The spark-pac4j
project is an easy and powerful security library for Sparkjava web applications which supports authentication and authorization, but also logout and advanced features like session fixation and CSRF protection.
It's based on Java 8, Spark 2.7 and on the pac4j security engine v2. It's available under the Apache 2 license.
- A client represents an authentication mechanism. It performs the login process and returns a user profile. An indirect client is for UI authentication while a direct client is for web services authentication:
▸ OAuth - SAML - CAS - OpenID Connect - HTTP - OpenID - Google App Engine - LDAP - SQL - JWT - MongoDB - Stormpath - IP address
- An authorizer is meant to check authorizations on the authenticated user profile(s) or on the current web context:
▸ Roles / permissions - Anonymous / remember-me / (fully) authenticated - Profile type, attribute - CORS - CSRF - Security headers - IP address, HTTP method
-
The
SecurityFilter
protects an url by checking that the user is authenticated and that the authorizations are valid, according to the clients and authorizers configuration. If the user is not authenticated, it performs authentication for direct clients or starts the login process for indirect clients -
The
CallbackRoute
finishes the login process for an indirect client -
The
LogoutRoute
logs out the user from the application.
Just follow these easy steps to secure your Sparkjava web application:
You need to add a dependency on:
- the
spark-pac4j
library (groupId: org.pac4j, version: 2.3.0) - the appropriate
pac4j
submodules (groupId: org.pac4j, version: 2.2.1):pac4j-oauth
for OAuth support (Facebook, Twitter...),pac4j-cas
for CAS support,pac4j-ldap
for LDAP authentication, etc.
All released artifacts are available in the Maven central repository.
The configuration (org.pac4j.core.config.Config
) contains all the clients and authorizers required by the application to handle security.
It can be built via a configuration factory (org.pac4j.core.config.ConfigFactory
) for example:
public class DemoConfigFactory implements ConfigFactory {
private final String salt;
private final TemplateEngine templateEngine;
public DemoConfigFactory(final String salt, final TemplateEngine templateEngine) {
this.salt = salt;
this.templateEngine = templateEngine;
}
@Override
public Config build(final Object... parameters) {
final OidcConfiguration oidcConfiguration = new OidcConfiguration();
oidcConfiguration.setClientId("343992089165-sp0l1km383i8cbm2j5nn20kbk5dk8hor.apps.googleusercontent.com");
oidcConfiguration.setSecret("uR3D8ej1kIRPbqAFaxIE3HWh");
oidcConfiguration.setDiscoveryURI("https://accounts.google.com/.well-known/openid-configuration");
oidcConfiguration.setUseNonce(true);
oidcConfiguration.addCustomParam("prompt", "consent");
final OidcClient oidcClient = new OidcClient(oidcConfiguration);
oidcClient.setAuthorizationGenerator((ctx, profile) -> { profile.addRole("ROLE_ADMIN"); return profile; });
final SAML2ClientConfiguration cfg = new SAML2ClientConfiguration("resource:samlKeystore.jks", "pac4j-demo-passwd",
"pac4j-demo-passwd", "resource:metadata-okta.xml");
cfg.setMaximumAuthenticationLifetime(3600);
cfg.setServiceProviderEntityId("http://localhost:8080/callback?client_name=SAML2Client");
cfg.setServiceProviderMetadataPath("sp-metadata.xml");
final SAML2Client saml2Client = new SAML2Client(cfg);
final FacebookClient facebookClient = new FacebookClient("145278422258960", "be21409ba8f39b5dae2a7de525484da8");
final TwitterClient twitterClient = new TwitterClient("CoxUiYwQOSFDReZYdjigBA", "2kAzunH5Btc4gRSaMr7D7MkyoJ5u1VzbOOzE8rBofs");
final FormClient formClient = new FormClient("http://localhost:8080/loginForm", new SimpleTestUsernamePasswordAuthenticator());
final IndirectBasicAuthClient indirectBasicAuthClient = new IndirectBasicAuthClient(new SimpleTestUsernamePasswordAuthenticator());
final CasConfiguration casConfiguration = new CasConfiguration("https://casserverpac4j.herokuapp.com/login");
final CasClient casClient = new CasClient(casConfiguration);
ParameterClient parameterClient = new ParameterClient("token", new JwtAuthenticator(new SecretSignatureConfiguration(salt)));
parameterClient.setSupportGetRequest(true);
parameterClient.setSupportPostRequest(false);
