I used this guide to built a showcase - sign in with LinkedIn into a specific site.
Everything worked perfectly until I demonstarted it in front of a wide audience and it broke down :-( It was a great FAIL and I want to know why. Here is what I do:
1.On the sign in page the user may click a Sign in with LinkedIn button and is redirected to similar link:
https://www.linkedin.com/oauth/v2/authorization?redirect_uri=[my_callback]&client_id=[my_client_id]&response_type=code&state=[securely_random]&scope=r_basicprofile%20r_emailaddress
2.The user allows the application and is sent back to my_callback
3.In my_callback I make a POST to https://www.linkedin.com/oauth/v2/accessToken in order to obtain an access token. I use the code sent by LinkedIn, correct client ID and secret. Everything is OK, e.g the response might be:
{
"access_token": [access_token],
"expires_in": 5184000
}
4.I make authenticated requests to fetch the profile data from endpoint https://www.linkedin.com/v1/people/~:(firstName,lastName,email_address)
Headers:
x-li-format: json
Authorization: Bearer [access_token]
I started to get an error 401 occasionally, e.g.:
{
"errorCode": 0,
"message": "Unable to verify access token",
"requestId": "YX21AN6NZG",
"status": 401,
"timestamp": 1483732371224
}
It seems that some of the requests randomly passed nevertheless...
Additional details:
The user is logged in LinkedIn
The user is administrator for the LinkedIn application
I have checked the limitations (throttle limits) at in the application. Available at https://www.linkedin.com/developer/apps. Everything which can be seen is green.
I have tried all advices and hacks from this question
My app is not live
I'm puzzled!
Question: Any obvious mistake?
Question: Is there any hidden throttle limits (or security instruments) for the limitation of the number of access tokens for specific user/app combination? (I'm always using the same user and I tested pretty aggressively before the big FAIL)
UPDATE: In the next two days the Sign in started working smoothly again as described above. No 401-s anymore... :-X I've made no changes to the code base. So is this some kind of throttle limit or just LI was in a bad mood on Friday?
In case someone is curious I got an answer to my problem from LI support:
Unfortunately, we really can't assist with API issues and 3rd party apps. My guess is that there was a hiccup on Friday and you were the victim of bad timing.
I accept the explanation that I was a victim so this answers my question...
I have an access-token that worked to get data through the API, however now it has stopped working. I've carefully read LinkedIn's documentation: https://developer.linkedin.com/docs/oauth2 and have come up with why this can happen.
The docs state, that the user's session is linked with the access-token. Therefore, logging out of the session means the access-token is invalidated. This makes sense because it's exactly what I see happening.
The oauth2 expired-at is just a timestamp of the ultimate time this access-token will be valid. But it can be invalidated at any moment apparently.
Other oauth2 implementations show features for refreshing the access-token, Linkedin does not provide such feature. Therefore a user has to refresh it manually every time. Not sure if this is by design or they haven't got around to it yet. Overall their API feels pretty out-dated.
Related
I have an app that uses the Twitter API where users can authenticate via twitter and retweet/like/follow through my app. Randomly this week the logs are showing "code: 89 Invalid or expired token".
Naturally, I go login to twitter to see the status of my app, and nothing seems out of the ordinary. I saw others with this issue had success regenerating their keys and replacing them in their application.
This didn't help.
One important thing to note is nothing has changed in the code of my application for the last 3-4 months, so I doubt it's anything in there. It's been working for over two years without any issue.
The thing I suspect the most is perhaps Twitter decided to suspend my app; Although, I don't see anywhere that is the case, and I thought I'd receive an email from them about it if it were.
I'm at a loss and would appreciate some possible solutions or alternative avenues I can pursue to find the culprit.
The keys associated with your app are the API Key (Consumer Token) and API secret key (Consumer Secret). The error you're getting is for the Access token, which belongs to the user. It sounds like the user associated with that request needs to authorize your app again before it can operate again with their access key. This can happen if the user removes authorization for your app by visiting their Settings/Privacy and safety/Apps and sessions.
If you were using your own access token in a scenario like single-user authorization, then regenerating the key might work, but in this case, the only way to get new keys for that user is for them to go through the sign-in process to authorize your app again. e.g. you could log who the user was that the error occurred on and send them a notification to re-authorize.
I want my app to chat.postMessages to Slack, using the icon_emoji and username parameters so that I sometimes change the icon and name displayed.
From the doc of the method, it seems I have to set as_user to false. But when I do so, I'm getting a missing_scope error: I need scope chat:write:bot. (Same error whether I use my OAuth access token or my OAuth bot user access token)
But from this page I got that they are disappearing? So are the parameters I want to use (icon_emoji and username) going to disappear as well?
