RE: Integrating JAX-RS runtime into DOSGi
Hi I've implemented just now on the trunk. I have the unit test only assuming that the calling BundleContext will be used to load the classes - can you please verify it? I'm not sure when exactly the Hudson build at http://hudson.zones.apache.org/hudson/job/CXF-DOSGi/ Will start and finish, hopefully soon enough, tomorrow will be ready for sure. Alternatively please rebuild dsw/cxf-dsw and then the distribution (single or multi) that you use... thanks, Sergey -Original Message- From: jmholtz...@gmail.com [mailto:jmholtz...@gmail.com] On Behalf Of Josh Holtzman Sent: 31 August 2009 10:38 To: dev@cxf.apache.org Subject: Re: Integrating JAX-RS runtime into DOSGi I'm using declarative services to register my endpoints, so it would be useful for me to specify the provider property using class names rather than instances in the service properties. Perhaps something like this? property name=org.apache.cxf.rs.provider org.whatever.FooXmlReaderWriter, org.whatever.SomeOtherReaderOrWriter /property Or is there some way to reference an object instance in DS that I'm unaware of? Thanks, Josh On Wed, Aug 26, 2009 at 6:45 PM, Sergey Beryozkin sbery...@progress.comwrote: Oh, this is exactly the sort of message I was in need for in the end of the long and difficult day :-) Cool. If you could verify (later on today or next week when you're back) that you could do Dictionary props = new Hashtable(); props.put(org.apache.cxf.rs.provider, new Object[]{new FooReaderWriter()}); bundleContext.registerService(YourApplicationService.class.getName, new YourApplicationServiceImpl(), properties); This option will allow users to register per-service tuned FooReaderWriters, something one can do today with registering different FooReaderWriter instances on a per-endpoint basis with jaxrs:endpoints then it would be of help. thanks for your help Sergey - Original Message - From: Josh Holtzman jholtz...@berkeley.edu To: dev@cxf.apache.org Sent: Wednesday, August 26, 2009 5:37 PM Subject: Re: Integrating JAX-RS runtime into DOSGi Bah, my isReadable and isWriteable were wrong. Sorry, false alarm. This works like a charm! Thanks, Josh On Wed, Aug 26, 2009 at 6:16 PM, Sergey Beryozkin sbery...@progress.com wrote: Hi Josh Thanks, this is exactly how providers are expected to be registered. http://hudson.zones.apache.org/hudson/job/CXF-DOSGi shows my changes have been picked up so it's a bug then. Have you tries the lastest build from snapshots or built the trunk yourself, including a trunk/distributuon ? I'll look into it... cheers, Sergey P.S. was about to ping you on #cxf but somehow I lost the connection - Original Message - From: Josh Holtzman jholtz...@berkeley.edu To: dev@cxf.apache.org Sent: Wednesday, August 26, 2009 5:02 PM Subject: Re: Integrating JAX-RS runtime into DOSGi Great, thanks Sergey. I just tried this, and wasn't able to read/write an arbitrary object. I've registered a MessageBodyReader and a MessageBodyWriter (actually, the same object) like so: FooXmlReaderWriter fooReaderWriter = new FooXmlReaderWriter(); context.registerService(MessageBodyReader.class.getName(), fooReaderWriter, null); context.registerService(MessageBodyWriter.class.getName(), fooReaderWriter, null); I see that these services are in fact registered: --- objectClass = javax.ws.rs.ext.MessageBodyReader service.id = 41 objectClass = javax.ws.rs.ext.MessageBodyWriter service.id = 42 But when I try to access a Foo via JAX-RS, I get this: [9853...@qtp-998044-0 - /inspection/rest/foo] WARN org.apache.cxf.jaxrs.interceptor.JAXRSOutInterceptor - .No message body writer found for response class : Foo. Please let me know if I misunderstood the recipe for registering JAX-RS providers. I'll be out of the office until Monday, and will start digging into the CXF code to see what's going on when I return. Many thanks, Josh On Tue, Aug 25, 2009 at 7:21 PM, Sergey Beryozkin sergey.beryoz...@iona.com wrote: Hi Josh I've updated the JAX-RS layer in DOSGi such that it will now discover JAXRS (and CXF specific providers) which have been registered as (global) OSGI services. At the moment I've decided not to use a ServiceTracker and instead a calling BundleContext is asked to exercise a filter expression which should catch JAXRS MessageBodyReader, MessageBodyWriter, ExceptionMapper, as well as CXF RequestHandler, ResponseHandler ParameterHandler. I'll attempt to optimize it later on One can disable such queries for such providers and also insist that only those global providers which have identified themselves (through a specific property) that they will work reliably with CXF can be used. Alternatively, one can register an array of service/endpoint -specific providers by using org.apache.cxf.rs.provider property, when