final DirectBasicAuthClient directBasicAuthClient = new DirectBasicAuthClient(new SimpleTestUsernamePasswordAuthenticator());
final HeaderClient headerClient = new HeaderClient("Authorization", (credentials, ctx) -> {
final String token = ((TokenCredentials) credentials).getToken();
if (CommonHelper.isNotBlank(token)) {
final CommonProfile profile = new CommonProfile();
profile.setId(token);
credentials.setUserProfile(profile);
}
});
final Clients clients = new Clients("http://localhost:8080/callback", oidcClient, saml2Client, facebookClient,
twitterClient, formClient, indirectBasicAuthClient, casClient, parameterClient, directBasicAuthClient, new AnonymousClient(),
headerClient);
final Config config = new Config(clients);
config.addAuthorizer("admin", new RequireAnyRoleAuthorizer("ROLE_ADMIN"));
config.addAuthorizer("custom", new CustomAuthorizer());
config.addMatcher("excludedPath", new PathMatcher().excludeRegex("^/facebook/notprotected$"));
config.setHttpActionAdapter(new DemoHttpActionAdapter(templateEngine));
return config;
}
}
http://localhost:8080/callback
is the url of the callback endpoint, which is only necessary for indirect clients.
Notice that you can define specific matchers via the addMatcher(name, Matcher)
method.
You can protect (authentication + authorizations) the urls of your Sparkjava application by using the SecurityFilter
and defining the appropriate mapping. It has the following behaviour:
-
If the HTTP request matches the
matchers
configuration (or nomatchers
are defined), the security is applied. Otherwise, the user is automatically granted access. -
First, if the user is not authenticated (no profile) and if some clients have been defined in the
clients
parameter, a login is tried for the direct clients. -
Then, if the user has a profile, authorizations are checked according to the
authorizers
configuration. If the authorizations are valid, the user is granted access. Otherwise, a 403 error page is displayed. -
Finally, if the user is still not authenticated (no profile), he is redirected to the appropriate identity provider if the first defined client is an indirect one in the
clients
configuration. Otherwise, a 401 error page is displayed.
The following parameters are available:
-
config
: the previous security configuration -
clients
(optional): the list of client names (separated by commas) used for authentication:
- in all cases, this filter requires the user to be authenticated. Thus, if the
clients
is blank or not defined, the user must have been previously authenticated - if the
client_name
request parameter is provided, only this client (if it exists in theclients
) is selected.
authorizers
(optional): the list of authorizer names (separated by commas) used to check authorizations:
- if the
authorizers
is blank or not defined, no authorization is checked - the following authorizers are available by default (without defining them in the configuration):
isFullyAuthenticated
to check if the user is authenticated but not remembered,isRemembered
for a remembered user,isAnonymous
to ensure the user is not authenticated,isAuthenticated
to ensure the user is authenticated (not necessary by default unless you use theAnonymousClient
)hsts
to use theStrictTransportSecurityHeader
authorizer,nosniff
forXContentTypeOptionsHeader
,noframe
forXFrameOptionsHeader
,xssprotection
forXSSProtectionHeader
,nocache
forCacheControlHeader
orsecurityHeaders
for the five previous authorizerscsrfToken
to use theCsrfTokenGeneratorAuthorizer
with theDefaultCsrfTokenGenerator
(it generates a CSRF token and saves it as thepac4jCsrfToken
request attribute and in thepac4jCsrfToken
cookie),csrfCheck
to check that this previous token has been sent as thepac4jCsrfToken
header or parameter in a POST request andcsrf
to use both previous authorizers.
-
matchers
(optional): the list of matcher names (separated by commas) that the request must satisfy to check authentication / authorizations -
multiProfile
(optional): it indicates whether multiple authentications (and thus multiple profiles) must be kept at the same time (false
by default).
For example:
before("/facebook", new SecurityFilter(config, "FacebookClient"));
For indirect clients (like Facebook), the user is redirected to an external identity provider for login and then back to the application.