No, as of Dec 19 2019, you cannot use icon_emoji, etc
I ran into the same issue, and so I opened a support ticket with Slack. After a few messages back and forth, I got confirmation that this is likely a bug:
With your help I was able to replicate the issue. I suspect this is a bug but I've escalated it to the Granular Bot Token team for a second opinion. As soon as I have more details I will let you know.
Edit:
I've received another update, which I've trimmed down to the essentials:
When building the Granular Bot Token model our team really wanted to use this as an opportunity to slowly deprecate certain API behaviour. The as_user parameter is one of them. See the following comment from (https://api.slack.com/authentication/quickstart).
One effect worth noting: chat.postMessage and other chat.* methods no longer mess around with the as_user parameter. You're granted a single chat:write scope (no :user or :bot is appended). If you call the chat.postMessage method with your bot token, you post as the bot. If you've obtained a user token through the new install flow, and you call the method with your user token, you post as the user.
So in regards to the action you are wanting to take, I'm afraid it's not possible with Granular Bot Token. We also recognized that the error this is throwing is completely unhelpful and confusing for developers. For this, we've created a bug to fix the error that's returned when Granular Bot tokens try to make restricted API calls.
I've requested clarification on whether it is just the as_user part that is deprecated, or the icon_emoji part as well.
Edit 2 (Dec 19, 2019 13:38 EST):
One, probably final, update:
The option for app to change information such as username, avatar and icon via the API is up for debate as well. Currently with Granular Bot Tokens you can only update that information via your API management page.
Which is really unfortunate.
I had to add a Scope to "Bot Token Scopes" with the value of "chat:write.customize".
https://api.slack.com/scopes/chat:write.customize
then I had to re-install my app. And it worked like I expected.
Note: I tried this as it was a suggestion from a comment. but it worked so i am adding it as an answer.
I am facing an issue regarding account linking in Actions on Google:
I am able to authenticate the user and access his email address and username however after this how can I redirect the user back to the google assistant and close the browser where he was authenticated?
Any help will be appreciated!
Update: Hey Prisoner thanks a lot for that.
I did what you said and yeah now it does redirect to google.com but without result_code=SUCCESS when I test it in the simulator.
The link is:
https://www.google.co.in/?gws_rd=cr&dcr=0&ei=z77fWbjQGIXxvATs_oqwBA
Now if I type talk to... again it shows me the message you need to link your account!
In the device the browser automatically closes and it shows SIGNING_IN however when I type an intent it is not recognized.
It would be great if you could point me in the right direction! (I am not sure but I might be at the token exchange stage that you mentioned, but I don't have a clue how to proceed!)
Update 2: As requested the entire flow that I am following:
This is the URL that I receive from debugInfo:
https://assistant.google.com/services/auth/handoffs/auth/start?account_name=cha***#gmail.com&provider=***_dev&scopes=email&return_url=https://www.google.com/
When I paste this in the browser the request that I receive at the authorization endpoint is:
ImmutableMultiDict([
('response_type', 'code'),
('client_id', ****.apps.googleusercontent.com'),
('redirect_uri', 'https://oauth-redirect.googleusercontent.com/r/****'),
('scope', 'email'),
('state', ' CtcCQUxWM2ROU3hNMjl4LUItVXhQSGd4THRMLU4yNExnb3lYbGRKQnQwa3NwTVFva19NUWpYNE5jNGJURzIyZFN3RDBXd2d4enFGVWJGb0Q0ZW1vaS1OaFdkaHdhb05HZ2xlWTR6SllKVlRWYktwd09faklyUTVheFhQbGw2dmVKYzVFTk05N3B1QkxaZG41RVdHN0wyTktvRFdCYzFPVFBzM1dQUlFtN2RmM1VtRU4****(state)')
])
The response (redirect_url) that I send back:
https://accounts.google.com/o/oauth2/v2/auth?scope=email&response_type=code&redirect_uri=https%3A%2F%2F******.herokuapp.com%2Fcallback%2Fgoogle&client_id=****.apps.googleusercontent.com
When it reaches my endpoint again the request arguments are:
ImmutableMultiDict([
('code', '4/***********')
])
Now I am able to access the email address and other details
The url that I redirect to from here:
https://oauth-redirect.googleusercontent.com/r/****?code=abcdefgh&state=CtcCQUxWM2ROU3hNMjl4LUItVXhQSGd4THRMLU4yNExnb3lYbGRKQnQwa3NwTVFva19NUWpYNE5jNGJURzIyZFN3RDBXd2d4enFGVWJGb0Q0ZW1vaS1OaFdkaHdhb05HZ2xlWTR6SllKVlRWYktwd09faklyUTVheFhQbGw2dmVKYzVFTk05N3B1QkxaZG41RVdHN0wyTktvRFdCYzFPVFBzM1dQUlFtN2RmM1VtRU4****(state)
This redirects me to :
https://www.google.co.in/?gws_rd=cr&dcr=0&ei=5c_fWdfKNYndvASO7o6ACA
Edit 3: I checked the network logs:
result_code=FAILURE&result_message=Account+linking+failed
I also added /token/google as the token URL in AoG. It is detected in heroku however I never receive this request in my code.
