Is protected-views flawed in wildfly 8.x? - security

In JSF 2.2 new <f:viewAction> component was introduced and more importantly along with it a way to protect pages, that leverage this functionality, from CSRF attacks.
The feature is mentioned in the JSF 2.2 specification:
Call ViewHandler.getProtectedViewsUnmodifiable() to determine if the view for this viewId is protected. If not,
assume the requested view is not protected and take no additional view protection steps. Obtain the value of the
value of the request parameter whose name is given by the value of
ResponseStateManager.NON_POSTBACK_VIEW_TOKEN_PARAM. If there is no value, throw
ProtectedViewException. If the value is present, compare it to the return from
ResponseStateManager.getCryptographicallyStrongTokenFromSession(). If the values do not match, throw
ProtectedViewException.
...
What got me curious is how is the token used by getCryptographicallyStrongTokenFromSession() method generated in Mojarra, since that is the JSF implementation we use for some of our applications.
I was quite surprised to find out that the implementation in versions prior 2.2.11 looks like this:
private String createCryptographicallyStrongToken() {
// PENDING: http://java.net/jira/browse/JAVASERVERFACES-2204
String result = "" + System.currentTimeMillis();
return result;
}
Since version 2.2.11 it seems to be implemented more robustly with the currentTimeMillis actually being encrypted using AES.
Since the non-encrypting versions of this implementation made their way into final releases of Wildfly 8.x server (8.0.0.Final, 8.1.0.Final, 8.2.0.Final and 8.2.1.Final) does this mean that applications deployed on these servers and relying on <protected-views> functionality are vulnerable?

Related

Problems de-serializing System.Security.Claims.Claim

I'm implementing an oAuth server and need to store refresh tokens, to do this I have (at the moment) chosen to serialize the tokens into JSON.
While I can see that the JSON includes everything that would be needed to rehydrate, when I de-serialize with token.FromJson() the embedded claims are not being reconstructed correctly.
So far I've considered inheriting from JsonConverter to create a claims converter but don't see a way of adjusting the global JsConfig to utilise it :(
Can any one point me in a good direction?
So...
Walking away from the code and returning did the trick!
Instead of using a JsonConverter you need to utilise a generic version of JsConfig when changing/overriding the behaviour of ServiceStack on a specific class, just stick the following in your services start-up code for example.
JsConfig<Claim>.SerializeFn = claim => string.Format("{0}|{1}", claim.Type, claim.Value);
JsConfig<Claim>.DeSerializeFn = claimDetails =>
{
var values = claimDetails.Split('|');
return new Claim(values[0], values[1]);
};

CDI extension, altering processed type

Using Weld 1.1.13.Final in test with Arquillian....
Let's say I inject into a field something volatile. Something like a property subject to change that I want the bean owning the injection point to receive change events. Thought about creating a CDI extension.
Caught ProcessAnnotatedType event and looking for all fields that have an custom annotation on field injection points:
<T> void pat(#Observes ProcessAnnotatedType<T> event, BeanManager bm) {
final AnnotatedType<T> target = event.getAnnotatedType();
for (AnnotatedField<? super T> field : target.getFields())
if (field.isAnnotationPresent(Value.class)) { // ignore that I don't check #Inject here for the moment
CtClass wrapper = pool.get(target.getJavaClass().getName());
ConstPool cp = wrapper.getClassFile().getConstPool();
CtMethod m = CtNewMethod.make(....)
....
wrapper.addMethod(m);
event.setAnnotatedType(bm.createAnnotatedType(wrapper.toClass()));
}
}
Had even grabbed thereafter all the injection points for fields and replaced the underlying WeldField with a new Field corresponding the "wrapper" type. Otherwise bean validation fails.
But this only works for stuff setup during startup not when for example Arquillian uses the Bean Manager to initialize a class that injects one of my "wraps". Things fail since the Bean Resolver uses the Type as a hash key to find beans.
Basically I don't think I can "mask" a class that is annotated (made into a bean) by the CDI with an extra method to receive custom events. Would have been cool but a Type is a Type (i.e. no idea how to proxy or fake the equals/hashCode).
Got it. Turns out the compute value function (google extension) inside the TypeSafeBeanResolver resolver (at least the CDI Weld implementation) is smart. If I just extend the class:
CtClass wrapper = pool.makeClass(target.getJavaClass().getName()+"Proxy");
wrapper.setSuperclass(pool.get(target.getJavaClass().getName()));
.....
final AnnotatedType<T> other = bm.createAnnotatedType(wrapper
.toClass());
then everything works fine. Tested capturing an event in a bean. Will post the code on a Gist with a comment.

