I'm running an application with SignalR 2.2.0 on server side and signalr-java-client (self compiled, last GitHub version) on Android as client.
Currently, there are 4 clients connected to my hub. From time to time, it happens, that all 4 clients simultaneously receive the HTTP status 400 with the message "The connection id is in the incorrect format" (the clients were connected before). I analyzed this multiple times and am not able to find any information/pattern when or why this happens.
The connecten is secured via JWT, the token is definitely valid. When retrieving a new token, the connection is stopped and started again. Apart from this, it is very unlikely that the error is device-related, because the error is thrown at all 4 clients the same time.
I know, this error can occur when the client's Identity changes, but an Identity change for 4 clients the same time seems very unlikely to me.
This is the server-code used for authentication (Deepak asked).
The following method gets called in my Startup.cs:
public static void ConfigureOAuth(IAppBuilder app, string audienceID, string sharedSecret)
{
byte[] secret = TextEncodings.Base64Url.Decode(sharedSecret);
app.UseJwtBearerAuthentication(
new JwtBearerAuthenticationOptions
{
Provider = new MyOAuthBearerAuthenticationProvider(),
AuthenticationMode = AuthenticationMode.Active,
AllowedAudiences = new[] { audienceID },
IssuerSecurityTokenProviders = new IIssuerSecurityTokenProvider[]
{
new SymmetricKeyIssuerSecurityTokenProvider(Issuer, secret)
}
});
}
Here's the code of MyOAuthBearerAuthenticationProvider class:
class MyOAuthBearerAuthenticationProvider : OAuthBearerAuthenticationProvider
{
/// <summary>
/// Get's a JWT from querysting and puts it to context
/// </summary>
public override Task RequestToken(OAuthRequestTokenContext context)
{
if (context.Token == null)
{
string value = context.Request.Query.Get("auth_token");
if (!string.IsNullOrEmpty(value)) //token from queryString
{
context.Token = value;
}
}
return Task.FromResult<object>(null);
}
}
I have to retrieve the token from query string, because additionally to the java-client, a javascript client is used, which is not able to set headers.
Lastly, I secure my hub and some of it's methods with the Authorization attribute:
[Authorize(Roles = "MyExampleRole")]
This is the client-code for connection:
public boolean connect(String url, String token) {
if (connected) {
return true;
}
try {
this.hubConnection = new HubConnection(url, "auth_token=" + token, true, logger);
this.hubProxy = hubConnection.createHubProxy("MyHub");
this.hubProxy.subscribe(this.signalRMethodProvider);
this.hubConnection.stateChanged(stateChangedCallback);
SignalRFuture<Void> awaitConnection = this.hubConnection.start();
awaitConnection.get(10000, TimeUnit.MILLISECONDS);
return true;
}
catch (InterruptedException | TimeoutException | ExecutionException e) {
log.error("connect", e);
return false;
}
}
Does anybody have an Idea, how to fix this problem or where I may receive further information?
Thank you very much
-Lukas
seems fine...
possible alteration you can do is change
awaitConnection.get(10000, TimeUnit.MILLISECONDS);
to
awaitConnection.done(new Action<Void>() {
#Override
public void run(Void obj) throws Exception {
Log.d(TAG, "Hub Connected");
}
}).onError(new ErrorCallback() {
#Override
public void onError(Throwable error) {
error.printStackTrace();
Log.d(TAG, "SignalRServiceHub Cancelled");
}
}).onCancelled(new Runnable() {
#Override
public void run() {
Log.d(TAG, "SignalRServiceHub Cancelled");
}
});
Related
Is there any way to send a push notification from server when a user complete a task? For example: A todo app will notify on that date with push notification. I want to use firebase and firestore for storing user tokens.
Alarm manager can be a solution that I have found but I don't wanna use it.
Yes, you can use scheduler to send notification from server to your app:
You may follow my working code:
Emplement IJob:
public class SendNotificationViaFcm: IJob
{
public void Execute(IJobExecutionContext context)
{
bool isNotificationSent=false;
try
{
var taskToSendNotification = FirebaseCloudMessaging.SendMessage();
Task.WaitAll(taskToSendNotification);
isNotificationSent = taskToSendNotification.Result;
}
catch (Exception exception)
when (
exception is ObjectDisposedException || exception is ArgumentNullException ||
exception is AggregateException)
{
}
catch (Exception exception) when (exception is InvalidOperationException)
{
}
catch (Exception exception)
{
// ignored
}
}
}
Call FCM Api from your server:
public class FirebaseCloudMessaging
{
private static readonly Uri FcmUri = new Uri(
uriString: #"https://fcm.googleapis.com",
uriKind: UriKind.Absolute);
private const string FcmApiKey = "Your Legacy Server Key";
public static async Task<bool> SendMessage()
{
using (var httpClient = new HttpClient())
{
httpClient.BaseAddress = FcmUri;
httpClient.DefaultRequestHeaders.Clear();
httpClient.DefaultRequestHeaders.TryAddWithoutValidation("Authorization",
"key=" + FcmApiKey);
var response = await httpClient.PostAsJsonAsync(#"/fcm/send", new
{
to = "/topics/global",
priority = "high",
data = new
{
title = "Warning",
message = "Please start app to track movemoent!"
