I have followed the instructions for installing a password client for Laravel passport exactly as written in the Laravel docs and with default Laravel 6.0 composer versions of guzzle, etc. I have done the install on an existing project and as a clean install, both on local dev environment and live server, and every time I try to post to the example.com/oauth/token route, I am greeted with a crazy Guzzle error that seems to have no previous search history on the internet. The error is (summarized):
GuzzleHttp\Exception\ServerException
/var/task/vendor/guzzlehttp/guzzle/src/Exception/RequestException.php:113
"Return value of Zend\\Diactoros\\normalizeServer() must be of the type array, none returned"
I am running php 7.3 in all environments, but tried php 7.2 and 7.1 and got the same result. I'm running Laravel Valet locally, and have never seen anything like this on any other project. I am also running a staging server with Laravel Vapor, and I get the exact same error. My guzzle request is almost exactly the same as Taylor Otwell's example in the Laravel docs, and looks like this:
$http = new \GuzzleHttp\Client;
$response = $http->post(env('API_TOKEN_URL'), [
'form_params' => [
'grant_type' => 'password',
'client_id' => env('PASSPORT_CLIENT_ID'),
'client_secret' => env('PASSPORT_CLIENT_SECRET'),
'username' => $request['username'],
'password' => $request['password'],
],
]);
return json_decode((string) $response->getBody(), true);
I have data dumped all variables to verify that the username, password, client_id and client_secret are all accurate. It doesn't seem to be an authentication issue at all, but some issue with Guzzle passing proper server headers. I have no idea how to fix, as there is no previous record of this issue that I could find anywhere else on the internet. Any ideas???
if someone face this issue just update the package name: laminas/laminas-diactoros to latest version such as 2.2.2 by running
composer require laminas/laminas-diactoros
the problem comes from
normalize_server.legacy.php
its does not return anything.
Related
I use Guzzle in my Laravel 7 project and XAMPP 7.4.5, I'm trying to make a GET to my local API localhost/events_platforma/view/users It works fine but when I'm trying to make a POST request to https://localhost/events_platforma/register it fails and gives out that cURL error and My API Are on SLIM.
I have added this file
curl.cainfo = curl.cainfo="C:\xampp\php\extras\ssl\cacert.pem"
But still, give out an error
The quick solution for localhost is to turn off the certificate verification using options in guzzle of verify as false.
A quick small example below
use GuzzleHttp\Client;
$client = new Client([
'base_uri' => 'http://exmaple.org'
]);
$client->request('GET', '/', ['verify' => false]);
If you are using Http-client provided by laravel you can add guzzle options like this,
$response = Http::withOptions([
'verify' => false,
])->get('http://example.org/');
NOTE:
Though even guzzle suggests to not using it, but if you are testing your own apis it can work.
Though you can simple add your certificates as per request just by providing path.
Mozilla provides a commonly used CA bundle which can be downloaded here (provided by the maintainer of cURL).
// Use a custom SSL certificate on disk.
$client->request('GET', '/', ['verify' => '/path/to/cacert.pem']);
Read more about certificates from https://curl.se/docs/sslcerts.html .
Read more about verify from guzzle docs verify
I'm using Laravel 5.4
web.php
Route::delete('claim/{id?}', 'ClaimController#claimRemove');
myTest.php
$response = $this->json('delete', 'claim', [
'id' => $id
]);
When i run phpunit, i'm getting the
MethodNotAllowedHttpException
BUT if I run it via Postman or phpstorm rest client - it works fine, so the reason is somewhere in $this->json method. I also tried $this->call.
If I switch delete method to post in web.php and in my test file - test is passing well.
So, question is - why it's not working with DELETE method or how to test DELETE calls?:)
Thanks.
Seems like it was version issue. Didn't modify anything, but after two weeks just composer update and test passed fine.
If composer update simply solved your problem it sounds like that your route cache has not been updated with your latest route changes. Both composer update and composer install include normally a list of artisan commands such as route:clear as they are specified indirectly in the composer.json file when optimize is used.
Secondly, use this form below due to the fact the id is part of your route otherwise it will hit the route without an id. However it will also be acceptable because you have made the parameter optional.
$response = $this->json('delete', 'claim/' . $id, []);
The way you have defined your routes, the ID needs to be passed in the URL.
Replace
$response = $this->json('delete', 'claim', [
'id' => $id
]);
with
$response = $this->json('delete', 'claim/' . $id, []);
I'm trying to use mailgun to send notifications to users. I have verified domain, but even though my config\services.php file looks like this:
'mailgun' => [
'domain' => 'mg.mydomain.biz',
'secret' => 'key-3223423423n423j42jklkj23l',
],
all the emails go through sandbox domain.
php artisan config:clear does not help. It is the same with development and production environment. Also I need to point out that emails come from #maydomain.biz when config\mail.php clearly states 'from' => ['address' => 'no-reply#mg.mydomain.biz', 'name' => 'Sender'],. I channged it hours ago and cleared config cache un restarted web server since.
sudo service supervisor restart solved the problem.