Re: Handling collections with Aegis in JAX-RS
Sergey, If you want to me to tackle PADB for Aegis, you have a writing assignment. Please write up a comprehensive explanation in the javadoc. Just calling it an 'alternative' does not give an implementor much to go with. I think the top-level issue with namespaces is going to boil down to the existence, or nonexistence, of 'parts', but I'll report back presently. --benson On Mon, Aug 24, 2009 at 8:58 AM, Sergey Beryozkinsergey.beryoz...@iona.com wrote: Hi Benson Thanks for spending your time on these tests, and fixing CXF 2401. I was also able to add few more tests, including the one which writes/reads a complex Map to/from JSON. AegisJSONProvider tries its best for users to avoid setting up a namespace map manually, but in cases when it does not guess properly the users would be able to overwrite namespaces and their prefixes as needed. This isssue would likely arise on the read side only, when CXF reads it. JSON is namespace-unaware so when reading, one needs to setup a namespace map for Jettison to report values like @ns1.bar to the JAXB reader... (but only if AegisJSONProvider has not guessed how to map prefixes to namespaces) There's a couple of issues I'd like to discuss. First one is that AegisJsonProviderTest.testReadWriteComplexMap does have to setup a namespace for a map root element, but it does not have to do it for testWrite/ReadCollection tests. I'm wondering, can it be avoided in cases when Maps are being written/read ? That is, can we modify createReader/Writer methods such that QNames for containers like Map/Collections and its member types are available ? Another one is that DataBindingJSONProvider test (aegis tests) still can not handle collections. I think the problem there is that when DataBindingProvider (the one DataBindingJSONProvider extends) is being initialized, it uses a workaround which we discussed in the other thread. Namely, it attempts to convert a JAXRS model info into WSDL-like one and set calls DataBinding.initialize(Service). It's quite limiting for a number of reasons, and one of them is that the generic types are not visible to data bindings So we introduced PropertiesAwareDataBinding interface and I've just updated the JAXRS code to call it like this : MapClass?, Type allClasses = getAllJAXRSResponseInputTypes(); MapString, Object props = new HashMapString, Object(); props.put(PropertiesAwareDataBinding.TYPES_PROPERTY, allClasses); ((PropertiesAwareDataBinding)db).initialize(props); At the moment no CXF DataBindings implement this newly introduced interface. I think the only way for DataBindingJSONProvider aegis tests which handle collections to start passing is for Aegis DataBinding to implement PropertiesAwareDataBinding and initialize itself using the provided MapClass?, Type, as opposed to Service. Would you be open to updating the Aegis databinding ? thanks, Sergey bimargulies wrote: I have a rather clear memory of working on these, there wasn't enough passing of Generic classes around. I'll go have a look. On Fri, Aug 21, 2009 at 6:38 AM, Sergey Beryozkinsergey.beryoz...@iona.com wrote: Hi Benson if you could look at any of these tests or at least point me in the right direction then it would be great. I know you're busy - so just look at it whenever you get a chance, not urgent... cheers, Sergey Sergey Beryozkin-2 wrote: Hi Benson I can't make the Aegis tests writing/reading collections working in CXF JAX-RS. I've found that AegisProviderTest#testReadWriteComplexMap is still @Ignored, it might've passed for you because it was @Ignored :-) I've also added testWriteCollections() (which writes ListAegisTestBean) to AegisJSONProviderTest. I also updated DataBindingJSONProviderTest, one of its internal classes to return ListBook. AegisJSONProvider extends AegisElementProvider, DataBindingJSONProvider extends DataBindingProvider which actually (in this case) delegates to Aegis DataBinding. AegisJSONProviderTest fails at the write time, it can't find the mapping for List. DataBindingJSONProviderTest fails early at the Aegis DataBinding initialization time for the same reason. I thought Lists were supported by default ? I haven't found any exam[le showing how a type mapping for Lists can be created. Can you please, whenever you have a chance, have a look at these tests ? thanks, Sergey -- View this message in context: http://www.nabble.com/Handling-collections-with-Aegis-in-JAX-RS-tp24933144p25076146.html Sent from the cxf-dev mailing list archive at Nabble.com. -- View this message in context: http://www.nabble.com/Handling-collections-with-Aegis-in-JAX-RS-tp24933144p25115676.html Sent from the cxf-dev mailing list archive at Nabble.com.