}
//to = "/topics/global",
//priority = "high",
//notification = new
//{
// title = "Warning!",
// body = "Please start app to track movemoent!"
//}
});
Debug.Write(response.Content.ReadAsStringAsync());
var ck = response.IsSuccessStatusCode;
return response.IsSuccessStatusCode;
}
}
}
Implement schedular for your time interval:
public class Scheduler
{
public static void Start()
{
try
{
IScheduler scheduler = StdSchedulerFactory.GetDefaultScheduler();
scheduler.Start();
// scheduler.Shutdown();
var sentCloudNotification = JobBuilder.Create<SendNotificationViaFcm>().Build();
var cloudNotificationTrigger = TriggerBuilder.Create().WithSimpleSchedule(x => x.WithIntervalInMinutes(1).RepeatForever()).Build();
scheduler.ScheduleJob(sentCloudNotification, cloudNotificationTrigger);
}
catch (SchedulerException exception)
{
Debug.Write(exception.Message);
}
catch (Exception exception)
{
Debug.Write(exception.Message);
}
}
}
Finally Run in from your Global.asax.cs
protected void Application_Start()
{
Scheduler.Start();
}
It sounds like you're looking for a tool that allows you to schedule transactional notifications. What sort of server technology are you using?
From a high level you could do something like this:
1) user adds a task in the Android application
2) android application sends request to server to save the task
3) you have some code that runs in some sort of on task save callback that schedules a block of code to run in the future using crontab, celery or something similar.
4) the block of code that runs in the future is an api call to twilio to send a push notification
relevant links: https://www.twilio.com, https://firebase.google.com/docs/cloud-messaging/, http://www.celeryproject.org/, https://en.wikipedia.org/wiki/Cron
I have an app connected with Azure backend. I created a login and some api calls 2 months ago. They worked fine until a few days ago and then it starts to fail "sometimes".
The login log onFailure says: Error while authenticating user
The callback log onFailure says: Error while processing request
And the cause of both says : stream was reset: PROTOCOL_ERROR
This post is to similar to this but didn't work.
Some code here:
LoginFragment.java
private void login(String email, String password){
loginProgressBar.setVisibility(View.VISIBLE);
try {
JsonObject params = new JsonObject();
params.addProperty("Username", email);
params.addProperty("Password", password);
ListenableFuture<MobileServiceUser> listenable = Client.logIn(getContext(), params);
Futures.addCallback(listenable, new FutureCallback<MobileServiceUser>() {
#Override
public void onSuccess(MobileServiceUser mobileServiceUser) {
loginProgressBar.setVisibility(View.GONE);
SharedPreferences settings = getActivity().getSharedPreferences(Client.MS_USER,0);
SharedPreferences.Editor editor = settings.edit();
Client.clientId = mobileServiceUser.getUserId();
Client.token = mobileServiceUser.getAuthenticationToken();
editor.putString(Client.MS_USER_ID, Client.clientId);
editor.putString(Client.MS_AUTH_TOKEN, Client.token);
editor.apply();
Client.getInstance(getContext()).setCurrentUser(mobileServiceUser);
Intent i = new Intent(getContext(), MainActivity.class);
startActivity(i);
}
#Override
public void onFailure(Throwable t) {
loginProgressBar.setVisibility(View.GONE);
Throwable t2 = t.getCause();
Throwable t3 = t2.getCause();
Log.e("LoginFail", t.getMessage());
Log.e("LoginFail", t2.getMessage());
if(t3 != null){
Log.e("LoginFail", t3.getMessage());
}
Toast.makeText(getContext(), getResources().getString(R.string.bad_login), Toast.LENGTH_LONG).show();
}
}, MoreExecutors.directExecutor());
} catch (MalformedURLException e) {
e.printStackTrace();
}
}
Client.java
public class Client {
public static final String MS_USER = "MS_USER";
public static final String MS_USER_ID = "MS_USER_ID";
public static final String MS_AUTH_TOKEN = "MS_AUTH_TOKEN";
public static String clientId;
public static String token;
private static MobileServiceClient instance = null;
public static MobileServiceClient getInstance(Context context) {
if (instance ==null){
try {
instance = new MobileServiceClient(Env.AZURE_URL, context);
instance.setAndroidHttpClientFactory(() -> {
OkHttpClient client = new OkHttpClient();
client.setReadTimeout(20, TimeUnit.SECONDS);
client.setWriteTimeout(20, TimeUnit.SECONDS);
return client;
});
} catch (MalformedURLException e) {
e.printStackTrace();
}
} else{
instance.setContext(context);
}
return instance;
}