Implementing a Locale provider that works in JSF and JAX-RS

I've been joyfully using omnifaces' Faces.getLocale() to aquire the locale used by the currently logged in user (which in turn gets this from a <f:view> definition). I really like the fallback approach from view to client to system default locale as it fits the requirements for locale selection in my application:
If a user is logged in, use his language preference (obtained from the backend entity)
If no user preference can be found, use the highest ranking language from the Accept-Languages HTTP header
If no locale has been selected by now, use the system default.
Now I've started using JAX-RS (resteasy implementation) and find it quite difficult to write a service that will provide my backend code with the current user's locale.
I can't use Faces.getLocale(), since that requires a FacesContext which isn't present during JAX-RS request processing.
I can't use the #Context SecurityContext annotation in a #Provider (which would give me the user preferred locale) or #Context HttpHeaders (access to the client locale) since JAX-RS only injects those when it uses the provider itself, not when my backend code instantiates the class.
And I don't want to litter my method signatures with Locale parameters, since virtually everything requires a locale to be present.
To have a concrete example: I have a vcard generator that generates little NOTE fields depending on the user's preferred locale. I can both call the vcard generating method via JSF/EL:
<h:commandLink action="#{vcfGenerator.forPerson(person)}"
value="Go" target="_blank" />
And via a REST service:
#GET #Path('person/{id:[1-9][0-9]*}/vcard')
#Produces('text/vcard')
String exportVcard(#PathParam('id') Long personId, #Context HttpHeaders headers) {
VcfGenerator exporter = Component.getInstance(VcfGenerator) as VcfGenerator
Person person = entityManager.find(Person, personId)
if (! person)
return Response.noContent().build()
def locale = headers.acceptableLanguages[0] ?: Locale.ROOT
return exporter.generateVCF(person, locale).toString()
}
This works (VcfGenerator has a set of JSF-only methods that use Faces.getLocale()), but is a pain to maintain. So instead of passing the Locale object, I'd like to say:
Vcard generateVCF(Person person) {
Locale activeLocale = LocaleProvider.instance().getContext(VcfGenerator.class)
ResourceBundle bundle = ResourceBundle.getBundle("messages", activeLocale, new MyControl())
// use bundle to construct the vcard
}
Has anyone done similar work and can share insights?
I know this has been posted a while ago, but as it has not been marked as resolved, here is how I got a workaround working for this specific case:
First I got a custom ResourceBundle working, as #BalusC described here: http://balusc.blogspot.fr/2010/10/internationalization-in-jsf-with-utf-8.html
Then I updated the constructor in order to detect if a FacesContext is currently being in use, from this :
public Text() {
setParent(ResourceBundle.getBundle(BUNDLE_NAME,
FacesContext.getCurrentInstance().getViewRoot().getLocale(), UTF8_CONTROL));
}
To This:
public Text() {
FacesContext ctx = FacesContext.getCurrentInstance();
setParent(ResourceBundle.getBundle(BUNDLE_NAME,
ctx != null ? ctx.getViewRoot().getLocale() : Locale.ENGLISH, UTF8_CONTROL));
}
This now works both in JSF and JAX-RS context.
Hope this help,