Re: Handling collections with Aegis in JAX-RS
The more I look at this, the less I understand it. Aegis has a perfectly good API for passing in a set of root classes. Why is it better to pass that collection through this property map instead of calling setRootClasses? So that you can get to the other data bindings the same way? It still won't work for generics. The problem is this. Aegis has a pre-generic concept of root classes, because they are *classes*, not reflection types. Classes don't carry generic information. While Aegis has a default mappings, it won't use it for root types. Root types have to be in the root classes or derived from the service model. My memory isn't good enough to recall how JAXB deals. When a bean has a generic type as a property, the code is using reflection, so it can do generic type processing. So, to get what you want, we'd need to rework your interface to use java.lang.reflect.Type. I'm willing to look into the Aegis-specific angle here for now and see how to add generic roots. I think the new interface needs more thinks. On Mon, Aug 24, 2009 at 8:58 AM, Sergey Beryozkinsergey.beryoz...@iona.com wrote: Hi Benson Thanks for spending your time on these tests, and fixing CXF 2401. I was also able to add few more tests, including the one which writes/reads a complex Map to/from JSON. AegisJSONProvider tries its best for users to avoid setting up a namespace map manually, but in cases when it does not guess properly the users would be able to overwrite namespaces and their prefixes as needed. This isssue would likely arise on the read side only, when CXF reads it. JSON is namespace-unaware so when reading, one needs to setup a namespace map for Jettison to report values like @ns1.bar to the JAXB reader... (but only if AegisJSONProvider has not guessed how to map prefixes to namespaces) There's a couple of issues I'd like to discuss. First one is that AegisJsonProviderTest.testReadWriteComplexMap does have to setup a namespace for a map root element, but it does not have to do it for testWrite/ReadCollection tests. I'm wondering, can it be avoided in cases when Maps are being written/read ? That is, can we modify createReader/Writer methods such that QNames for containers like Map/Collections and its member types are available ? Another one is that DataBindingJSONProvider test (aegis tests) still can not handle collections. I think the problem there is that when DataBindingProvider (the one DataBindingJSONProvider extends) is being initialized, it uses a workaround which we discussed in the other thread. Namely, it attempts to convert a JAXRS model info into WSDL-like one and set calls DataBinding.initialize(Service). It's quite limiting for a number of reasons, and one of them is that the generic types are not visible to data bindings So we introduced PropertiesAwareDataBinding interface and I've just updated the JAXRS code to call it like this : MapClass?, Type allClasses = getAllJAXRSResponseInputTypes(); MapString, Object props = new HashMapString, Object(); props.put(PropertiesAwareDataBinding.TYPES_PROPERTY, allClasses); ((PropertiesAwareDataBinding)db).initialize(props); At the moment no CXF DataBindings implement this newly introduced interface. I think the only way for DataBindingJSONProvider aegis tests which handle collections to start passing is for Aegis DataBinding to implement PropertiesAwareDataBinding and initialize itself using the provided MapClass?, Type, as opposed to Service. Would you be open to updating the Aegis databinding ? thanks, Sergey bimargulies wrote: I have a rather clear memory of working on these, there wasn't enough passing of Generic classes around. I'll go have a look. On Fri, Aug 21, 2009 at 6:38 AM, Sergey Beryozkinsergey.beryoz...