public static ListenableFuture<MobileServiceUser> logIn(Context context, JsonObject parameters) throws MalformedURLException {
String deviceID = "gcm:" + Settings.Secure.getString(context.getContentResolver(), Settings.Secure.ANDROID_ID);
parameters.addProperty("device_id", deviceID);
parameters.addProperty("device_dateTime", Env.DATE_FORMAT.format(new Date()));
parameters.addProperty("device_timeZone", API.getTimezone());
parameters.addProperty("device_language", Env.LANGUAGE);
parameters.addProperty("app", Env.APP_NAME);
return getInstance(context).login("auth", parameters);
}
public static ListenableFuture<JsonElement> callApi(Context context, String apiName, JsonObject parameters, String httpMethod){
if(httpMethod.equals("POST")){
String deviceID = "gcm:" + Settings.Secure.getString(context.getContentResolver(), Settings.Secure.ANDROID_ID);
parameters.addProperty("user_id", Client.clientId);
parameters.addProperty("device_id", deviceID);
parameters.addProperty("device_dateTime", Env.DATE_FORMAT.format(new Date()));
parameters.addProperty("device_timeZone", API.getTimezone());
parameters.addProperty("device_language", Env.LANGUAGE);
parameters.addProperty("app", Env.APP_NAME);
parameters.addProperty("role", "Patient");
return getInstance(context).invokeApi(apiName, parameters, httpMethod, null);
} else {
return getInstance(context).invokeApi(apiName, null, httpMethod, null);
}
}
This is probably related to an issue in Azure App Service that is weirdly enough not reported on the public Azure status page.
The message that affected Azure client received was (quoted from the link above):
Starting at 02:00 UTC on 3 Apr 2018, you have been identified as a
customer using App Services who may have received connection failure
notifications when using Android apps with older HTTP clients or
desktop browsers using cross-site scripting calls. Engineers have
identified an issue with a recent deployment and are investigating
mitigation options. Customers experiencing this issue can
self-mitigate by updating the site config setting "http20Enabled" to
false via resources.azure.com. Instructions on how to update site
config can be found here:
https://azure.microsoft.com/en-us/blog/azure-resource-explorer-a-new-tool-to-discover-the-azure-api/
Go to https://resources.azure.com/
Make sure you are in Read/Write mode by clicking in the option to the
left of your name.
Find the affected site and browse to Config > Web:
https://resources.azure.com/subscriptions//resourceGroups//providers/Microsoft.Web/sites//config/web
Change the property: "http20Enabled": from true to false by clicking
in Edit properties, Update to “false” and then clicking PUT to save
change.
If you have tried these steps and are continuing to experience issues
with your App Service, please create a technical support ticket to
further troubleshoot: aka.ms/azsupt. This message will be closed in 7
days.
I am trying to use 'Authenticate using Cognito-Identity with Cognito user pool' in my Android application. My Cognito user pool authentication works well, when I run that separately and I had seen a JWTToken as well. When I run the the 'PubSub' sample application with Unauthenticated role, it worked as expected. When I integrate these two features in one application, the application threw following error.
W/System.err: MqttException (0) - java.io.IOException: Already connected
W/System.err: at org.eclipse.paho.client.mqttv3.internal.ExceptionHelper.createMqttException(ExceptionHelper.java:38)
W/System.err: at org.eclipse.paho.client.mqttv3.internal.ClientComms$ConnectBG.run(ClientComms.java:664)
W/System.err: at java.lang.Thread.run(Thread.java:761)
W/System.err: Caused by: java.io.IOException: Already connected
W/System.err: at java.io.PipedOutputStream.connect(PipedOutputStream.java:100)
W/System.err: at java.io.PipedInputStream.connect(PipedInputStream.java:195)
W/System.err: at org.eclipse.paho.client.mqttv3.internal.websocket.WebSocketReceiver.<init>(WebSocketReceiver.java:42)
W/System.err: at org.eclipse.paho.client.mqttv3.internal.websocket.WebSocketSecureNetworkModule.start(WebSocketSecureNetworkModule.java:78)
W/System.err: at org.eclipse.paho.client.mqttv3.internal.ClientComms$ConnectBG.run(ClientComms.java:650)
W/System.err: ... 1 more
I have been trying to resolve this issue since last Thursday and still stuck at the same place. Really No idea where should i check.!
I am adding my Authentication(Cognito user pool authentication) activity and Connect activity.
