Red Hat Bugzilla – Bug 780751
Wise SOAPClient SOAPFaultException lost
Last modified: 2011-11-03 07:32:58 EDT
Help Desk Ticket Reference: https://c.na7.visual.force.com/apex/Case_View?id=500A0000007BBvM&sfdc.override=1
Steps to Reproduce: 1. Deploy the attached war file on SOA-P 5.1
2. Unzip the webservice_consumer_wise_smooks.zip into the quickstarts directory
3. ant deploy
4. ant runtest
Wise will allow expected exceptions to be returned as an InvocationResult but will trap anything it does not recognise and throw it back to the invoker as an exception. This terminates the pipeline and the remainder of the pipeline is prevented from processing the *valid* SOAP response because wise is turning it into an exception, thereby treating it in a different manner from the exceptions that are defined within the WSDL (and expected). It is still visible through the exception processing but shouldn't be, it should be treated in the same manner as other SOAP faults.
Attachment: Added: HelloWS.jar
Attachment: Added: webservice_consumer_wise_smooks.zip
Link: Added: This issue is related to JBESB-3643
This is an updated version fixing a small build.xml bug that would lose the build/META-INF/smooks*.xml files if you did a clean.
Link: Added: This issue relates to SOA-3227
The fix for this has been committed, but it seems as if there's a regression caused by the Smooks upgrade from ESB 4.9->ESB 4.10. PropertyChecker.java is not finding the generated bean class the first time the quickstart is run, but any subsequent time, this works fine.
Link: Added: This issue is related to JBESB-3666
So, is the issue associated with this JIRA fixed then? Sounds like Tom is saying it is, but something else (unrelated to the fix made for this JIRA) is broken because of the upgrade to Smooks v1.4, right? Or is the regression specifically related to the fix made for this JIRA? If so, do we know if it fails or not with the previous version of Smooks?
Tom... can you provide more info on the regression please? Which test is failing etc?
This may be related to SOA-3321, it would be worth retesting
The classloading appears to have been addressed by the fix for SOA-3321, the quickstart ran correctly.
One problem I noticed us that the fix is passing information between the SOAPClient and aspect using a system property, which is not safe. We need to fix this part.
Link: Added: This issue depends JBESB-3678
Fixed the system property issue, replaced with ThreadLocal.
Release Notes Docs Status: Added: Documented as Resolved Issue
Writer: Added: dlesage
Release Notes Text: Added: https://issues.jboss.org/browse/SOA-3205
Release Notes Text: Removed: https://issues.jboss.org/browse/SOA-3205 Added: https://issues.jboss.org/browse/SOA-3205
If Wise encountered an error it could not handle, it would terminate the pipeline. A code fix has been applied so that exceptions are now handled in a consistent manner.
Verified in 5.2.ER6