March 2007 - jboss-dev-forums - Jboss List Archives

[Design of JBoss Portal] - JACC issues with the portal

by Wyoming

Hello We use JBoss (4.0.5.GA) with JACC Authorisation. When I tried to use the portal (2.6-DR) with JACC I found out some strange behaviour! The first issue is: I deployed my own portlet web application called "testportlet.war". If I call "Request.isUserInRole()" in my portlet and debug down into "DelegatingPolicy.implies()" then i see that the JACC contextID is not "testportlet.war" what I expected. Instead the contextID is: - "portal-server.war" when I have called the portlet directly in the portal - "portal-wsrp.war" when I have called the portlet via WSRP This doesn't make sense because like this the roles used for security checking have to be defined in one or even two of the portal's web.xml files. If I fix the value of the contextID in the debugger then this leads me to the next problem: The permissions loaded to the contextID "testportlet.war" do all have a servlet name "" (empty string). But the comparing permission that is created in "JaccAuthorizationRealm.hasRole()" has another servlet name: - "PortalServletWithPathMapping" calling the portlet directly in the portal - "MarkupService" when I have called the portlet via WSRP If I change the servlet name to "" in the debugger then it runs perfectly in both calling cases. So on my opinion these are bugs or does anyone have an idea how to change the behaviour by changing the configuration. regards Marcel View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=4031671#4031671 Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=4031671

17 years, 3 months

1
0
0 / 0

[Design of POJO Server] - Re: trunk testsuite status

by kabir.khan＠jboss.com

"clebert.suconic(a)jboss.com" wrote : this should be compatible... I think there is no problem on upgrading the dependency. | | | BTW: Is there a way of woven format being fixed between AOP 1.x and AOP 2.0? (at least aop classes compiled using 1.X should run fine on AOP 2.0 IMO. Let's discuss this here http://www.jboss.com/index.html?module=bb&op=viewtopic&p=4031647#4031647 View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=4031649#4031649 Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=4031649

17 years, 3 months

1
0
0 / 0

[Design of AOP on JBoss (Aspects/JBoss)] - Woven class compatibility

by kabir.khan＠jboss.com

Regarding http://www.jboss.com/index.html?module=bb&op=viewtopic&p=4031144#4031144 I've added a test target to the aop 2.0.0 testsuite to ensure that code woven with the "ClassicInstrumentor" in JBoss AOP 1.5.5.GA works when run with AOP 2.0.0.alphax. It runs the "basic" test. Clebert, do you have any examples of things not working across versions? View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=4031647#4031647 Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=4031647

17 years, 3 months

1
0
0 / 0

[Design of POJO Server] - Re: trunk testsuite status

by kabir.khan＠jboss.com

"scott.stark(a)jboss.org" wrote : "kabir.khan(a)jboss.com" wrote : I fixed some leaks in aop 2 this morning related to the use of PER_INSTANCE scoped aspects, which I see messaging is using. I've added a new aop snapshot containing this fix to the repository. This *should* be fine to use with the 1.2.0.SP.snapshot.aop2beta. Unless I'm totally off the mark there are no changes in the woven code created by aop alpha3 and the current snapshot. | | This is beyond the 2.0.0.alpha4 and in the current snapshot? Sorry did not see this before now. Yes, the fix is in the current snapshot, which is newer than alpha4 View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=4031641#4031641 Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=4031641

17 years, 3 months

1
0
0 / 0

[Design of JBoss internal QA (Test Suite)] - Re: JBoss 4.2.0 - jboss.examples:service=HASingletonMBeanExa

by rrajesh

You might want to take a look at the files that go into the configs the test is run against, in this case cluster-UDP-0 and cluster-UDP-1 The testsuite/build.xml should tell you how these configs are created. In this case take a look at the "create-cluster-node" target in testsuite/imports/server-config.xml View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=4031622#4031622 Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=4031622