AmazonCognitoIdentityProviderClient identityProviderClient = new
AmazonCognitoIdentityProviderClient(new AnonymousAWSCredentials(), new ClientConfiguration());
identityProviderClient.setRegion(Region.getRegion(Regions.US_WEST_2));
CognitoUserPool userPool = new CognitoUserPool(getApplicationContext(), "us-west-2_ghtcc6ho9", "4t0mk45hNso69dp2j4jvel5ghm", "1jmq0lhhq721oif9k6nug31c29i760vihua8hvrgu5umfr2a1vd7", identityProviderClient);
cogUser = userPool.getUser();
authenticationHandler = new AuthenticationHandler() {
#Override
public void onSuccess(CognitoUserSession userSession, CognitoDevice newDevice) {
String ids = userSession.getIdToken().getJWTToken();
Log.d("MyToken","session id___"+userSession.getIdToken().getExpiration()+"___"+userSession.getIdToken().getIssuedAt());
Intent pubSub = new Intent(MainActivity.this, PubSubActivity.class);
pubSub.putExtra("token",""+ids);
startActivity(pubSub);
//MainActivity.this.finish();
}
#Override
public void getAuthenticationDetails(AuthenticationContinuation authenticationContinuation, String userId) {
Log.d("MyToken","getAuthenticationDetails");
AuthenticationDetails authenticationDetails = new AuthenticationDetails("shone", "172737", null);
authenticationContinuation.setAuthenticationDetails(authenticationDetails);
// Allow the sign-in to continue
authenticationContinuation.continueTask();
}
#Override
public void getMFACode(MultiFactorAuthenticationContinuation multiFactorAuthenticationContinuation) {
Log.d("MyToken","getMFACode");
multiFactorAuthenticationContinuation.continueTask();
}
#Override
public void authenticationChallenge(ChallengeContinuation continuation) {
Log.d("MyToken","authenticationChallenge"+continuation.getChallengeName());
newPasswordContinuation.continueTask();
}
#Override
public void onFailure(Exception exception) {
exception.printStackTrace();
Log.d("MyToken","onFailure");
}
};
cogUser.getSessionInBackground(authenticationHandler);
When It reaches 'OnSuccess' I am launching my connect activity and passing my session token along with the Intent. Moving to the next activity
private static final String COGNITO_POOL_ID = "us-west-2:a153a090-508c-44c0-a9dd-efd450298c4b";
private static final Regions MY_REGION = Regions.US_WEST_2;
AWSIotMqttManager mqttManager;
String clientId;
AWSCredentials awsCredentials;
CognitoCachingCredentialsProvider credentialsProvider;
#Override
protected void onCreate(Bundle savedInstanceState) {
super.onCreate(savedInstanceState);
setContentView(R.layout.activity_main);
Intent intent = getIntent();
if(null == intent){
Toast.makeText(getApplicationContext(), "Token is null", Toast.LENGTH_SHORT).show();
}else {
token = intent.getStringExtra("token");
}
clientId = UUID.randomUUID().toString();
credentialsProvider = new CognitoCachingCredentialsProvider(
getApplicationContext(),
COGNITO_POOL_ID,
MY_REGION
);
mqttManager = new AWSIotMqttManager(clientId, CUSTOMER_SPECIFIC_ENDPOINT);
Map loginsMap = new HashMap();
loginsMap.put("cognito-idp.us-west-2.amazonaws.com/us-west-2_ghtcc6ho9", token);
credentialsProvider.setLogins(loginsMap);
Log.d("SESSION_ID", ""+token);
new Thread(new Runnable() {
#Override
public void run() {
credentialsProvider.refresh();
awsCredentials = credentialsProvider.getCredentials();
Log.d("SESSION_ID B: ", ""+awsCredentials.getAWSAccessKeyId());
Log.d("SESSION_ID C: ", ""+awsCredentials.getAWSSecretKey());
}
}).start();
}
View.OnClickListener connectClick = new View.OnClickListener() {
#Override
public void onClick(View v) {
Log.d(LOG_TAG, "clientId = " + clientId);
try {
mqttManager.connect(credentialsProvider, new AWSIotMqttClientStatusCallback() {
#Override
public void onStatusChanged(final AWSIotMqttClientStatus status,
final Throwable throwable) {
Log.d(LOG_TAG, "Status = " + String.valueOf(status)+"______"+((null !=throwable)?throwable.getMessage():""));
runOnUiThread(new Runnable() {
#Override
public void run() {
if (status == AWSIotMqttClientStatus.Connecting) {
tvStatus.setText("Connecting...");
} else if (status == AWSIotMqttClientStatus.Connected) {
tvStatus.setText("Connected");
} else if (status == AWSIotMqttClientStatus.Reconnecting) {
if (throwable != null) {
Log.e(LOG_TAG, "Connection error.", throwable);
}
tvStatus.setText("Reconnecting");
} else if (status == AWSIotMqttClientStatus.ConnectionLost) {
if (throwable != null) {
Log.e(LOG_TAG, "Connection error.", throwable);
throwable.printStackTrace();
}
tvStatus.setText("Disconnected");
} else {
tvStatus.setText("Disconnected");
}
}
});
}
});
} catch (final Exception e) {
Log.e(LOG_TAG, "Connection error.", e);
}
}
};
What is wrong in my code? Why it throws exception when the MQTT connect is being invoked? Any help would be appreciated.