Using 5.0
in config/session.php I have set 'domain' => '.example.com' but it is not working. I cannot persist a session on even one domain like this.
My site has many subdomains:
vancouver.example.com
newyork.example.com
etc... they are hosted on the same server and are the same Laravel app (share the same storage directory)
I login with the correct credentials, upon which the app redirects to another page on the site, and I have no session at that point. var_dump(Auth::user()) shows null even though I logged in with the correct credentials.
storage/framework/sessions shows 14 different files there, they are all for me and I cleared them out before I started testing this.
I'll attach my AuthController#postLogin method below, which works fine if session.php 'domain' => null
public function postLogin(Request $request)
{
$this->validate($request, [
'email' => 'required|email', 'password' => 'required',
]);
$credentials = $request->only('email', 'password');
if ($this->auth->attempt($credentials, $request->has('remember'))) {
Session::flash('message', 'You are now logged in.');
Session::flash('status', 'success');
if (str_contains($_SERVER['HTTP_REFERER'], '?goto=')) {
$params = explode('?', $_SERVER['HTTP_REFERER'])[1];
$target = explode('=', $params)[1];
} else {
$target = '/';
}
return redirect($target);
}
return redirect($this->loginPath())
->withInput($request->only('email', 'remember'))
->withErrors([
'email' => $this->getFailedLoginMessage(),
]);
}
Figured it out. Update domain => '.example.com' in session.php and clear the cookies for the site in question.
#gadss
you need to add session table like this
php artisan session:table
composer dump-autoload
php artisan migrate
and change .env to
SESSION_DRIVER=database
also modify config/session.php
'driver' => env('SESSION_DRIVER', 'database') and
'domain' => '.yourdomain.com'
after that clear your browser's cache and cookies.
You'll need to update the session configuration to persist the session in domain-wide including subdomains. Follow the steps given below.
Go to config/session.php and update the domain with prefix . as config => '.your-domain.com'.
Then clear your application cache, Open the Chrome DevTool and Go to Application > Application > Clear Storage. You'll need to clear out the previous cookies also.
run artisan command php artisan config:cache or php artisan config:clear to drop previously cached laravel application configs.
If you are using database as the session driver, You need to create a session table for that. run command php artisan session:table to generate the session table migration and then migrate it using php artisan migrate. Then perform the three steps given above.
With Laravel 8 it becomes more simplier :
Add SESSION_DOMAIN to your .env file :
SESSION_DOMAIN=.yourdomain.tld
Clear configuration cache :
php artisan config:cache
Delete your browser sessions cookies, then session become shared between all your subdomains.
In my case I used to AutoLogin user to subdomain once account is created on www. domain. Worked fine.
Have you tried storing the sessions in the database, memcached, or redis instead of in files? I had a similar situation to yours and storing sessions in the database solved the issue for me.
For some reason Laravel's session driver doesn't handle cross domain sessions correctly when using the file driver.
If someone still gets the problem with subdomain cookie. Try to change Session Cookie Name in config/session.php
If someone needs to sync session in subdomains with different laravel application sharing same database
Follow all the instructions of #Kiran Maniya
Then you have to keep same application name in order to get same session name. Or just change the cookie config in config/session.php
You can hardcode it if keeping same name is not possible.
'cookie' => env(
'SESSION_COOKIE',
Str::slug(env('APP_NAME', 'laravel'), '_').'_session'
)
to something like:
'cookie' => env(
'SESSION_COOKIE',
'session_sharing_application_session'
)
Everything in my Magento store is working ok, except for a route I created that calls the API:
$proxy = new SoapClient('SOAPCLIENTURL');
$sessionId = $proxy->login('USERNAME', 'PASSWORD');
$proxy->customerCustomerCreate($sessionId, array(
'email' => $email,
'firstname' => '',
'lastname' => '',
'password' => $password,
'website_id' => 7,
'store_id' => 7
));
When I comment out these lines, the route works fine. Any ideas why this 503s the page and how to fix it?
The code block that's causing your problem is a request to an external API that could fail for numerous reasons. The way you'd fix this is to monitor your server and Magento error and exception logs for errors, take a look at the error, and then fix the problem (or post the specific error to a site like this and ask for help).
You could also try running the above code snippet outside of a Magento context (in a stand along script) and see what sort of error you get.
If errors aren't showing up then you need to research how to setup your system for proper error handling.
Also, if you're setting up a Magento route and making an API call into the same system, there's no reason to use the SOAP or XML-RPC layer. Each Magento API section has a corresponding PHP object that contains all the logic. The above method is equivalent to
Mage::getModel('customer/customer_api_v2')->create(array(...));
With the real PHP class being at
app/code/core/Mage/Customer/Model/Customer/Api/V2.php
and the create method defined at
app/code/core/Mage/Customer/Model/Customer/Api.php