@iona.com wrote: Hi Benson if you could look at any of these tests or at least point me in the right direction then it would be great. I know you're busy - so just look at it whenever you get a chance, not urgent... cheers, Sergey Sergey Beryozkin-2 wrote: Hi Benson I can't make the Aegis tests writing/reading collections working in CXF JAX-RS. I've found that AegisProviderTest#testReadWriteComplexMap is still @Ignored, it might've passed for you because it was @Ignored :-) I've also added testWriteCollections() (which writes ListAegisTestBean) to AegisJSONProviderTest. I also updated DataBindingJSONProviderTest, one of its internal classes to return ListBook. AegisJSONProvider extends AegisElementProvider, DataBindingJSONProvider extends DataBindingProvider which actually (in this case) delegates to Aegis DataBinding. AegisJSONProviderTest fails at the write time, it can't find the mapping for List. DataBindingJSONProviderTest fails early at the Aegis DataBinding initialization time for the same reason. I thought Lists were supported by default ? I haven't found any exam[le showing how a type mapping for Lists can be created. Can you please,
RE: Integrating JAX-RS runtime into DOSGi
Missed 'would result in the array value being passed'... -Original Message- From: Sergey Beryozkin [mailto:sbery...@progress.com] Sent: 31 August 2009 13:10 To: dev@cxf.apache.org Subject: RE: Integrating JAX-RS runtime into DOSGi Ok, I didn't know property name=org.apache.cxf.rs.provider org.whatever.MyJaxRsReaderWriter, org.whatever.SomeOtherReaderOrWriter /property I'll look more into it... Thanks, Sergey -Original Message- From: jmholtz...@gmail.com [mailto:jmholtz...@gmail.com] On Behalf Of Josh Holtzman Sent: 31 August 2009 13:04 To: dev@cxf.apache.org Subject: Re: Integrating JAX-RS runtime into DOSGi This works when specifying a single property, but not as an array. In other words, this is fine: property name=org.apache.cxf.rs.provider value=org.whatever.MyJaxRsReaderWriter / but this does not work: property name=org.apache.cxf.rs.provider org.whatever.MyJaxRsReaderWriter, org.whatever.SomeOtherReaderOrWriter /property I seem to recall a general CXF issue with array properties in DS, so this may be completely unrelated to the JAX-RS stuff. Thanks, Josh On Mon, Aug 31, 2009 at 1:35 PM, Sergey Beryozkin sbery...@progress.comwrote: Hi I've implemented just now on the trunk. I have the unit test only assuming that the calling BundleContext will be used to load the classes - can you please verify it? I'm not sure when exactly the Hudson build at http://hudson.zones.apache.org/hudson/job/CXF-DOSGi/ Will start and finish, hopefully soon enough, tomorrow will be ready for sure. Alternatively please rebuild dsw/cxf-dsw and then the distribution (single or multi) that you use... thanks, Sergey -Original Message- From: jmholtz...@gmail.com [mailto:jmholtz...@gmail.com] On Behalf Of Josh Holtzman Sent: 31 August 2009 10:38 To: dev@cxf.apache.org Subject: Re: Integrating JAX-RS runtime into DOSGi I'm using declarative services to register my endpoints, so it would be useful for me to specify the provider property using class names rather than instances in the service properties. Perhaps something like this? property name=org.apache.cxf.rs.provider org.whatever.FooXmlReaderWriter, org.whatever.SomeOtherReaderOrWriter /property Or is there some way to reference an object instance in DS that I'm unaware of? Thanks, Josh On Wed, Aug 26, 2009 at 6:45 PM, Sergey Beryozkin sbery...@progress.comwrote: Oh, this is exactly the sort of message I was in need for in the end of the long and difficult day :-) Cool. If you could verify (later on today or next week when you're back) that you could do Dictionary props = new Hashtable(); props.put(org.apache.cxf.rs.provider, new Object[]{new FooReaderWriter()}); bundleContext.registerService(YourApplicationService.class.getName, new YourApplicationServiceImpl(), properties); This option will allow users to register per-service tuned FooReaderWriters, something one can do today with registering different FooReaderWriter instances on a per-endpoint basis with jaxrs:endpoints then it would be of help. thanks for your help Sergey - Original Message - From: Josh Holtzman jholtz...