I beat my head up with this almost a week.
Full course of action ->
After succesfull login you will have a jwt token
String idToken = cognitoUserSession.getIdToken().getJWTToken();
put it into a map
Map<String, String> logins = new HashMap<String, String>();
//fill it with Cognito User token
logins.put("cognito-idp.<REGION>.amazonaws.com/<COGNITO_USER_POOL_ID>", idToken);
then use it to set in two places (not stated in any documentation!)
CognitoCachingCredentialsProvider credentialsProvider = new
CognitoCachingCredentialsProvider(context, IDENTITY_POOL_ID, REGION);
credentialsProvider.setLogins(logins);
and
AmazonCognitoIdentity cognitoIdentity = new AmazonCognitoIdentityClient(credentialsProvider);
GetIdRequest getIdReq = new GetIdRequest();
getIdReq.setLogins(logins); //or if you have already set provider logins just use credentialsProvider.getLogins()
getIdReq.setIdentityPoolId(COGNITO_POOL_ID);
GetIdResult getIdRes = cognitoIdentity.getId(getIdReq);
after that you still nedd to make some call
AttachPrincipalPolicyRequest attachPolicyReq = new AttachPrincipalPolicyRequest(); //in docs it called AttachPolicyRequest but it`s wrong
attachPolicyReq.setPolicyName("allAllowed"); //name of your IOTAWS policy
attachPolicyReq.setPrincipal(getIdRes.getIdentityId());
new AWSIotClient(credentialsProvider).attachPrincipalPolicy(attachPolicyReq);
and only after that you can enable connect button and continue like that
mqttManager.connect(credentialsProvider, new AWSIotMqttClientStatusCallback() {
Really for this small piece of code i spent a lot of time...
I was also getting same error -
Feb 27, 2019 10:23:09 AM com.amazonaws.services.iot.client.mqtt.AwsIotMqttConnectionListener onFailure
WARNING: Connect request failure
MqttException (0) - java.io.IOException: Already connected
at org.eclipse.paho.client.mqttv3.internal.ExceptionHelper.createMqttException(ExceptionHelper.java:38)
at org.eclipse.paho.client.mqttv3.internal.ClientComms$ConnectBG.run(ClientComms.java:664)
at java.lang.Thread.run(Thread.java:748)
Caused by: java.io.IOException: Already connected
at java.io.PipedOutputStream.connect(PipedOutputStream.java:100)
but the problem was different.
First of all, you do not need to call attachPrincipalPolicy from code. You can use the command line as well. You can do something like -
aws iot attach-principal-policy --principal us-east-1:1c973d17-98e6-4df6-86bf-d5cedc1fbc0d --policy-name "thingpolicy" --region us-east-1 --profile osfg
You will get the principal ID from identity browser of your identity pool. Now lets come to the error -
To successfully connect to mqtt with authenticated Cognito credentials, you need 2 correct policies -
Authenticated role corresponding to your identity pool should allow all mqtt operations.
AWS IoT policy should allow the same operations and you need to associate your cognito identity with this policy. We use attachPrincipalPolicy to do so.
If anyone step is missed we get above error. I agree the error is misleading - Already connected makes no sense to me for this. I would normally think it has to do with clientId, which should be unique. But anyways hopefully AWS folks would make this better at some point.
For my particular case issue was point 1. Though my IoT policy had all the required permissions, the auth role corresponding to the identity pool did not. So make sure you do that.
I have created a youtube video to show this as well: https://www.youtube.com/watch?v=j2KJVHGHaFc
When a client connected to broker it has a unique client ID. If clients tried to connect with same client id then this error occur. Use different client IDs like foo1, foo2, foo3, etc.
I successed connection bitween client in android and server.
but, when I want to send message like "hello" or whatever, message was disappear.
this is my client code:
group = new OioEventLoopGroup();
Bootstrap b = new Bootstrap();
b.group(group);
b.channel(OioSocketChannel.class);
b.option(ChannelOption.SO_KEEPALIVE, true);
b.handler(new ChannelInitializer<SocketChannel>() {
#Override
public void initChannel(SocketChannel ch) throws Exception {
ch.pipeline().addLast(handler);
}
});
Channel ch = null;
ChannelFuture f = null;
try {
f = b.connect(new InetSocketAddress(host, port)).sync();
ch = f.channel();
} catch (InterruptedException e) {
e.printStackTrace();
}
ch.writeAndFlush("hello!");
and this is my server code:
#Override
public void channelActive(ChannelHandlerContext ctx){
channels.add(ctx.channel());
ctx.channel().writeAndFlush("Welcome My Server");
System.out.println(ctx.channel().remoteAddress());
}
#Override
public void channelRead(ChannelHandlerContext ctx, Object msg) {
ByteBuf in = (ByteBuf) msg;
try {
while (in.isReadable()) {
System.out.print((char) in.readByte());
System.out.flush();
}
} finally {
ReferenceCountUtil.release(msg);
}
}
when I connect, Server was printing 'connected client ip address'
but after that, 'hello' message is not printed in my server.
what is wrong? server? client?
