This is a distinct filter as apposed to reusing StaticHeaderWriter
since the specification specifies that the "Strict-Transport-Security"
header should only be set on secure requests. It would not make sense to
require DelegatingRequestMatcherHeaderWriter since this requirement is
in the specification.
- hf.doFilter is missing FilterChain argument
- response.headers does not contain the exact values for the headers so
should not be used for comparison (note it is a private member so this
is acceptable)
- hf does not need non-null check when hf.doFilter is invoked
- some of the configurations are no longer valid (i.e. ALLOW-FROM
requires strategy)
- Some error messages needed updated (some could still use improvement)
- No validation for missing header name or value
- rebased off master / merged
- nsa=frame-options-strategy id should use - not =
- FramewOptionsHeaderFactory did not produce "ALLOW-FROM " prefix of origin
- remove @Override on interface overrides to work with JDK5
- Implemented different ALLOW-FROM strategies as specified in the proposal.
Conflicts:
config/src/main/java/org/springframework/security/config/http/HeadersBeanDefinitionParser.java
config/src/test/groovy/org/springframework/security/config/http/HttpHeadersConfigTests.groovy
Created HeadersFilter for setting security headers added including a
bean definition parser for easy configuration of the headers. Enables
easy configuration for the X-Frame-Options, X-XSS-Protection and
X-Content-Type-Options headers. Also allows for additional headers to
be added.
Session fixation protection, whether by clean new session or
migrated session, now publishes an event when a session is
migrated or its ID is changed. This enables application developers
to keep track of the session ID of a particular authentication
from the time the authentication is successful until the time
of logout. Previously this was not possible since session
migration changed the session ID and there was no way to
reliably detect that.
Revised changes per Rob Winch's suggestions.
Previously Spring Security would disable automatically saving the
SecurityContext when the Thread was different than the Thread that
created the SaveContextOnUpdateOrErrorResponseWrapper. This worked for
many cases, but could cause issues when a timeout occurred. The problem
is that a Thread can be reused to process the timeout since the Threads
are pooled. This means that a timeout of a request trigger an apparent
logout as described in the following workflow:
- The SecurityContext was established on the SecurityContextHolder
- An Async request was made
- The SecurityContextHolder would be cleared out
- The Async request times out
- The Async request would be dispatched back to the container upon
timing out. If the container reused the same Thread to process the
timeout as the original request, Spring Security would attempt to
save the SecurityContext when the response was committed. Since the
SecurityContextHolder was still cleared out it removes the
SecurityContext from the HttpSession
Spring Security will now prevent the SecurityContext from automatically
being saved when the response is committed as soon as
HttpServletRequest#startAsync() or
ServletRequest#startAsync(ServletRequest,ServletResponse) is called.
Both overloads of
AbstractAuthenticationProcessingFilter.successfulAuthentication()
claimed to invoke SessionAuthenticationStrategy, which is not true, as
the invokation happens earlier in doFilter(). The Javadoc on these
methods are updated to reflect the actual code.
Previously DummyRequest implemented HttpServletRequest which caused complications
since Servlet 2.5 and Servlet 3 had non passive changes. While we were "safe" if the
Servlet 3 methods were never invoked reflective access of the methods would also
problems. We could prevent users from accessing the methods of DummyRequest by
returning new HttpServletRequestWrapper(DummyRequest), but a debugger could
potentially try to iterate over the methods triggering a NoClassDefFoundError.
DummyRequest now extends HttpServletRequestWrapper which will be dynamically
linked to the proper version of HttpServletRequest. We use a Dynamic Proxy that
throws UnsupportedOperationException to implement any methods we are not
interested in.
Previously SecurityContextCallableProcessingInterceptor used afterCompletion
to clear the SecurityContextHolder. This does not work since afterCompletion
is invoked on the Servlet Container thread.
Now SecurityContextCallableProcessingInterceptor clears the
SecurityContextHolder on postProcess which is invoked on the same thread
that the Callable is processed on.