17 years, 3 months

1
0
0 / 0

[Design of AOP on JBoss (Aspects/JBoss)] - Re: MethodInvocation.getArguments return value for 0 argumen

by kabir.khan＠jboss.com

OK, in the context of null vs non-null I get your point. Although I do actually prefer the empty array, as we discovered now for messaging people have aspects relying on the null behaviour. We should definitely use the same on both, so I think we should use null on both for backwards compatiblity reasons. View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=4031621#4031621 Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=4031621

17 years, 3 months

1
0
0 / 0

[Design of JBoss internal QA (Test Suite)] - JBoss 4.2.0 - jboss.examples:service=HASingletonMBeanExample

by ldimaggi＠redhat.com

I'm trying to debug this failure in the 4.2.0 testsuite: Suite: org.jboss.test.cluster.test.HASingletonElectionPolicyTestCase(Default-UDP) Test: testElectionPolicy Type: error Exception: javax.management.InstanceNotFoundException Message: jboss.examples:service=HASingletonMBeanExample_1 is not registered. In the server.log, I'm seeing this: ---------------------------------- 10:03:57,801 INFO [ServiceConfigurator] Problem configuring service jboss.examples:service=HASingletonMBeanExample-HASingletonController_1 org.jboss.deployment.DeploymentException: No Attribute found with name: HASingletonElectionPolicyMBean ---------------------------------- This is why the tests are failing. The thing that I'm stuck with is that the HASingletonElectionPolicyMBean class in question is present in jbossha.jar: ---------------------------------- [root@dhcp83-138 META-INF]# jar -tvf /var/lib/jbossas/server/cluster-UDP-0/lib/jbossha.jar | grep HASingletonElectionPolicyMBean 257 Mon Mar 19 19:01:36 EDT 2007 org/jboss/ha/singleton/HASingletonElectionPolicyMBean.class ---------------------------------- But - I guess that there is an attribute with the same name (HASingletonElectionPolicyMBean) defined where? Can anyone suggest how to debug this this? Thanks, Len DiMaggio ldimaggi(a)redhat.com View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=4031608#4031608 Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=4031608

17 years, 3 months

1
0
0 / 0

[Design of JBoss ESB] - Re: Advanced MessageStore implementation

by derek.adams

If you guys are interested, we implemented a basic JSR-170 (Jackrabbit) MessageStore a few weeks back. With work on MR1, it was put on the backburner, but I will push it into the trunk if there are not any objections. View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=4031604#4031604 Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=4031604

17 years, 3 months

1
0
0 / 0

[Design of AOP on JBoss (Aspects/JBoss)] - Re: MethodInvocation.getArguments return value for 0 argumen

by flavia.rainone＠jboss.com

I also think that it would be fine if we apply no arguments null to b/a/t/f. If we don't apply the same policy, it would make a difference. Look at this: | public void before(@Args Object arguments) | { | ... | } | | public Object around(Invocation invocation) throws Throwable | { | ... | } | Suppose both advices are applied to a POJO->method() execution joinpoint: | public class POJO | { | public void method() | { | ... | } | } | If we apply different policies to before and around, i.e.: - keep null @Args only for FieldRead (current policy), which means "before" advice would receive a non-null array in the example above - Invocation needs to contain a null array because the arguments length is 0 We would need to change JoinPointGenerator because it currently copies arguments array used on before advice to the invocation that will be passed as parameter to "around" advice, and this would break the null args invocation policy. This is just an example. But it illustrates how interaction between b/a/t/f and around makes a difference when there are no arguments, if we use different null policies. Nonetheless, I can also implement different policies on both, if that is the case. View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=4031593#4031593 Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=4031593

17 years, 3 months

1
0
0 / 0

[Design of Messaging on JBoss (Messaging/JBoss)] - Re: org.jboss.test.cts.test.MDBUnitTestCase

by clebert.suconic＠jboss.com

duhhh... I was kind of sleepy.... Also fixed the ConsumerAspect... good catch! :-) View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=4031581#4031581 Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=4031581

17 years, 3 months

1
0
0 / 0

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

jboss-dev-forums March 2007