I think encode, decode is not problem, cuz nothing received
please let me know how to do for that?
If you want to write a String you need to put StringEncoder in the ChannelPipeline (on the client side). If you check the returned ChannelFuture of writeAndFlush(...) you will see it was failed.
We have contacted Google about this and we are on chat
The issue seems to be fixed for devices except Samsung phones.
I'm adding a Google+ sign in option to an app per the official instructions. Once the user has selected their account I would like my server to retrieve their Google+ profile info and update their profile on our site to match.
The first part - having the user select a Google account locally - seems to work just fine. When I try to request a token for the selected account, the Google auth dialog displays with the appropriate parameters; however, when I authorize the app using that dialog and re-request the token, GoogleAuthUtil.getToken(...) again throws a UserRecoverableAuthException (NeedPermission, not GooglePlayServicesAvailabilityException) and I get the same dialog asking me to approve!
This behavior is present on a Samsung S3 running Android 4.1.1 (with 3 Google accounts) and an Acer A100 running 4.0.3. It is NOT present on an HTC Glacier running 2.3.4. Instead, the HTC Glacier gives me a valid auth code. All devices have the latest iteration of Google Play Services installed and are using different Google+ accounts.
Anyone seen this before? Where can I start with debugging?
Here's the complete code - is anything obviously awry?
public class MyGooglePlusClient {
private static final String LOG_TAG = "GPlus";
private static final String SCOPES_LOGIN = Scopes.PLUS_LOGIN + " " + Scopes.PLUS_PROFILE;
private static final String ACTIVITIES_LOGIN = "http://schemas.google.com/AddActivity";
private static MyGooglePlusClient myGPlus = null;
private BaseActivity mRequestingActivity = null;
private String mSelectedAccount = null;
/**
* Get the GPlus singleton
* #return GPlus
*/
public synchronized static MyGooglePlusClient getInstance() {
if (myGPlus == null)
myGPlus = new MyGooglePlusClient();
return myGPlus;
}
public boolean login(BaseActivity requester) {
Log.w(LOG_TAG, "Starting login...");
if (mRequestingActivity != null) {
Log.w(LOG_TAG, "Login attempt already in progress.");
return false; // Cannot launch a new request; already in progress
}
mRequestingActivity = requester;
if (mSelectedAccount == null) {
Intent intent = AccountPicker.newChooseAccountIntent(null, null, new String[]{GoogleAuthUtil.GOOGLE_ACCOUNT_TYPE}, false,
null, GoogleAuthUtil.GOOGLE_ACCOUNT_TYPE, null, null);
mRequestingActivity.startActivityForResult(intent, BaseActivity.REQUEST_GPLUS_SELECT);
}
return true;
}
public void loginCallback(String accountName) {
mSelectedAccount = accountName;
authorizeCallback();
}
public void logout() {
Log.w(LOG_TAG, "Logging out...");
mSelectedAccount = null;
}
public void authorizeCallback() {
Log.w(LOG_TAG, "User authorized");
AsyncTask<Void, Void, String> task = new AsyncTask<Void, Void, String>() {
#Override
protected String doInBackground(Void... params) {
String token = null;
try {
Bundle b = new Bundle();
b.putString(GoogleAuthUtil.KEY_REQUEST_VISIBLE_ACTIVITIES, ACTIVITIES_LOGIN);
token = GoogleAuthUtil.getToken(mRequestingActivity,
mSelectedAccount,
"oauth2:server:client_id:"+Constants.GOOGLE_PLUS_SERVER_OAUTH_CLIENT
+":api_scope:" + SCOPES_LOGIN,
b);
} catch (IOException transientEx) {
// Network or server error, try later
Log.w(LOG_TAG, transientEx.toString());
onCompletedLoginAttempt(false);
} catch (GooglePlayServicesAvailabilityException e) {
Log.w(LOG_TAG, "Google Play services not available.");
Intent recover = e.getIntent();
mRequestingActivity.startActivityForResult(recover, BaseActivity.REQUEST_GPLUS_AUTHORIZE);
} catch (UserRecoverableAuthException e) {
// Recover (with e.getIntent())
Log.w(LOG_TAG, "User must approve "+e.toString());
Intent recover = e.getIntent();
mRequestingActivity.startActivityForResult(recover, BaseActivity.REQUEST_GPLUS_AUTHORIZE);
} catch (GoogleAuthException authEx) {
// The call is not ever expected to succeed
Log.w(LOG_TAG, authEx.toString());
onCompletedLoginAttempt(false);
}
Log.w(LOG_TAG, "Finished with task; token is "+token);
if (token != null) {
authorizeCallback(token);
}
return token;
}
};
task.execute();
}
public void authorizeCallback(String token) {
Log.w(LOG_TAG, "Token obtained: "+token);
// <snipped - do some more stuff involving connecting to the server and resetting the state locally>
}
public void onCompletedLoginAttempt(boolean success) {
Log.w(LOG_TAG, "Login attempt "+(success ? "succeeded" : "failed"));
mRequestingActivity.hideProgressDialog();
mRequestingActivity = null;
}
}
I've had this issue for a while and came up with a proper solution.