Previously, if the Principal returned by getPreAuthenticatedPrincipal was not a String,
it prevented requiresAuthentication from detecting when the Principal was the same.
This caused the need to authenticate the user for every request even when the Principal
did not change.
Now requiresAuthentication will check to see if the result of
getPreAuthenticatedPrincipal is equal to the current Authentication.getPrincipal().
Previously a NullPointerException would occur if an HttpServletRequest.getMethod()
returned null.
Now AntPathRequestMatcher and RegexpRequestMatcher will handle if the
HttpServletRequest.getMethod() returns null. While under normal circumstances,
it is unlikely for the method to be null this can occur when using
DefaultWebInvocationPrivilegeEvaluator.isAllowed(String, Authentication).
Previously SaveContextOnUpdateOrErrorResponseWrapper would save the SecurityContext on a different
Threads than the one it was created on. This causes issues with Async Web requests which may write
to the response on a new Thread.
Now SaveContextOnUpdateOrErrorResponseWrapper will not save the SecurityContext when a different
Thread invokes any of the methods that commit the response. This prevents issues with Async
processing. However, explicit calls to SecurityContextRepository.save will still save the
SecurityContext since it invokes the saveRequest method rather than private doSave method within
the SaveContextOnUpdateOrErrorResponseWrapper which contains the logic to prevent saving from
another Thread.
Previously the SwitchUserFilter was logging as an error and then
throwing an Exception immediately after. This is not correct, since
whomever is catching the Exception should choose to log an error or not.
Now the log statement is at a debug level.
Previously there was a race condition could occur when the user attempts to access
a slow resource and then logs out which would result in the user not being logged
out.
SecurityContextLogoutHandler will now remove the Authentication from the
SecurityContext to protect against this scenario.
Previously Spring Security did not save the Security Context immediately prior
to the following methods being invoked:
- HttpServletResonse.flushBuffer()
- HttpServletResonse.getWriter().close()
- HttpServletResonse.getWriter().flush()
- HttpServletRespose.getOutputStream().close()
- HttpServletRespose.getOutputStream().flush()
This meant that the client could get a response prior to the SecurityContext
being stored. After the client got the response, it would make another request
and this would not yet be authenticated. The reason this can occur is because
all of the above methods commit the response, which means that the server can
signal to the client the response is completed. A similar issue happened in
SEC-398.
Now the previously listed methods are wrapped in order to ensure the SecurityContext
is persisted prior to the response being committed.
When SEC-1950 was introduced it caused problems when a <filter-mapping> was mapped
to multiple dispatchers (i.e. REQUEST and FORWARD) since when the second dispatcher
completed execution it cleared the SecurityContext and the original FilterChain
would then save the cleared out SecurityContext.
We now use a pattern similar to the OncePerRequestFilter to only invoke
SecurityContextHolder.clearContext() on the first invocation of the Filter. We do not simply extend
OncePerRequestFilter because we want to invoke the delegate filters for every request.
Previously communication errors with LDAP were only logged at debug level.
Communication errors (along with other non-authenticated related NamingExceptions)
are now logged as error messages. We created an InternalAuthetnicationServiceException
to represent errors that should be logged as errors to distinguish between internal
and external authentication failures. For example, we do not want an OpenID Provider
being able to report errors that cause our logs to fill up. However, an LDAP system is
internal and should be trusted so logging at an error level makes sense.
Previously subclasses of AbstractSecurityInterceptor did not restore the original
Authentication when RunAsManager was used and an Exception was thrown in the
original method.
AbstractSecurityInterceptor has added a new method finallyInvocation which
should be invoked in a finally block immediately after the original invocation
which will restore the original Authentication. All existing sub classes have
been updated to use this new method.
Rob Winch
Marten Deinum
Keesun Baik
Nicholas Williams
Balazs Zagyvai
mpigg