@berkeley.edu To: dev@cxf.apache.org Sent: Wednesday, August 26, 2009 5:37 PM Subject: Re: Integrating JAX-RS runtime into DOSGi Bah, my isReadable and isWriteable were wrong. Sorry, false alarm. This works like a charm! Thanks, Josh On Wed, Aug 26, 2009 at 6:16 PM, Sergey Beryozkin sbery...@progress.com wrote: Hi Josh Thanks, this is exactly how providers are expected to be registered. http://hudson.zones.apache.org/hudson/job/CXF-DOSGi shows my changes have been picked up so it's a bug then. Have you tries the lastest build from snapshots or built the trunk yourself, including a trunk/distributuon ? I'll look into it... cheers, Sergey P.S. was about to ping you on #cxf but somehow I lost the connection - Original Message - From: Josh Holtzman jholtz...@berkeley.edu To: dev@cxf.apache.org Sent: Wednesday, August 26, 2009 5:02 PM Subject: Re: Integrating JAX-RS runtime into DOSGi Great, thanks Sergey. I just tried this, and wasn't able to read/write an arbitrary object. I've registered a MessageBodyReader and a MessageBodyWriter (actually, the same object) like so: FooXmlReaderWriter fooReaderWriter = new FooXmlReaderWriter(); context.registerService(MessageBodyReader.class.getName(), fooReaderWriter, null); context.registerService(MessageBodyWriter.class.getName(), fooReaderWriter, null); I see that these services are in fact registered: --- objectClass = javax.ws.rs.ext.MessageBodyReader service.id = 41 objectClass = javax.ws.rs.ext.MessageBodyWriter service.id = 42 But when I try to access a Foo via JAX-RS, I get this:
Missing OM* classes
Hi all, I'm having a problems compiling CXF 2.2.3. Very early, when it reaches Apache CXF API it hits errors for missing the AXIOM classes: [INFO] [INFO] Building Apache CXF API [INFO]task-segment: [install] [INFO] [INFO] [dependency:unpack {execution: unpack-schemas}] [INFO] Configured Artifact: org.apache.cxf:cxf-common-schemas:2.2.3:jar [INFO] Unpacking D:\.m2\repository\org\apache\cxf\cxf-common-schemas\2.2.3\cxf-common-schemas-2.2.3.jar to D:\aiq\jbossws\projects\cxf-2.2.3\api\target\schemas with includes null and excludes:null [INFO] [cxf-xml2fastinfoset:xml2fastinfoset {execution: xml2fastinfoset}] [INFO] [cxf-common-xsd:xsdtojava {execution: generate-sources}] [INFO] [remote-resources:process {execution: default}] [INFO] [resources:resources {execution: default-resources}] [INFO] Using 'UTF-8' encoding to copy filtered resources. [INFO] Copying 15 resources [INFO] skip non existing resourceDirectory D:\aiq\jbossws\projects\cxf-2.2.3\api\src\main\resources [INFO] skip non existing resourceDirectory D:\aiq\jbossws\projects\cxf-2.2.3\api\src\main\resources-filtered [INFO] skip non existing resourceDirectory D:\aiq\jbossws\projects\cxf-2.2.3\api\target\generated\src\main\resources [INFO] Copying 3 resources [INFO] [compiler:compile {execution: default-compile}] [INFO] Compiling 263 source files to D:\aiq\jbossws\projects\cxf-2.2.3\api\target\classes [INFO] [ERROR] BUILD FAILURE [INFO] [INFO] Compilation failure could not parse error message: org\apache\neethi\PolicyReference.java(org\apache\neethi:PolicyReference.java):35: package org.apache.axiom.om does not exist import org.apache.axiom.om.OMElement; ^ could not parse error message: org\apache\neethi\PolicyReference.java(org\apache\neethi:PolicyReference.java):36: package org.apache.axiom.om.impl.builder does not exist import org.apache.axiom.om.impl.builder.StAXOMBuilder; .GOES ON LIKE THIS
warning: not quite compatible change coming to Aegis in 2.3
Aegis takes a collection of root classes declared as SetClass?. I plan to change this to be Setjava.lang.reflect.Type to permit proper generic analysis. Given how types do and don't work in Java, this will be slightly incompatible.
Bigger proposed Aegis change: rename 'Type'
Aegis has a class named 'Type'. This leads to confusing code all over the place when organizing imports around java.lang.reflect.Type. Anyone care to object to changing it to AegisType for 2.3?