String token = GoogleAuthUtil.getToken(this, accountName, scopeString, appActivities);
This line will either return the one time token or will trigger the UserRecoverableAuthException.
On the Google Plus Sign In guide, it says to open the proper recovery activity.
startActivityForResult(e.getIntent(), RECOVERABLE_REQUEST_CODE);
When the activity returns with the result, it will come back with few extras in the intent and that is where the new token resides :
#Override
protected void onActivityResult(int requestCode, int responseCode, Intent intent) {
if (requestCode == RECOVERABLE_REQUEST_CODE && responseCode == RESULT_OK) {
Bundle extra = intent.getExtras();
String oneTimeToken = extra.getString("authtoken");
}
}
With the new oneTimeToken given from the extra, you can submit to the server to connect properly.
I hope this helps!
Its too late to reply but it may help to people having same concern in future.
They have mentioned in the tutorial that it will always throw UserRecoverableAuthException
when you invoke GoogleAuthUtil.getToken() for the first time. Second time it will succeed.
catch (UserRecoverableAuthException e) {
// Requesting an authorization code will always throw
// UserRecoverableAuthException on the first call to GoogleAuthUtil.getToken
// because the user must consent to offline access to their data. After
// consent is granted control is returned to your activity in onActivityResult
// and the second call to GoogleAuthUtil.getToken will succeed.
startActivityForResult(e.getIntent(), AUTH_CODE_REQUEST_CODE);
return;
}
i used below code to get access code from google.
execute this new GetAuthTokenFromGoogle().execute(); once from public void onConnected(Bundle connectionHint) and once from protected void onActivityResult(int requestCode, int responseCode, Intent intent)
private class GetAuthTokenFromGoogle extends AsyncTask<Void, Integer, Void>{
#Override
protected void onPreExecute()
{
}
#Override
protected Void doInBackground(Void... params) {
// TODO Auto-generated method stub
try {
accessCode = GoogleAuthUtil.getToken(mContext, Plus.AccountApi.getAccountName(mGoogleApiClient), SCOPE);
new ValidateTokenWithPhoneOmega().execute();
Log.d("Token -- ", accessCode);
} catch (IOException transientEx) {
// network or server error, the call is expected to succeed if you try again later.
// Don't attempt to call again immediately - the request is likely to
// fail, you'll hit quotas or back-off.
return null;
} catch (UserRecoverableAuthException e) {
// Recover
startActivityForResult(e.getIntent(), RC_ACCESS_CODE);
e.printStackTrace();
} catch (GoogleAuthException authEx) {
// Failure. The call is not expected to ever succeed so it should not be
// retried.
authEx.printStackTrace();
return null;
} catch (Exception e) {
throw new RuntimeException(e);
}
return null;
}
#Override
protected void onPostExecute(Void result)
{
}
}
I have got around this issue by using a web based login. I open a url like this
String url = "https://accounts.google.com/o/oauth2/auth?scope=" + Scopes.PLUS_LOGIN + "&client_id=" + webLoginClientId + "&response_type=code&access_type=offline&approval_prompt=force&redirect_uri=" + redirect;
The redirect url then handles the response and returns to my app.
In terms of my findings on using the Google Play Services, I've found:
HTC One is 3.1.59 (736673-30) - not working
Galaxy Note is 3.1.59 (736673-36) - not working
Nexus S is 3.1.59 (736673-34) - works
And I'd like to be involved in the chat that is occurring, however I don't have a high enough reputation to do so.
I've experienced the same issue recently - it appears to be device-specific (I had it happen every time on one S3, but on another S3 running the same OS it didn't happen, even with the same account). My hunch is that it's a bug in a client app, either the G+ app or the Google Play Services app. I managed to solve the issue on one of my devices by factory resetting it (a Motorola Defy), then reinstalling the Google Play Services app, but that's a completely useless solution to tell to users.