JSF 2 checkboxes and boolean getters

I'm generating a jaxws client based on webservice. Jaxb will generate booleans using the java.lang.Boolean instead of the primitive type. In addition to this, it will generate the is() naming convention for beans.
However if I try to link the boolean (e.g. isOptional()) to a checkbox, it will throw the following exception:
value="#{property.optional}": Property 'optional' not readable on type java.lang.Boolean
My google skills have informed me that jsf works fine with:
boolean isOptional()
boolean getOptional()
Boolean getOptional()
But not with
Boolean isOptional()
However it is not feasible to update the beans manually due to the size and amount of the webservices, so is there any way to make jsf use the java.lang.Boolean isOptional() properly? Or can I somehow define a property in the jaxb bindings file at generation time which magically generates "getOptional()"?
On a sidenote, the following does work:
<h:selectBooleanCheckbox value="#{property.isOptional()}"/>
However I can't actually update the value presumably because it can't find the setter.
EDIT: I'm running the latest jdk 7, the output of "java -version":
java version "1.7.0_05"
Java(TM) SE Runtime Environment (build 1.7.0_05-b05)
Java HotSpot(TM) Client VM (build 23.1-b03, mixed mode, sharing)
The output of "wsimport -version":
JAX-WS RI 2.2.4-b01
Generated code:
public Boolean isOptional() {
return optional;
}
Jaxb will generate booleans using the java.lang.Boolean instead of the primitive type. In addition to this, it will generate the is() naming convention for beans.
Using the is getter prefix for java.lang.Boolean was a known major mistake of JAXB. It has been fixed in version 2.1.13 which was released April 2010 already. Keep your libraries up to date.
See also this blog article for some background.
The Great JAXB API Blunder
September 15, 2006
You've got to hand it to Sun for screwing this one up. It's one thing to write software that doesn't adhere to a specification when the documentation is as thick as a textbook. Take, for example, just about anything created by the W3C. However, it's really bad when it is your own spec that you can't follow, especially when it is the most well known part of it. That's right, Sun missed by a mile on their own spec when they created the JAXB 2.0 API. The JAXB 2.0 compiler (XJC) incorrectly uses the prefix "is" rather than "get" when generating the getter method for a java.lang.Boolean property. While the JavaBean spec states that read methods for primitive booleans can use the alternate "is" prefix, this flexibility does not extend to its boolean wrapper counterpart.
8.3.2 Boolean Properties
In addition, for boolean properties, we allow a getter method to match the pattern:
public boolean is();
This "is" method may be provided instead of a "get" method, or it may be provided in addition to a "get" method. In either case, if the "is" method is present for a boolean property then we will use the "is" method to read the property value.
An example boolean property might be:
public boolean isMarsupial();
public void setMarsupial(boolean m);
Given that JAXB is a code generation framework, and the idea behind code generation frameworks is that the code is to be used "as is" and not modified thereafter, this is a pretty big "oops". While this issue has been reported, the response from Sun is "sorry, its too late".
This behavior is governed by the spec, and unfortunately it's just too late for the spec to change now.
In terms of the user experience, thanks to auto-boxing, I don't think this will be a real issue for people. Is the problem that you are using Introspector and it's missing the property?
Too late? Not a real issue? It's BROKEN. FIX IT! I also don't like the naive statement that it probably won't affect frameworks. Um, yes it will, considering other projects did happen to adhere to the spec (hibernate, spring, myfaces, etc.)
UPDATE: Stevo Slavic informed me that this has been fixed in JAXB 2.1.13. See JAXB-131 for details. Yeah!
JSF/EL is not at fault here. It's doing its job properly conform the JavaBeans spec.
I'm not sure why the latest and greatest JAXB version still generates the wrong method but I finally fixed it by adding "-B-enableIntrospection" (as per http://jaxb.java.net/2.2.4/docs/xjc.html) to the wsimport call. This results in:
public Boolean getOptional() {
return optional;
}

Grails Spring Security Configuration thru xml

i am new at grails spring security, and i wanted to know if it is possible to enable/disable security of in my app thru an xml entry.
rather than having
grails.plugins.springsecurity.active = true
I d like to read the value from a bean that i declared in my resources.xml file.
grails.plugins.springsecurity.active = com.myorg.util.BeanUtil.getBean("repositorySettings").getIsSecured()
Using this approach throws errors. is there a way that i can accomplish this, and read the true/false value from a bean in the resources.xml
It's somewhat possible. resources.xml and resources.groovy are loaded after the plugin is, but parsed before. So you can put code in resources.groovy that will run while it's being parsed, before the plugin loads the config, and determines whether it's enabled. But you can't use a bean for that since it wouldn't be ready until it's too late. This wouldn't be possible in resources.xml since that's just XML but resources.groovy allows Groovy code and bean definitions:
import com.pbbi.mimgr.util.BeanUtil
beans = {
def grailsApplication = springConfig.unrefreshedApplicationContext.grailsApplication
def securityConfig = grailsApplication.config.grails.plugins.springsecurity
securityConfig.active = BeanUtil.getBean('repositorySettings').isSecured
}
If BeanUtil accesses the ApplicationContext then this won't work. But if it's accessing singletons that aren't Spring beans it should be fine.

Resources