Using generated page access_token in multiple regions - facebook

I am trying to subscribe my facebook chatbot to the user's page, using the manage_pages permission.
I already have the permission approved to my app, but I am having trouble, because the token is generated in the browser(in Brazil), but I send it to aws lambda in US west 2, and I get the following error
The \'manage_pages\' permission must be granted before impersonating a user\'s page.
In my test, I run the subscribe process in my machine, and reuse this token in AWS Lambda(US West 2). After researching and making some tests, I figured out that it just dont work in USA, in Brazil it works.
Using the same token generated in the process, I run the following test.
var request = require("request")
request.post("https://graph.facebook.com/v2.9/me/messages", {
qs: {
access_token: "GENERATED TOKEN"
},
json: {
"recipient": {
"id": "USER_ID"
},
"message": {
"text": "hello!"
}
}
},function(e){
if(e){
return console.log("ERROR",e)
}
console.log("FINISH",arguments)
})
When I run with my internet, it works fine, but If I use a VPN in Miami, the error occurs.
I believe it has something to do with the data replication that facebook does, but how can I make this token be accessible by the USA region?
EDIT
I found this error when starting the facebook login process with the VPN online:
I didnt find any restrictions in my app, where can I check this?

The problem was occurring because my app had coutry restriction to only Brazil, I added United States to restrictions and it worked.
The restriction can be changed in Settings -> Advanced, App Restrictions

Related

Google Sign-In with Passportjs not getting authenticated

I'm using Sails with Passport for authentication. I'm using passport-google-oauth(OAuth2Strategy) and passport-facebook for enabling Google Sign-in.
I'm not too well-versed with Passport, so pardon me if this is a rookie question. I've set up login via Facebook and it works just fine. With Google, I do receive an authorization code after allowing access to the app, but the I'm eventually not authenticated. I'm guessing the same code should work for both Facebook and Google since the strategies are both based on oauth2.
I'm not even sure what code to share, since I'm using the auto-generated code from sails-generate-auth, but do let me know if there's anything else I can share.
Any ideas on why this might be happening? The app is locally hosted but that's unlikely to be the problem since I am getting to the authorization stage anyway.
I faced the same problem and it was located here in in api/services/passport.js:
// If the profile object contains a list of emails, grab the first one and
// add it to the user.
if (profile.hasOwnProperty('emails')) {
user.email = profile.emails[0].value;
}
// If the profile object contains a username, add it to the user.
if (profile.hasOwnProperty('username')) {
user.username = profile.username;
}
// If neither an email or a username was available in the profile, we don't
// have a way of identifying the user in the future. Throw an error and let
// whoever's next in the line take care of it.
if (!user.username && !user.email) {
return next(new Error('Neither a username nor email was available'));
}
The Google service was not returning a profile.username property.
Because of it, the user is not saved in the database and cannot be authenticated. Then the passport callback receives an empty user, so the function that handles errors is fired and the user is redirected to the login page.
This change allows to use the displayName property as the username:
// If the profile object contains a list of emails, grab the first one and
// add it to the user.
if (profile.hasOwnProperty('emails')) {
user.email = profile.emails[0].value;
}
// If the profile object contains a username, add it to the user.
if (profile.hasOwnProperty('username')) {
user.username = profile.username;
}
/** Content not generated BEGIN */
// If the username property was empty and the profile object
// contains a property "displayName", add it to the user.
if (!user.username && profile.hasOwnProperty('displayName')) {
console.log(profile); // <= Use it to check the content given by Google about the user
user.username = profile.displayName;
}
/** Content not generated END */
// If neither an email or a username was available in the profile, we don't
// have a way of identifying the user in the future. Throw an error and let
// whoever's next in the line take care of it.
if (!user.username && !user.email) {
return next(new Error('Neither a username nor email was available'));
}
You could also use the profile.id property because profile.displayName is not necessarily unique (ie: two Google accounts can have an identical displayName). But it is also true accross different services: a Twitter account could also have the same username than a Facebook account. If both register on your application, you will have a bug. This is a problem from the code generated by sails-generate-auth and you should adapt it with the behavior that you want.
I will propose a PR if this solution works for you too.
Alright, so this ultimately turned out to be a known issue with the API.
TL;DR: Enable the Google+ API and the Contacts API as mentioned here. (The Contacts API isn't required, as #AlexisN-o pointed out in the comments. My setup worked as desired with Contacts API disabled. This obviously depends on what scope you're using.)
I believe it's not a nice way of failing since this was an API error that was prevented from bubbling up. Anyway, I dug into passport.authenticate to figure out what was going wrong. This eventually calls the authenticate method defined in the package corresponding to the strategy (oauth2 in this case). In here (passport-google-oauth/lib/passport-google-oauth/oauth2.js) I found that the accessToken was indeed being fetched from Google, so things should be working. This indicated that there was a problem with the requests being made to the token urls. So I ventured a little further into passport-oauth2/lib/strategy.js and finally managed to log this error:
{ [InternalOAuthError: failed to fetch user profile]
name: 'InternalOAuthError',
message: 'failed to fetch user profile',
oauthError:
{ statusCode: 403,
data: '{
"error": {
"errors": [{
"domain": "usageLimits",
"reason": "accessNotConfigured",
"message": "Access Not Configured. The API (Google+ API) is not enabled for your project. Please use the Google Developers Console to update your configuration.",
"extendedHelp": "https://console.developers.google.com"
}],
"code": 403,
"message": "Access Not Configured. The API (Google+ API) is not enabled for your project. Please use the Google Developers Console to update your configuration."
}
}'
} }
This was the end of the hunt for me and the first result for the error search led to the correct answer. Weird fix though.

