"http://%s", bindAddress, skywalkingPort, client, cmd.OutOrStdout(), browser)
		},
	}

	return cmd
}

// portForward first tries to forward localhost:remotePort to podName:remotePort, falls back to dynamic local port
func portForward(podName, namespace, flavor, urlFormat, localAddress string, remotePort int,
	client kube.CLIClient, writer io.Writer, browser bool,
) error {
	// port preference:

	}

	// Attempt to create the folder in case it doesn't exist
	if err := os.MkdirAll(filepath.Dir(path), 0o750); err != nil {
		// If we cannot create it, just warn here - we will fail later if there is a real error
		log.Warnf("Failed to create directory for %v: %v", path, err)
	}

	var err error
	listener, err := net.Listen("unix", path)
	if err != nil {

		{"-t", iptablesconstants.MANGLE, "-D", iptablesconstants.OUTPUT, "-j", ChainInpodOutput},
		{"-t", iptablesconstants.NAT, "-D", iptablesconstants.OUTPUT, "-j", ChainInpodOutput},
	}

	// these sometimes fail due to "Device or resource busy"
	optionalDeleteCmds := [][]string{
		// flush-then-delete our created chains
		{"-t", iptablesconstants.MANGLE, "-F", ChainInpodPrerouting},

	}

	// Create a kube client
	client := kube.NewFakeClient(objects...)

	mockRedir := &mockInterceptRuleMgr{}
	err = doAddRun(args, conf, client.Kube(), mockRedir)
	if err == nil {
		t.Fatal("expected to fail, but did not!")
	}

	return mockRedir
}

func testDoAddRun(t *testing.T, stdinData, nsName string, objects ...runtime.Object) *mockInterceptRuleMgr {
	args := buildCmdArgs(stdinData, testPodName, nsName)

				runtime.UnlockOSThread()
			}
		}()

		return toRun()
	}

	var wg sync.WaitGroup
	wg.Add(1)

	// Start the callback in a new green thread so that if we later fail
	// to switch the namespace back to the original one, we can safely
	// leave the thread locked to die without a risk of the current thread
	// left lingering with incorrect namespace.
	var innerError error
	go func() {

			// netns == ""; at this point netns should have been added via the initial snapshot,
			// or via the cni plugin. If it happens to get here before the cni plugin somehow,
			// then we will just fail to add the pod to the mesh, and it will be retried later when cni plugin adds it.

			// We need a pod IP - if the pod was added via the CNI plugin, that plugin told us the IPs

	deserializer := codecFactory.UniversalDeserializer()

	fail := admitv1.Fail
	fakeClient := kube.NewFakeClient(&admitv1.ValidatingWebhookConfiguration{
		TypeMeta: metav1.TypeMeta{},
		ObjectMeta: metav1.ObjectMeta{
			Name: "istiod-default-validator",
		},
		Webhooks: []admitv1.ValidatingWebhook{
			{
				Name: "random",
			},
			{
				FailurePolicy: &fail,
				Name:          "validation.istio.io",
			},
		},

	// once the pod restarts (in CrashLoopBackoff), which can take some time.
	// We don't want to constantly try to apply the iptables rules, which is unneeded and will fail.
	// Instead, we track which UIDs we repaired and skip them if already repaired.
	//
	// An alternative would be to write something to the Pod (status, annotation, etc).

			expectedString: "Error: expecting pod name",
			wantException:  true, // "istioctl experimental inspect pod" should fail
		},
		{ // case 3 unknown pod
			args:           strings.Split("po not-a-pod", " "),
			expectedString: "pods \"not-a-pod\" not found",
			wantException:  true, // "istioctl experimental describe pod not-a-pod" should fail
		},
		{ // case 8 unknown service
			args:           strings.Split("service not-a-service", " "),

			expectedString: "NAME     CLUSTER     CDS     LDS     EDS     RDS     ECDS     ISTIOD",
			revision:       "canary",
		},
		{ // case 7: supplying type that doesn't select pods should fail
			args:          strings.Split("serviceaccount/sleep", " "),
			wantException: true,
		},
	}