Re: warning: not quite compatible change coming to Aegis in 2.3
On Mon August 31 2009 10:50:42 am Benson Margulies wrote: Aegis takes a collection of root classes declared as SetClass?. I plan to change this to be Setjava.lang.reflect.Type to permit proper generic analysis. Given how types do and don't work in Java, this will be slightly incompatible. I actually think this is OK since Class implements Type. In anycase, just add it to:http://cxf.apache.org/23-migration-guide.html -- Daniel Kulp dk...@apache.org http://www.dankulp.com/blog
Re: recent changes
I'm going to revert this change, at least for now until things are fully settled and we get the major stuctural changes in systest pushed back through the fixes. Then, we can probably revisit it on trunk only. The major issue is that those classes are really there to help ANYONE that is using CXF to write unit/system tests for their own work. I know Progress uses it all over the place. Thus, to move it out of rt-core will cause major headaches and not something I want to do on the 2.1/2.2 branches. For trunk, it may make sense, but I'd like to get the rest of the refactoring of the test done first and merged before adding to my misery. :-) Dan On Sun August 30 2009 8:51:51 pm Benson Margulies wrote: While I was working on splitting the systests, I set out to clean up some other test-related stuff, perhaps rashly. There is some test-only code in rt-core. I moved it from src/main/java to src/test/java, and changed the POM for rt-core to build a test jar. Then I 'fixed' the dependencies to have a scope=test dependency on rt-core with classifier 'tests'. For this to work, I had to change the fastinstall profile to use maven.test.skip.exec instead of plain maven.test.skip. Elsewise, the tests don't get compiled, and the jar doesn't happen, and the dependent project refuses to build. To really do this cleanly, someone aught to split up rt-core so that the stuff that the test classes depend on was in one slice and the other stuff that depends on the test classes was in another, but I think I've run out of courage for the moment. -- Daniel Kulp dk...@apache.org http://www.dankulp.com/blog
RE: Integrating JAX-RS runtime into DOSGi
Hi Josh This should be fixed now. I've confirmed it by running the locally-modified greeter_rest demo on Eclipse Equinox with its DS service implementation... Unit test has also been there. Now, Equinox DS passes a Spring[] so at the moment I'm assuming that it has to be String array, as opposed to Object[]. Given that DS is likely implemented by Equinox only it should work just fine. Can you confirm it's working for you (after the build has been done) ? So we have 3 options now for registering providers : - registering providers as OSGI services - providing an array of provider instances as an application service property - using declarative services I'm not sure it is worth investigating the option of class-scanning the application bundle - it would be more expensive. But may be I will later on... Thanks, Sergey -Original Message- From: jmholtz...@gmail.com [mailto:jmholtz...@gmail.com] On Behalf Of Josh Holtzman Sent: 31 August 2009 13:04 To: dev@cxf.apache.org Subject: Re: Integrating JAX-RS runtime into DOSGi This works when specifying a single property, but not as an array. In other words, this is fine: property name=org.apache.cxf.rs.provider value=org.whatever.MyJaxRsReaderWriter / but this does not work: property name=org.apache.cxf.rs.provider org.whatever.MyJaxRsReaderWriter, org.whatever.SomeOtherReaderOrWriter /property I seem to recall a general CXF issue with array properties in DS, so this may be completely unrelated to the JAX-RS stuff. Thanks, Josh On Mon, Aug 31, 2009 at 1:35 PM, Sergey Beryozkin sbery...@progress.comwrote: Hi I've implemented just now on the trunk. I have the unit test only assuming that the calling BundleContext will be used to load the classes - can you please verify it? I'm not sure when exactly the Hudson build at http://hudson.zones.apache.org/hudson/job/CXF-DOSGi/ Will start and finish, hopefully soon enough, tomorrow will be ready for sure. Alternatively please rebuild dsw/cxf-dsw and then the distribution (single or multi) that you use... thanks, Sergey -Original Message- From: jmholtz...@gmail.com [mailto:jmholtz...@gmail.com] On Behalf Of Josh Holtzman Sent: 31 August 2009 10:38 To: dev@cxf.apache.org Subject: Re: Integrating JAX-RS runtime into DOSGi I'm using declarative services to register my endpoints, so it would be useful for me to specify the provider property using class names rather than instances in the service properties. Perhaps something like this? property name=org.apache.cxf.rs.provider org.whatever.FooXmlReaderWriter, org.whatever.SomeOtherReaderOrWriter /property Or is there some way to reference an object instance in DS that I'm unaware of? Thanks, Josh On Wed, Aug 26, 2009 at 6:45 PM, Sergey Beryozkin sbery...