Is there a way to fully test “publish_actions” to comment (other user posts) without submitting app for review?

I created an app on the FB developer page, and correctly configured it to allow “publish_actions”. Also tested to comment my own posts using the graph API and everything went OK.
https://graph.facebook.com/v2.2/xxx/comment
Sending the field message “some comment” (as post)
The problem is whenever that “Comment ID” does not belong to me, for example a friend or a page post that I am trying to comment, I always get:
{
"error": {
"message": "(#200) Permissions error",
"type": "OAuthException",
"code": 200
}
}
(The token generation, and permission requests worked flawlessly, also even added all the permission available to test, but no luck)
Is there any way to test the comment function on other person post or page without submitting the app? Or is there a way to create “test” pages (as there is for test users) to test this?

Posting on user's timeline on the behalf of the user

I created and APP on the Facebook.
Now I am trying to post a comment on the user's timeline using the Graph API but I cannot undestand what I need to do.
I already authorized the Application (I authorized this permissions: 'user_status,publish_stream,user_photos,email').
If I look on my profile -> privacy I can see that the APP can:
This app can: Post on your behalf
This app may post on your behalf, including status updates, photos and more.
Last data access:
Basic InformationToday
See details · Learn more
Posts on your behalf:
Who can see posts this app makes for you on your Facebook timeline?
Public
Notifications:
When to notify you?
The app sends you a notification
So this part seems OK.
On my application I do the following to try to post somenthing on my timeline (I need to post it when I am off line).
1) GET : https://graph.facebook.com/oauth/access_token?client_id=APPID&client_secret=*APP_SECRET*&grant_type=client_credentials
RESPONSE:
access_token = 422828347771671|UdQELQIf0N7krF4JUo7VwtPLTkk
2) GET: https://graph.facebook.com/search?q=myemail&type=user
RESPONSE:
_"error":_{
______"message":_"A_user_access_token_is_required_to_request_this_resource_",
______"type":_"OAuthException",
______"code":_102
___}
But reading the documentation (https://developers.facebook.com/docs/reference/api/#searching) it seems that this request does not need any kind of access token.
I also tried to add the access token, but the result is still the same.
3) I know my facebookid so I tried to use it directly:
POST: https://graph.facebook.com/100001139132403/feed
ARGUMENT:
access_token=422828347771671|UdQELQIf0N7krF4JUo7VwtPLTkk,
message=Hello
RESPONSE:
"error": {
"message": "(#200) This API call requires a valid app_id.",
"type": "OAuthException",
"code": 200
}
}
So I tried to add client_id=APPID and app_id=APPID, I tried to put one, the other and also both arguments as GET or POST but nothing changed.
I checked the APPId ad it is correct.
Do someone have any idea?
Thank you!
You cant simply generate a user access_token by querying to a URL. What you did gives the App access token and as the error say, you are trying to do something, that needs user access_token.
Refer and implement this : https://developers.facebook.com/docs/howtos/login/server-side-login/
And use that access_token for the rest of the processes.

