT_URL") != "" {
 				break

e2e/framework/command.go

@@ -286,22 +286,22 @@ func (f *Framework) DevPodProviderUpdate(ctx context.Context, args ...string) er
 	return nil
 }
 
-func (f *Framework) DevPodMachineCreate(args []string) error {
+func (f *Framework) DevPodMachineCreate(ctx context.Context, args []string) error {
 	baseArgs := []string{"machine", "create"}
 	baseArgs = append(baseArgs, args...)
-	err := f.ExecCommand(context.Background(), false, false, "", baseArgs)
+	err := f.ExecCommand(ctx, false, false, "", baseArgs)
 	if err != nil {
-		return fmt.Errorf("devpod nachine create failed: %s", err.Error())
+		return fmt.Errorf("devpod machine create failed: %s", err.Error())
 	}
 	return nil
 }
 
-func (f *Framework) DevPodMachineDelete(args []string) error {
+func (f *Framework) DevPodMachineDelete(ctx context.Context, args []string) error {
 	baseArgs := []string{"machine", "delete"}
 	baseArgs = append(baseArgs, args...)
-	err := f.ExecCommand(context.Background(), false, false, "", baseArgs)
+	err := f.ExecCommand(ctx, false, false, "", baseArgs)
 	if err != nil {
-		return fmt.Errorf("devpod nachine delete failed: %s", err.Error())
+		return fmt.Errorf("devpod machine delete failed: %s", err.Error())
 	}
 	return nil
 }
@@ -325,31 +325,31 @@ func (f *Framework) DevPodWorkspaceDelete(
 
 func (f *Framework) SetupGPG(tmpDir string) error {
 	if _, err := exec.LookPath("gpg"); err != nil {
-		err := exec.Command("sudo", "apt-get", " install", "gnupg2", "-y").Run()
-		if err != nil {
-			return nil
+		if installErr := exec.Command("sudo", "apt-get", "install", "gnupg2", "-y").
+			Run(); installErr != nil {
+			return fmt.Errorf("gpg not found and failed to install gnupg2: %w", installErr)
 		}
 	}
 
-	err := exec.Command("gpg", "--import", filepath.Join(tmpDir, "gpg-public.key")).Run()
-	if err != nil {
-		return nil
+	// #nosec G204 -- gpg with fixed arguments for test GPG key setup
+	if err := exec.Command("gpg", "--import", filepath.Join(tmpDir, "gpg-public.key")).
+		Run(); err != nil {
+		return fmt.Errorf("failed to import gpg public key: %w", err)
 	}
 
-	err = exec.Command("gpg", "--import", filepath.Join(tmpDir, "gpg-private.key")).Run()
-	if err != nil {
-		return nil
+	// #nosec G204 -- gpg with fixed arguments for test GPG key setup
+	if err := exec.Command("gpg", "--import", filepath.Join(tmpDir, "gpg-private.key")).
+		Run(); err != nil {
+		return fmt.Errorf("failed to import gpg private key: %w", err)
 	}
 
-	err = exec.Command("gpgconf", "--kill", "gpg-agent").Run()
-	if err != nil {
-		return nil
+	if err := exec.Command("gpgconf", "--kill", "gpg-agent").Run(); err != nil {
+		return fmt.Errorf("failed to kill gpg-agent: %w", err)
 	}
 
-	err = exec.Command("gpg-agent", "--homedir", "$HOME/.gnupg", "--use-standard-socket", "--daemon").
-		Run()
-	if err != nil {
-		return nil
+	if err := exec.Command("gpg-agent", "--homedir", "$HOME/.gnupg", "--use-standard-socket", "--daemon").
+		Run(); err != nil {
+		return fmt.Errorf("failed to start gpg-agent: %w", err)
 	}
 
 	return exec.Command("gpg", "-k").Run()
@@ -456,3 +456,19 @@ func (f *Framework) DevPodIDEList(ctx context.Context, extraArgs ...string) (str
 	baseArgs := []string{"ide", "list"}
 	return f.ExecCommandOutput(ctx, append(baseArgs, extraArgs...))
 }
+
+// SetupDockerProvider creates a new framework, removes any existing docker provider,
+// adds a fresh one with the given docker path, and sets it as the active provider.
+func SetupDockerProvider(binDir, dockerPath string) (*Framework, error) {
+	f := NewDefaultFramework(binDir)
+	_ = f.DevPodProviderDelete(context.Background(), "docker")
+	if err := f.DevPodProviderAdd(
+		context.Background(),
+		"docker",
+		"-o",
+		"DOCKER_PATH="+dockerPath,
+	); err != nil {
+		return nil, fmt.Errorf("failed to add docker provider: %w", err)
+	}
+	return f, f.DevPodProviderUse(context.Background(), "docker")
+}

e2e/framework/server_utils.go

@@ -24,8 +24,9 @@ func ServeAgent() {
 	// Create a file server handler for the specified directory
 	fileServer := http.FileServer(http.Dir(dir))
 
-	// Register the file server handler to serve files under the /files route
-	http.Handle("/files/", http.StripPrefix("/files", fileServer))
+	// Use a dedicated ServeMux to avoid conflicts with http.DefaultServeMux
+	mux := http.NewServeMux()
+	mux.Handle("/files/", http.StripPrefix("/files", fileServer))
 
 	ip := getIP()
 
@@ -40,10 +41,10 @@ func ServeAgent() {
 		log.Fatal(err)
 	}
 
-	// Start the HTTP server on port 8080
 	log.Printf("Server started on %s", addr)
 
-	err = http.Serve(listener, nil)
+	// #nosec G114 -- test-only agent file server, no timeout needed
+	err = http.Serve(listener, mux)
 	if err != nil {
 		log.Fatal(err)
 	}