Note: I am using python flask and hosting my app on heroku
Once you have authenticated the user, you'll need to return a temporary auth code back to Google. Later, Google will exchange this auth code for an access token and a refresh token, but you're not there yet. The important part is that this code needs to be unique and that, later, you'll be able to recognize what user it is for. The code should be valid for a limited time - 10 minutes is a generally accepted time frame.
In the request Google sent to you as part of the login, they've provided a redirect_uri and a state as parameters. You'll need to use these in your reply. (state can be anything - you shouldn't care what it is, you're just going to send it back with your redirect. Its purpose is to improve security by preventing replay attacks.)
Verify that the redirect_uri has the form
https://oauth-redirect.googleusercontent.com/r/YOUR_PROJECT_ID
Where YOUR_PROJECT_ID is... you guessed it, the ID of your project. You can find this in the cloud console.
You'll then redirect the user to this URL with a few additional parameters:
https://oauth-redirect.googleusercontent.com/r/YOUR_PROJECT_ID?code=AUTHORIZATION_CODE&state=STATE_STRING
Where YOUR_PROJECT_ID is as noted above, AUTHORIZATION_CODE is the code you've generated, and STATE_STRING is the value of the state parameter that you were sent in the request.
For details, you can see https://developers.google.com/actions/identity/oauth2-code-flow#handle_user_sign-in
A significant number of users are getting the error message of 'GetAccessToken - Unauthorized error: Your network is not allowed to request an OAuth token for this Application' whilst trying to log in to our application.
I have checked and tested, and they are accessing the correct network with the right credentials. This is happening to almost every person, so I don't believe this to be a coincidence.
This has only begun to happen in the past 24-36 hours. Are there any suggestions as to why this error is occurring, has something changed that I may have missed?
Thanks!
Couple of possible scenarios:
It is likely that the new simplified SSO is rolling out for these
users, which means they will use their O365 credentials to login to
Yammer from now on. Every user will need to reauth external apps (I
had to for our own apps also when this change occurred for us last
week).
Tokens also can expire when user's passwords expire, so if
they have a mandatory time interval to reset this could trigger an en masse need to reauth
Make sure your JS Origins are specified in your app details if you are using the Login with Yammer button and/or the JS SDK. http://naomimoneypenny.com/2015/02/11/yammer-apps-javascript-origins-update/
There isn't clear and definitive guidance regrettably from Yammer as to why and when tokens expire. Just that they do e.g. from https://developer.yammer.com/v1.0/docs/oauth-2
Once the token expires, you will need to re-run the steps above to generate a new code and access_token.
** UPDATE **
It truly seems that Google has just screwed every single person on the planet by absolutely requiring user interaction to upload a video. Of course I know, they are free. Exactly what I warned the client years ago about, so I don't need to be reminded. Thank You.
So I would like to try to take this in a different direction and just find a loophole and a workaround to still keep doing what we are doing in spite of Google's complete lack of support or caring in any way about the developers and what they have to deal with.
It would be different if you can actually call a phone number and talk to a human being about YouTube Partner access, but you can more quickly get access to the Illuminati.
OAuth 2.0 is now the only supported authentication method period. It does require user interaction.
But what about that token? Does anybody know how long the token lasts?
If I can obtain a token just once using user interaction and place it in the database, I can automate possibly hundreds or thousands of interactions afterwards.
In other words, I'm trying to turn the user interaction into a speed bump instead of a concrete wall.
If anybody has any examples of obtaining that token, caching it, and using it afterwards, that would be a godsend to me right now.
Thanks for the comments and the help. I'm not surprised that the YouTube Developers Forum just folded and said to come here instead :)
It seems that Google has completely pulled the plug on the existing dashboard.
https://code.google.com/apis/youtube/dashboard/gwt/index.html
That link is now 404'd. Tried from several different browsers on different systems.
Registered under the new Google APIs Console already, but still get the problem.