How to verify Facebook access token?

There's only thing that server has to do; just check any access token's validity.
Clients send to the server user id and access token obtained by FB.getLoginStatus. As I expected, there would be any URL that checks access token's validity, like http://xxx.facebook.com/access_token?=xxxxxxxxxxxxxxxxxxxxxxxxxxxx.
That returns whether it's available one or not or is there any API (server side) for that?
The officially supported method for this is:
GET graph.facebook.com/debug_token?
input_token={token-to-inspect}
&access_token={app-token-or-admin-token}
See the check token docs for more information.
An example response is:
{
"data": {
"app_id": 138483919580948,
"application": "Social Cafe",
"expires_at": 1352419328,
"is_valid": true,
"issued_at": 1347235328,
"metadata": {
"sso": "iphone-safari"
},
"scopes": [
"email",
"publish_actions"
],
"user_id": 1207059
}
}
You can simply request https://graph.facebook.com/me?access_token=xxxxxxxxxxxxxxxxx if you get an error, the token is invalid. If you get a JSON object with an id property then it is valid.
Unfortunately this will only tell you if your token is valid, not if it came from your app.
Just wanted to let you know that up until today I was first obtaining an app access token (via GET request to Facebook), and then using the received token as the app-token-or-admin-token in:
GET graph.facebook.com/debug_token?
input_token={token-to-inspect}
&access_token={app-token-or-admin-token}
However, I just realized a better way of doing this (with the added benefit of requiring one less GET request):
GET graph.facebook.com/debug_token?
input_token={token-to-inspect}
&access_token={app_id}|{app_secret}
As described in Facebook's documentation for Access Tokens here.
Simply request (HTTP GET):
https://graph.facebook.com/USER_ID/access_token=xxxxxxxxxxxxxxxxx
That's it.
The app token can be found from this url.
https://developers.facebook.com/tools/accesstoken
I found this official tool from facebook developer page, this page will you following information related to access token - App ID, Type, App-Scoped,User last installed this app via, Issued, Expires, Data Access Expires, Valid, Origin, Scopes.
Just need access token.
https://developers.facebook.com/tools/debug/accesstoken/
Exchange Access Token for Mobile Number and Country Code (Server Side OR Client Side)
You can get the mobile number with your access_token with this API https://graph.accountkit.com/v1.1/me/?access_token=xxxxxxxxxxxx. Maybe, once you have the mobile number and the id, you can work with it to verify the user with your server & database.
xxxxxxxxxx above is the Access Token
Example Response :
{
"id": "61940819992708",
"phone": {
"number": "+91XX82923912",
"country_prefix": "91",
"national_number": "XX82923912"
}
}
Exchange Auth Code for Access Token (Server Side)
If you have an Auth Code instead, you can first get the Access Token with this API - https://graph.accountkit.com/v1.1/access_token?grant_type=authorization_code&code=xxxxxxxxxx&access_token=AA|yyyyyyyyyy|zzzzzzzzzz
xxxxxxxxxx, yyyyyyyyyy and zzzzzzzzzz above are the Auth Code, App ID and App Secret respectively.
Example Response
{
"id": "619XX819992708",
"access_token": "EMAWdcsi711meGS2qQpNk4XBTwUBIDtqYAKoZBbBZAEZCZAXyWVbqvKUyKgDZBniZBFwKVyoVGHXnquCcikBqc9ROF2qAxLRrqBYAvXknwND3dhHU0iLZCRwBNHNlyQZD",
"token_refresh_interval_sec": XX92000
}
Note - This is preferred on the server-side since the API requires the APP Secret which is not meant to be shared for security reasons.
Good Luck.

