]
Mark Little updated JBESB-583:
------------------------------
Component/s: Web Services
Investigate the classpath issue re StAX and inclusion of
jboss-messaging-client.jar in the classpath
----------------------------------------------------------------------------------------------------
Key: JBESB-583
URL:
http://jira.jboss.com/jira/browse/JBESB-583
Project: JBoss ESB
Issue Type: Task
Security Level: Public(Everyone can see)
Components: Web Services
Affects Versions: 4.2 Milestone Release 2
Reporter: Tom Fennelly
Assigned To: Tom Fennelly
Fix For: 4.2
The SOAPClient supports using XStream to deserialize messages. In order to support
namespaced messages, we have to use the StaxDriver for XStream. This means including a
StAX impl in the classpath. We've included Woodstox. The StAX impl is picked up via
a factory class that uses a System property. The System prop is set via services defs in
the impl jar i.e. the woodstox jar set the impl via a services def.
So the first problem is:
When jboss-messaging-client.jar is on the classpath, the stax core is not able to find
the impl via the System property. Remove jboss-messaging-client.jar from the
classpath(and replace with jboss-remoting.jar) and all is fine for those tests.
This however causes the following mysterious issue:
When the jboss-message-client.jar is removed from the classpath and replaced with
jboss-remoting.jar, the JBossRemotingGatewayListener is unable to start because it
can't find the "org.jboss.remoting.transport.coyote.CoyoteInvoker" (or
something like that). This is totally bizarre because I've checked the classpath many
times and the class is def there. It is picking up the jar because I can assert that
other classes from the same jar are visible. I did a number of tests on this, all based
on removing everything else from the classpath and only having the jboss-remoting.jar
file.
Need to get to the bottom of these classpath issues. I've gotten around (for now) it
in the tests by running the unit tests twice in the build and modifying the classpath on
both runs. Don't worry, I exclude and include so as we're not running the same
tests twice.
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: