I'm building a page for a Flutter web application using Firebase Auth where the user can change his password. For security reasons, I want the user having to enter his recent password before being able to set a new password.
To achieve this, I tried to use the reauthenticateWithCredential()-method from Firebase.
Problem is, this only seems to work if the user hasn't signed in recently. If, on the other hand, he has just signed in, he can change his password without giving the right credentials (or even without giving any credentials at all).
So is there a possibility to prevent the user from changing his password without providing his correct recent password, no matter if he just signed in half a minute ago? Did I overlook something?
As far as I know it isn't possible to get the password from Firebase, so the only two solutions I can think of atm are
To sign out and re-sign-in the user, given he has entered the correct credentials, and then change the password, or
pass his password from the login to a variable on the change-password-screen, just in case it should be needed...
However, both methods don't seem to be appropriate for practical and/or security reasons.
We have a web application in .NET Core in which you can reset your password (such functionality much wow) using the ResetPasswordAsync(user, token, password) method of the UserManager class. We have the default settings about password validation (should be more than 6 symbols long, containing a digit, upper and lower character, a special symbol, etc). Everything works fine until you try to reset your password to one containing a + symbol. In that case the method returns a success but then you can't login with that password.
For example, if we have the password "Someth1ng_" it works, but if we have "Someth1ng+" it doesn't work. There are problems with other special symbols but we haven't taken the time to find out exactly which ones (the following are ok: *=_).
We've already checked whether the problem is with parameter passing from the client to the server and it's not - on the server the password is received as it was sent.
Could the problem be in the ResetPasswordAsync method? Has anybody had a similar issue?
I am having a similar issue, in my case I found that some passwords may affect login as ssnaky described before, in my case the password failed when I entered durga%118 at first I thought that the char % was the responsable but then I entered the new password durga%%%% and it worked fine, so I asume that maybe is something within the ResetPasswordAsync method.
NOTE: I found my problem, it was an encoding issue at the login section.
When the user clicks the Confirm Email link in the email they receive, the url directs the user to the ConfirmEmailAsync method in the Account Controller.
This method then returns the result of the call to the Email Token Provider.
var result = await _userManager.ConfirmEmailAsync(user, code);
I have captured the generated email token when the user registers and the token that is passed in the code variable above and they are identical.
I have tried Base64 encoding and decoding with no success, although I believe this is done automatically by the asp-net-core-identity-token-providers.
Any ideas ?
Generated tokens through identity are tied to the SecurityStamp on the user. If the stamp is updated for any reason, all previously sent tokens are invalidated. Changing or setting a password would cause this to happen.
This could be an order of operations problem. Do you by chance save the register password after you send the email to validate the email they provided?
For instance, this will fail:
Send email with token
Save the register info like the password
Try to use the token.
Instead, this should work:
Save the register info
Send email with token
Try to use token
That scenario is just an example as I can't be sure that is the way your code is setup. That said, if the tokens are the same, it is likely the securitystamp that is the problem. Keep tabs on it in your database throughout the process and see if something is updating it.
I'm using the Instagram API and want to get the access_token in order to throw api requests over my own account. When I try to follow the first step and get the authorization code programmatically using RestTemplate I can't get it work.
String AUTHORIZE_URL = "https://api.instagram.com/oauth/authorize/?client_id=<CLIENT_ID>&redirect_uri=<REDIRECT_URI>&response_type=code";
String url = String.format(AUTHORIZE_URL, clientId, redirectUri);
String o = restTemplate.getForObject(url, String.class);
The response is the html code of the login page because Instagram requires the user to be logged in to check if the app is authorized (of course it is, since the app an the user belongs to my own account).
How can I authenticate before throwing that request so they return the code to my redirectUri and not complain about login?
Note: I tried simulating the request to their login form but it returned a 403 Forbidden.
NOTE: I already got a valid access_token, manually generated, and it works perfectly but I want to implement also a process to re-generate a new access_token automatically since they may invalidate it at any time in the future.
Even though our access tokens do not specify an expiration time, your app should handle the case that either the user revokes access, or Instagram expires the token after some period of time. If the token is no longer valid, API responses will contain an “error_type=OAuthAccessTokenError”. In this case you will need to re-authenticate the user to obtain a new valid token.
In other words: do not assume your access_token is valid forever.
Instagram is upgrading their APIs and the flows. Earlier we needed access token to bypass forced login screen. Since yesterday, they have removed that.
Now if you call this code, it will check if you are already logged in or not. If so, it will call the AUTHORIZE_URL of yours and will send a response code. The format will be either:
On success validation - http://your-redirect-uri?code=CODE
On error - http://your-redirect-uri?error=access_denied&error_reason=user_denied&error_description=The+user+denied+your+request
Now what I'm doing is I'm directly calling the above URL of yours every time. Now if the user is logged in, a response code will be sent to you, else user will be asked to login and validate your app and then the code will be sent. Technically, you are eliminating the possibility of the error case! So no need of overhead of storing access token in your database or verifying its validity.
Just try and check now what happens.
PS: If you want to check API behavior, simply type the URL on the browser and check what it returns! It helped me a lot while coding and debugging! :)
What are the possible reasons that can cause token to become expired (besides having the user un-authorising the app)?
My problem is that I have an app with several thousands of users, all API communication works perfectly but for some users I am getting the invalid or expired token error, my initial though was that they are users who canceled the authentication to the app but I've contacted some of them and they haven't revoked the access.
Any ideas what other issues can cause that error?
Check the integrity of an access token at any time by calling the GET account/verify_credentials while using that access token.
Its mentioned and by research I came to know that:
Your access token will be invalid if a user explicitly rejects your
application from their settings or if a Twitter admin suspends your
application. If your application is suspended there will be a note on
your application page saying that it has been suspended.
Why is my twitter oauth access token invalid / expired ?
Check this post: invalid / expired access tokens.
There is one post in google groups that says:
You don't get a second chance, and this is by design. OAuth requests
have a unique signature; once a particular request is submitted, it
can't be submitted again.
If they enter the pin correctly, all is well, you get an access token.
If they enter the pin wrong, you get 401 Unauthorized - which is
expected.
But if they then try again to enter the pin, even the correct pin
shows as unauthorized.
Check this link for the above reference.
Some suggestions by twitter employee for the same problem:
I guess there are two things I would suggest at this point: 1.) Go to
your application settings and use the "Reset keys" tab to reset your
consumer key and secret, then update those values in the app and
verify that you still see the same behavior. 2.) Try passing
oauth_callback in your request_token call. Honestly I don't think this
will make a difference, but I want to try and be as rigorous as I can
here.
Also check this discussion saying:
You need to use the oauth_token and oauth_token_secret returned from
the oauth/access_token call instead of the one in your app's settings
in dev.twitter.com
I was getting same error then I changed (access_token) to (access_token_key) and it worked for me.
I hope it will help someone.
In addition to the comments everyone else has made, sometimes the twitter api will return a "invalid token" error when the token isn't the problem. I've noticed it the most when I've built a request string that doesn't parse correctly. For example, once I was getting that error when I was passing in screen_name's that had symbols that weren't URI-encodable. I also have gotten it when I passed in empty values like this (where the cursor is empty):
https://api.twitter.com/1/followers.json?cursor=&screen_name=whatevah
Could you give us the specifics of the calls that are returning this error?
Have you confirmed that the tokens worked at one time? In an OAuth system I worked on, there was an error in how tokens were securely stored and retrieved that caused a small percentage of them to become corrupted. If you can confirm that the tokens worked in the past, that's a good first step.
When you retrieve the tokens from storage, are they unchanged? Is it possible for them to get corrupted with the way you're managing them?
Put some logging in place to keep track of when tokens work and fail. Does a token ever start working again after it has failed once? If you fail to use a token for 30 days, does it expire? With a detailed log, you can start identifying the expired tokens and look for patterns in use to point to what might cause them to expire.
Be sure to explore other possibilities as well. How do users revoke tokens in Twitter? Is it easy to accidentally do that? For users with failed tokens, do they have other authorized apps that have stopped working as well?
First of all nice question Ran.
I want to ask you that have you gone through Twitter developers??
Sometimes it becomes ambiguous that which token to use since Twitter provides two pairs of tokens and the library.One of them is a secret key.
You need to select those token which starts with your Twitter ID followed by a hyphen.
Now your question is this error happens with some of yours users. So here is the answer that an app itself finds ambiguous to choose the token.
Though I might not be completely right, but I recommend you to try this solution at least once.
It might be possible these users have not revoked access. But in my experience an access token can also get expired after the user (in test cases: me) changed his/her password.
When the user does that, you can no longer use the REST API of stream API on that user's scope. Please adapt your application to handle with that situation. Revoke the user's session, so when he comes back to your application, he/she can be redirected to Twitter again to start a new OAuth access token process. Or send him/her an e-mail to kindly ask to reconnect. Vimeo/Windows/... are some of the people handling expired tokens with e-mails.
Have fun!
My God's answer is correct but I will share my answer from another question explaining how it could be your computer's clock:
If your OAuth flow was working one day and failing the next, check your computer's clock. I was running a Vagrant box that somehow had its time set to the day before, which caused the Twitter API to return {"code":89,"message":"Invalid or expired token."}. This may also appear as 401 timestamp out of bounds. You can use this command to update your clock in Ubuntu:
sudo ntpdate time.nist.gov
Alternative method if ntpdate isn't available on your system:
sudo date -s "$(wget -qSO- --max-redirect=0 google.com 2>&1 | grep Date: | cut -d' ' -f5-8)Z"
if your Access Token=738629462149844993-FcWHjfcucCLGEosyGGQ38qI******iC then don't forget to mention hyphen (-) followed by your USERID.
May be this will be helpful for you.I faced the same problem.
Please find the below piece of code snippet
$code = $tmhOAuth->user_request(array(
'method' => 'POST',
'url' => $tmhOAuth->url('oauth/access_token', ''),
'params' => array(
'oauth_verifier' => trim($params['oauth_verifier']),
)
));
if ($code == 200) {
$oauth_creds = $tmhOAuth->extract_params($tmhOAuth->response['response']);
// echo '<pre>';print_r($oauth_creds);exit;
$tmhOAuth->reconfigure(array_merge($tmhOAuth->config, array(
'token' => $oauth_creds['oauth_token'],
'secret' => $oauth_creds['oauth_token_secret'],
)));
$code = $tmhOAuth->user_request(array(
'url' => $tmhOAuth->url('1.1/account/verify_credentials')
));
}
The error invalid or expired token can be linked with the fact that one is not paying.
Without paying one will only be able to create the dev environment (sandbox).
As I have answered here:
Counts is only available to paid premium accounts, and one needs to pay for premium access.
Use this link to Apply for access.
Try to regenerate the keys again and save them properly.
For me, it happened because after regenerating one of the keys I did not update other keys. Therefore removed and regenerated all 4 keys again (CONSUMER_KEY, CONSUMER_SECRET, ACCESS_KEY, ACCESS_SECRET). And tried to execute it again and it worked this time.