Clients: (v6.1)

A Client represents an authentication mechanism. It performs the login process and returns (if successful) a user profile. Many clients are available for the:

• OAuth protocol
• SAML protocol
• CAS protocol
• OpenID Connect protocol
• HTTP protocol
• OpenID protocol
• Google App Engine support
• Kerberos (SPNEGO Negotiate) protocol

While most clients are self-sufficient, the HTTP clients require defining an Authenticator to handle the credentials validation.

Clients (like Authorizers) are generally defined in a security configuration.

Each client has a name which is by default the class name (like FacebookClient), but it can be explicitly set to another value with the setName method.

Understand the main features:

• Direct vs indirect clients
• Compute roles and permissions
• The callback URL
• Profile definition
• AJAX requests
• The Client methods

1) Direct vs indirect clients

Clients are of two kinds: direct for web services authentication and indirect for UI authentication. Here are their behaviors and differences:

2) Compute roles and permissions

To compute the appropriate roles and permissions of the authenticated user profile, you need to define an AuthorizationGenerator and attach it to the client.

Example:

AuthorizationGenerator authGen = (ctx, profile) -> {
  String roles = profile.getAttribute("roles");
  for (String role: roles.split(",")) {
    profile.addRole(role);
  }
  return profile;
};
client.addAuthorizationGenerator(authGen);

And you can add as many authorization generators as you want using the addAuthorizationGenerator method or a list of authorization generators using the setAuthorizationGenerators method.

In fact, the AuthorizationGenerator component can be used to do more than just computing roles and permissions, like defining the remember-me nature of a profile based on a remember-me checkbox of a form (see: RememberMeAuthorizationGenerator).

3) The callback URL

For an indirect client, you must define the callback URL which will be used in the login process: after a successful login, the identity provider will redirect the user back to the application on the callback URL.

On this callback URL, the “callback filter” must be defined to finish the login process.

As the callback URL can be shared between multiple clients, the callback URL can hold the information of the client (to be able to distinguish between the different clients), as a query parameter or as a path parameter.

Example:

FacebookClient facebookClient = new FacebookClient(fbKey, fbSecret);
TwitterClient twitterClient = new TwitterClient(twKey, twSecret);
Config config = new Config("http://localhost:8080/callback", facebookClient, twitterClient);

In that case, the callback URL of the FacebookClient is http://localhost:8080/callback?client_name=FacebookClient and the callback URL of the TwitterClient is http://localhost:8080/callback?client_name=TwitterClient.

This is the callback URL you must define on the identity provider side.

This happens because the default CallbackUrlResolver of the clients is the QueryParameterCallbackUrlResolver.

You can change the client_name parameter using the setClientNameParameter method of the QueryParameterCallbackUrlResolver.

But you can also use the PathParameterCallbackUrlResolver, which adds the client name as a path parameter.

Example:

OidcConfiguration configuration = new OidcConfiguration();
configuration.setClientId("788339d7-1c44-4732-97c9-134cb201f01f");
configuration.setSecret("we/31zi+JYa7zOugO4TbSw0hzn+hv2wmENO9AS3T84s=");
configuration.setDiscoveryURI("https://login.microsoftonline.com/38c46e5a-21f0-46e5-940d-3ca06fd1a330/.well-known/openid-configuration");
AzureAdClient azureAdClient = new AzureAdClient(configuration);
client.setCallbackUrlResolver(new PathParameterCallbackUrlResolver());
Clients clients = new Clients("http://localhost:8080/callback", azureAdClient);
Config config = new Config(clients);

In that case, the callback URL will be http://localhost:8080/callback/AzureAdClient for the AzureAdClient.

You may even use the NoParameterCallbackUrlResolver which left the callback URL untouched. In that case, no parameter will be added to the callback URL and no client will be retrieved on the callback endpoint. You will be forced to define a “default client” at the CallbackLogic level.

Example:

defaultCallbackLogic.setClient("FacebookClient");

The CallbackUrlResolver relies on a UrlResolver to complement the URL according to the current web context. The UrlResolver can be retrieved via the getUrlResolver() method of the client.

You can use the DefaultUrlResolver and handle relative URLs by using: defaultUrlResolver.setCompleteRelativeUrl(true). Or provide your own UrlResolver using the setUrlResolver method.

4) Profile definition

Most clients rely on the Authenticator and ProfileCreator components to validate credentials and create the user profile.

At the end of the login process, the returned user profile is created by the (internal) Authenticator or ProfileCreator, which holds a profile definition.

This profile definition can be overridden using the setProfileDefinition method.

5) AJAX requests

For an indirect client, if the user tries to access a protected URL, he will be redirected to the identity provider for login.

Though, if the incoming HTTP request is an AJAX one, no redirection will be performed and a 401 error page will be returned.

The HTTP request is considered to be an AJAX one if the value of the X-Requested-With header is XMLHttpRequest or if the is_ajax_request parameter or header is true. This is the behaviour of the DefaultAjaxRequestResolver.

But you can provide your own AjaxRequestResolver with: client.setAjaxRequestResolver(myAjaxRequestResolver);.

6) The Client methods

The Client interface has the following methods:

Clients are generally populated with default sub-components: RedirectActionBuilder, CredentialsExtractor, ProfileCreator, LogoutActionBuilder and Authenticator, except for HTTP clients where the Authenticator must be defined. Sub-components can of course be changed for various customizations.