Get application id from user access token (or verify the source application for a token)

I found this question, which has an answer, but facebook changed the token format since then, now it is something like:
AAACEdEose0cBACgUMGMCRi9qVbqO3u7mdATQzg[more funny letters]ig8b3uss9WrhGZBYjr20rnJu263BAZDZD
In short, you cannot infer anything from it.
I also found the access token debugger, which shows the information I am looking for if you paste a token in, which is nice, but does not help me do it programmatically.
Point is, if someone gets a token for a user, he can use it to access the graph, which is what I do in my application - I want to be sure that people are forwarding the token that was issued to them by my application, and not another.
My application flow is:
Get access token from facebook (nothing special, in the way it is described in here , Server-side Flow. (also iPhone and android and used, but they have similar flows if I recall correctly))
[device] <-> [facebook]
With that access token, the device will access my application server with the token
[device] <-> [Jonathan's application]
At my server I attach the access token to the user and use that to give permissions to that user in my application. (using the facebook connect to authenticate users)
My application is secured, and the access done is also authenticated regardless of facebook, BUT! in this flow, the a weak link I identified is that I cannot authenticate for sure that the access token I got was signed for my application - I do not like it because I cache the tokens for offline use, I want to be 100% sure they are for my application, with my permissions.
So what will be the (best) way to authenticate that the token I got is related to my application (for relation to user, I use the token to access /me and see which user this token is for)
I do not need to decrypt the token (i guess its some sort of AES), I am just looking for an endpoint that will tell me the token matched my application id.
(EDIT: Using the C# SDK, if it matters.. But a graph/rest call to give that info is just as good as well :) )
https://graph.facebook.com/app/?access_token=[user_access_token]
This will return the app this token was generated for, you can compare that against your app's id.
The official graph endpoint for inspecting access tokens is:
GET graph.facebook.com/debug_token?
input_token=[user_access_token]&
access_token=[app_token_or_admin_token]
Example response:
{
"data": {
"app_id": 138483919580948,
"application": "Social Cafe",
"expires_at": 1352419328,
"is_valid": true,
"issued_at": 1347235328,
"metadata": {
"sso": "iphone-safari"
},
"scopes": [
"email",
"publish_actions"
],
"user_id": 1207059
}
}
app_token_or_admin_token can be obtained using the Graph API call:
GET graph.facebook.com/oauth/access_token?
client_id={app-id}
&client_secret={app-secret}
&grant_type=client_credentials
The debug_token endpoint will fail if that user_access_token doesn't belong to the app that generated the app_token_or_admin_token.
Relevant facebook documentation:
Inspecting access tokens:
https://developers.facebook.com/docs/facebook-login/login-flow-for-web-no-jssdk/#checktoken
App Tokens:
https://developers.facebook.com/docs/facebook-login/access-tokens/#apptokens
A documented way to ensure this is to use appsecret_proof.
GET graph.facebook.com/v2.5/me?access_token=[TOKEN]&appsecret_proof=[PROOF]
This verifies not only that it is a valid token, but also that the token belongs to the app. It also gets you user data in one go.
You can derive PROOF above in C# using this (from here):
public static string ComputeHmacSha256Hash(string valueToHash, string key)
{
byte[] keyBytes = Encoding.ASCII.GetBytes(key);
byte[] valueBytes = Encoding.ASCII.GetBytes(valueToHash);
byte[] tokenBytes = new HMACSHA256(keyBytes).ComputeHash(valueBytes);
valueBytes = null;
keyBytes = null;
StringBuilder token = new StringBuilder();
foreach (byte b in tokenBytes)
{
token.AppendFormat("{0:x2}", b);
}
tokenBytes = null;
return token.ToString();
}
ComputeHmacSha256Hash(accessToken, appSecret);
Why not to use official way of doing things? Here's the request from FB's own video about security.
Request:
https://graph.facebook.com/debug_token?input_token={token-to-check}&access_token={app_id}|{app_secret}
Response:
"data": { "app_id": {token-app-id}, "user_id": {token-user-id}, ... }
Link to an official video: https://www.facebook.com/FacebookforDevelopers/videos/10152795636318553/
I made a screenshot so that time is visible, and you can find more info if you are interested.