@progress.comwrote: Oh, this is exactly the sort of message I was in need for in the end of the long and difficult day :-) Cool. If you could verify (later on today or next week when you're back) that you could do Dictionary props = new Hashtable(); props.put(org.apache.cxf.rs.provider, new Object[]{new FooReaderWriter()}); bundleContext.registerService(YourApplicationService.class.getName, new YourApplicationServiceImpl(), properties); This option will allow users to register per-service tuned FooReaderWriters, something one can do today with registering different FooReaderWriter instances on a per-endpoint basis with jaxrs:endpoints then it would be of help. thanks for your help Sergey - Original Message - From: Josh Holtzman jholtz...@berkeley.edu To: dev@cxf.apache.org Sent: Wednesday, August 26, 2009 5:37 PM Subject: Re: Integrating JAX-RS runtime into DOSGi Bah, my isReadable and isWriteable were wrong. Sorry, false alarm. This works like a charm! Thanks, Josh On Wed, Aug 26, 2009 at 6:16 PM, Sergey Beryozkin sbery...@progress.com wrote: Hi Josh Thanks, this is exactly how providers are expected to be registered. http://hudson.zones.apache.org/hudson/job/CXF-DOSGi shows my changes have been picked up so it's a bug then. Have you tries the lastest build from snapshots or built the trunk yourself, including a trunk/distributuon ? I'll look into it... cheers, Sergey P.S. was about to ping you on #cxf but somehow I lost the connection - Original Message - From: Josh Holtzman jholtz...@berkeley.edu To: dev@cxf.apache.org Sent: Wednesday, August 26, 2009 5:02 PM Subject: Re: Integrating JAX-RS runtime into DOSGi Great, thanks Sergey. I just tried this, and wasn't able to read/write an arbitrary object. I've registered a MessageBodyReader and a MessageBodyWriter (actually, the same object) like so: FooXmlReaderWriter fooReaderWriter = new FooXmlReaderWriter(); context.registerService(MessageBodyReader.class.getName(), fooReaderWriter, null);
generics and type mapping in Aegis
Sergey's issues with JAX-RS have started me looking at how Aegis maps types. The following temptation now occurs: allow the fundamental aegis type mechanism to map generics. Somene could thus add a mapping from any Xy,x,q - XML via a custom type. Of course, with type erasure, this only will pay attention to the declared type, not the dynamic type. Somehow, I feel as if this is a slippery slope we don't want to start down. However, the defined APIs for JAX-RS have started us there, with their extra arguments for generic type specs. Apparently, someone thinks that it's so important to be able to write a CollectionX to the top of JAX-RS to justify adding an asymmetry. Anybody else got a thought here?
Re: Bigger proposed Aegis change: rename 'Type'
On Mon August 31 2009 10:55:16 am Benson Margulies wrote: Aegis has a class named 'Type'. This leads to confusing code all over the place when organizing imports around java.lang.reflect.Type. Anyone care to object to changing it to AegisType for 2.3? Nope. Migration guide it it affects people. Though people who are accessing the Aegis Type things are probably few and far between. -- Daniel Kulp dk...@apache.org http://www.dankulp.com/blog
Re: Reliable messaging: Connecting to Oracle
Thanks for the patch, Dan. Your efforts are much appreciated. I've committed the fix just now in r809738. Cheers, Eoghan 2009/8/27 Daniel Kulp dk...@apache.org Please file a JIRA and submit the changes as a patch. This is excellent! https://issues.apache.org/jira/browse/CXF Dan On Thu August 27 2009 10:17:52 am Dan Ryazansky wrote: The current version of CXF uses a Derby database when Reliable messaging is turned on. However, we required use of ORACLE. I had to make the following changes to to make the code database agnostic in RMTxStore: from CUR_MSG_NO DECIMAL(31, 0) NOT NULL DEFAULT 1, to CUR_MSG_NO DECIMAL(31, 0) DEFAULT 1 NOT NULL, Derby doesn't care about the order of NOT NULL / DEFAULT, Oracle does from EXPIRY BIGINT, to EXPIRY DECIMAL(31, 0), Oracle doesn't have BIGINT. DECIMAL is the one data type that seems to work across different DBs. from X0Y32.equals(ex.getSQLState()) to X0Y32.equals(ex.getSQLState()) || 955 == ex.getErrorCode() When checking for table creation (955 is the Oracle error code if the table already exists). Is this something that should be committed into the repository? -- Daniel Kulp dk...@apache.org http://www.dankulp.com/blog
Seeming JAX-RS mistake
This can't work right for generic types (like collections), since it doesn't use java.lang.reflect.Type. public interface ContextResolverT { /** * Get a context of type codeT/code that is applicable to the supplied * type. * @param type the class of object for which a context is desired * @return a context for the supplied type or codenull/code if a * context for the supplied type is not available from this provider. */ T getContext(Class? type); }