Edit (6th Aug 2013): This seems to have been fixed for me without any changes to my code.
The first potential issue I can see is that you are calling GoogleAuthUtil.getToken() after you get the onConnected() callback. This is a problem because requesting an authorization code for your server using GoogleAuthUtil.getToken() will always show a consent screen to your users. So you should only get an authorization code for new users and, to avoid showing new users two consent screens, you must fetch an authorization code and exchange it on your server before resolving any connection failures from PlusClient.
Secondly, make sure you actually need both a PlusClient and an authorization code for your servers. You only need to get a PlusClient and an authorization code if you are intending to make calls to the Google APIs from both the Android client and your server. As explained in this answer.
These issues would only result in two consent dialogs being displayed (which is clearly not an endless loop) - are you seeing more than two consent dialogs?
I had a similar problem where an apparent auth loop kept creating {read: spamming} these "Signing In..." and Permission request dialogs while also giving out the discussed exception repeatedly.
The problem appears in some slightly-modified example code that I (and other like me, I suspect) "cargo-culted" from AndroidHive. The solution that worked for me was ensuring that only one background token-retrieval task runs at the background at any given time.
To make my code easier to follow, here's the auth flow in my app (that is almost identical to the example code on AndoidHive): Activity -> onConnected(...) -> getProfileInformation() -> getOneTimeToken().
Here's where getOneTimeToken() is called:
private void getProfileInformation() {
try {
if (Plus.PeopleApi.getCurrentPerson(mGoogleApiClient) != null) {
Person currentPerson = Plus.PeopleApi
.getCurrentPerson(mGoogleApiClient);
String personName = currentPerson.getDisplayName();
String personPhotoUrl = currentPerson.getImage().getUrl();
String personGooglePlusProfile = currentPerson.getUrl();
String email = Plus.AccountApi.getAccountName(mGoogleApiClient);
getOneTimeToken(); // <-------
...
Here's my getOneTimeToken():
private void getOneTimeToken(){
if (task==null){
task = new AsyncTask<Void, Void, String>() {
#Override
protected String doInBackground(Void... params) {
LogHelper.log('d',LOGTAG, "Executing background task....");
Bundle appActivities = new Bundle();
appActivities.putString(
GoogleAuthUtil.KEY_REQUEST_VISIBLE_ACTIVITIES,
ACTIVITIES_LOGIN);
String scopes = "oauth2:server" +
":client_id:" + SERVER_CLIENT_ID +
":api_scope:" + SCOPES_LOGIN;
String token = null;
try {
token = GoogleAuthUtil.getToken(
ActivityPlus.this,
Plus.AccountApi.getAccountName(mGoogleApiClient),
scopes,
appActivities
);
} catch (IOException transientEx) {
/* Original comment removed*/
LogHelper.log('e',LOGTAG, transientEx.toString());
} catch (UserRecoverableAuthException e) {
/* Original comment removed*/
LogHelper.log('e',LOGTAG, e.toString());
startActivityForResult(e.getIntent(), AUTH_CODE_REQUEST);
} catch (GoogleAuthException authEx) {
/* Original comment removed*/
LogHelper.log('e',LOGTAG, authEx.toString());
} catch (IllegalStateException stateEx){
LogHelper.log('e',LOGTAG, stateEx.toString());
}
LogHelper.log('d',LOGTAG, "Background task finishing....");
return token;
}
#Override
protected void onPostExecute(String token) {
LogHelper.log('i',LOGTAG, "Access token retrieved: " + token);
}
};
}
LogHelper.log('d',LOGTAG, "Task setup successful.");
if(task.getStatus() != AsyncTask.Status.RUNNING){
task.executeOnExecutor(AsyncTask.SERIAL_EXECUTOR); //double safety!
} else
LogHelper.log('d',LOGTAG,
"Attempted to restart task while it is running!");
}
Please note that I have a {probably redundant} double-safety against the task executing multiple times:
if(task .getStatus() != AsyncTask.Status.RUNNING){...} - ensures that the task isn't running before attempting to execute it.
task.executeOnExecutor(AsyncTask.SERIAL_EXECUTOR);- makes sure that copies of this task are "synchronized" (i.e. a queue is in place such that only one task of this type can executed at a given time).
P.S.
Minor clarification: LogHelper.log('e',...) is equivalent to Log.e(...) etc.
you should startactiviy in UI thread
try {
....
} catch (IOException transientEx) {
....
} catch (final UserRecoverableAuthException e) {
....
runOnUiThread(new Runnable() {
public void run() {
startActivityForResult(e1.getIntent(), AUTH_CODE_REQUEST);
}
});
}
Had the same bug with infinite loop of permission request. For me it was because time on my phone was shifted. When I check detect time automatically this bug disappeared. Hope this helps!