Doug Schnelzer wrote:
I'm getting the following error running the latest Jetspeed2 from CVS
under JBoss 4.0.1sp1 in the second and subsequent portlets in the
portal. The first portlet and the portal display fine.
The error int the portlet window is:
VelocityViewServlet : Error processing the template
Invocation of method 'getTitle' in class
org.apache.jetspeed.velocity.JetspeedPowerTool threw exception class
java.lang.IllegalStateException : No classloader has been defined for
portlet application 2
java.lang.IllegalStateException: No classloader has been defined for
portlet application 2
Do you think this issue is related to this thread and the refactoring
that Ate is doing? If so, any recommendations on a workaround in
JBoss or do I just need to sit tight and wait till JS2-210 is
committed?
Doug,
You *can* try the deployment_refactoring branch I created for JS2-210.
It is fully functional under Tomcat 5.0.30 and 5.5.8.
But, I didn't test it out with JBoss yet so it might just work or not at all.
If you would have time to try it out though I would very much appreciate it.
I do expect it to be easier to get working with other servers though as my
refactored
solution only relies on the server its own handling of web application
deployment
conforming to the servlet specs.
As I already mentioned before, I would like to request a vote for merging my
branch with CVS HEAD later this week. What does worry me though is that nobody
seems
to have time or interest in testing this thing out...
Ate
Thanks for the help,
- Doug
On Tue, 01 Mar 2005 14:29:30 -0800, David Sean Taylor
<[EMAIL PROTECTED]> wrote:
Peter Meier wrote:
I noticed that the classloader that is stored in PortletDefinition under
portletClassLoader is *not* the classloader that is actually used (or
has been used) to load the portlet class itself. Instead, there is a
classloader stored that has been taken from a map that has been
previously created. Although in this case the URL classloader is
pointing to the right class path, it could cause ClassCastExceptions if
it used to create classes instead of the classloader from the portlet
class.
My question is what was the initial intention of having a classloader
map, when it is not used for portlet class loading itself?
Is there any objection if the code in JetspeedPortletFactory is altered
thus that it *always* stores the actual classloader of the portlet in
portletClassLoader of Portletdefinition as the name suggests?
Ate's upcoming commit will fix that
http://issues.apache.org/jira/browse/JS2-210
By doing the latter, I also avoid classloader issues in JBoss.
Regards,
Peter
Architect,
TEC
Wellington, New Zealand
email: [EMAIL PROTECTED]
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
--
David Sean Taylor
Bluesunrise Software
[EMAIL PROTECTED]
[office] +01 707 773-4646
[mobile] +01 707 529 9194
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