Thus, a callback endpoint is required in the application. It is managed by the CallbackRoute
which has the following behaviour:
-
the credentials are extracted from the current request to fetch the user profile (from the identity provider) which is then saved in the web session
-
finally, the user is redirected back to the originally requested url (or to the
defaultUrl
).
The following parameters are available:
-
config
: the previous security configuration -
defaultUrl
(optional): it's the default url after login if no url was originally requested (/
by default) -
multiProfile
(optional): it indicates whether multiple authentications (and thus multiple profiles) must be kept at the same time (false
by default) -
renewSession
(optional): it indicates whether the web session must be renewed after login, to avoid session hijacking (true
by default).
For example:
CallbackRoute callback = new CallbackRoute(config, null, true);
get("/callback", callback);
post("/callback", callback);
You can get the profile of the authenticated user using profileManager.get(true)
(false
not to use the session, but only the current HTTP request).
You can test if the user is authenticated using profileManager.isAuthenticated()
.
You can get all the profiles of the authenticated user (if ever multiple ones are kept) using profileManager.getAll(true)
.
Example:
WebContext context = new SparkWebContext(request, response);
ProfileManager manager = new ProfileManager(context);
Optional<CommonProfile> profile = manager.get(true);
The retrieved profile is at least a CommonProfile
, from which you can retrieve the most common attributes that all profiles share. But you can also cast the user profile to the appropriate profile according to the provider used for authentication. For example, after a Facebook authentication:
FacebookProfile facebookProfile = (FacebookProfile) commonProfile;
The LogoutFilter
can handle:
- the local logout by removing the pac4j profiles from the session (it can be used for the front-channel logout from the identity provider in case of a central logout)
- the central logout by calling the identity provider logout endpoint.
It has the following behaviour:
-
If the
localLogout
property istrue
, the pac4j profiles are removed from the web session (and the web session is destroyed if thedestroySession
property istrue
) -
A post logout action is computed as the redirection to the
url
request parameter if it matches thelogoutUrlPattern
or to thedefaultUrl
if it is defined or as a blank page otherwise -
If the
centralLogout
property istrue
, the user is redirected to the identity provider for a central logout and then optionally to the post logout redirection URL (if it's supported by the identity provider and if it's an absolute URL). If no central logout is defined, the post logout action is performed directly.
The following parameters are available:
-
defaultUrl
(optional): the default logout url if nourl
request parameter is provided or if theurl
does not match thelogoutUrlPattern
(not defined by default) -
logoutUrlPattern
(optional): the logout url pattern that theurl
parameter must match (only relative urls are allowed by default) -
localLogout
(optional): whether a local logout must be performed (true
by default) -
destroySession
(optional): whether we must destroy the web session during the local logout (false
by default) -
centralLogout
(optional): whether a central logout must be performed (false
by default).
Example:
get("/logout", new LogoutRoute(config, "/?defaulturlafterlogout"));
The ApplicationLogoutRoute
has been renamed as LogoutRoute
with more options.
Use context.getSparkResponse().body()
instead of context.getBody()
.
The RequiresAuthenticationFilter
is now named SecurityFilter
.
The ApplicationLogoutRoute
behaviour has slightly changed: even without any url
request parameter, the user will be redirected to the defaultUrl
if it has been defined.
Authorizations are now handled by the library so the ClientFactory
can now longer be used and is replaced by a ConfigFactory
which builds a Config
which gathers clients (for authentication) and authorizers (for authorizations).
The application logout process can be managed with the ApplicationLogoutFilter
.
The demo webapp: spark-pac4j-demo is available for tests and implements many authentication mechanisms: Facebook, Twitter, form, basic auth, CAS, SAML, OpenID Connect, JWT...
See the release notes. Learn more by browsing the spark-pac4j Javadoc and the pac4j Javadoc.
If you have any question, please use the following mailing lists:
The version 2.3.1-SNAPSHOT is under development.
Maven artifacts are built via Travis: and available in the Sonatype snapshots repository. This repository must be added in the Maven pom.xml file for example:
<repositories>
<repository>
<id>sonatype-nexus-snapshots</id>
<name>Sonatype Nexus Snapshots</name>
<url>https://oss.sonatype.org/content/repositories/snapshots</url>
<releases>
<enabled>false</enabled>
</releases>
<snapshots>
<enabled>true</enabled>
</snapshots>
</repository>
</repositories>