// Set the authentication URL for this connection object
$authenticationURL= 'https://www.google.com/youtube/accounts/ClientLogin';
// Try to connect to YouTube with the channel credentials passed
try {
$httpClient =
Zend_Gdata_ClientLogin::getHttpClient(
$username = $channelfields['EMAIL_ADDRESS'],
$password = $channelfields['PASSCODE'],
$service = 'youtube',
$client = null,
$source = 'Redacted Data',
$loginToken = $channelfields['CACHED_TOKEN'],
$loginCaptcha = '',
$authenticationURL);
} catch (Zend_Gdata_App_HttpException $httpException) {
$update_error['response_body'] = $httpException->getRawResponseBody();
$update_error['error'] = 1;
} catch (Zend_Gdata_App_Exception $e) {
$update_error['message'] = $e->getMessage();
$update_error['error'] = 1;
}
This code has worked perfectly fine before, but does not work with the older API key, or the newer one generated inside the Google APIs console.
I'm attempting a simple upload and this concerns me greatly:
"The service account flow supports server-to-server interactions that do not access user information. However, the YouTube Data API does not support this flow. Since there is no way to link a Service Account to a YouTube account, attempts to authorize requests with this flow will generate a NoLinkedYouTubeAccount error."
From all reports it seems that Google has forced YouTube uploads to become interactive in all cases precluding all possibility of platforms that automatically upload generated content from working at all.
Any help or insights into the process is appreciated.
P.S - Ohhh, it's been awhile since I looked at that system and Google shut down the YouTube Developer Forums and said "YOU" were responsible for their support now :)
OAuth2 does support the ability to avoid user interaction through the offline access type parameter (ie, using access_type=offline). Check out Google documentation for details.
The solution is really rather simple. Your app needs to use oauth to request offline access. It will be given an access cide which you convert to a refresh token, which is the thing you store in your database. This doesn't expire. Well actually it sometimes does, but that's another story. Whenever you need to access the api, use the stored refresh token to request an access token which you include in each api call.
See https://developers.google.com/accounts/docs/OAuth2WebServer for details.
I don't know what you did but https://code.google.com/apis/youtube/dashboard/gwt/index.html works perfectly fine for me. Maybe it was a temporary issue. If you want no user interaction you HAVE to use YouTube API v2 OR you have to use v3 with methods that don't require authentification OR you have to provide your own youtube account credentials which is not recommended and probably not appropriate for you situation.
Several issues to respond here, I think.
1) The older API console has not been removed, but I've noticed intermittent outages to it and to the newer API console while Google is rolling out their new "cloud console."
2) ClientLogin was officially deprecated in April of 2012, not just 48 hours ago. Jeff Posnick has detailed all the changes over the months (and related ones, such as AuthSub, Youtube Direct, etc.) at his blog (apiblog.youtube.com).
3) You're right that, with v3 of the APIs, you cannot do automatic uploads across the board, as the oAuth2 flow requires user interaction. However, given the limited description of your use case, using refresh tokens is probably your best bet. If the content is user generated, somewhere they must be logging into your app, correct? (so that your app knows which credentials to leverage to do the uploads). At the point they're logging into your app, and you're starting the oAuth2 flow, you just have to hit the first oAuth endpoint and pass it the parameter access_type=offline (along with any other parameters). This will ensure that, when they grant that initial permission, you're returned a refresh token instead of an access token. With that refresh token, you can exchange it for multiple access tokens as needed (an access token lives for about an hour. I don't know how long a refresh token lives, but I've never had one expire before my own login cookies did, and then I just get a new one when my users re-login to my app).
Here's some more info on how to use the refresh token; note, too, that the various google api client libraries make it pretty smooth.
https://developers.google.com/accounts/docs/OAuth2WebServer#refresh
Also, this video tutorial from a Google Developers Live broadcast a couple of months ago might help illustrate the point: http://www.youtube.com/watch?v=hfWe1gPCnzc -- it's using the oAuth playground rather than a client library, but the concept is the same.
The answer is to use google-api-php-client, create an interactive auth page, and set up YouTube API v3 correctly with the new API console.
You can create a very simple page that will authenticate for the supplied channel and then store the correct token in your database. Is already working and uploading hundreds of videos on one channel. You do need to remember to fully activate yourself under the new API console and add the services required. Just keep authenticating and adding the services it says it needs. After that, the regular v3 upload process works just fine. On failure send a group an email and they can get a new token in 10 seconds.
Not the most elegant solution, but the documentation from Google is far from elegant anyways that Stack Overflow is now their front line support.
Just hang in there, a solution is always found. Don't give up!
I didn't get here by myself either, the other answers on this page helped me get all the way to this point. Thanks guys.
P.S - Don't forget the scopes
$client->setScopes("https://www.googleapis.com/auth/youtube https://www.googleapis.com/auth/youtube.upload");