It seems my understanding of how classloaders work is not as thorough as it could be. From what I can see, juli doesn't include a UniqueId class, so I'm not sure how we could get a conflict there. Could you explain how that could be happening? I could take a look at it for a bit today, though I think it'd be good to have Gianny weigh in when he returns.
On Fri, Jun 27, 2008 at 9:32 AM, Kevan Miller <[EMAIL PROTECTED]> wrote: > > On Jun 25, 2008, at 12:05 PM, Jason Warner wrote: > > I took a look at the classloader being used. It only contains one instance > of UniqueId in any of the parent classloaders. I took a quick look at the > changes between Wadi 2.0-M10 and 2.0, and there is indeed a significant > amount of change. In fact, this error stems from a class that didn't even > exist in 2.0-M10. I'll keep looking to see if I can figure out what's going > on, but any suggestions of other possible issues would definitely not be > ignored :) > > > I'm seeing two ClassLoaders in play when this error is occurring: > > org.apache.geronimo.configs/wadi-clustering/2.1.2-SNAPSHOT/car > org.apache.geronimo.configs/tomcat6/2.1.2-SNAPSHOT/car > > The following dependencies in plugins/clustering/wadi-clustering/pom.xml > would seem to be a problem: > > <dependency> > <groupId>org.apache.tomcat</groupId> > <artifactId>juli</artifactId> > </dependency> > > <dependency> > <groupId>org.apache.tomcat</groupId> > <artifactId>tribes</artifactId> > </dependency> > > They can cause us to load tribes or juli classes within two different > ClassLoaders and result in the ClassCastException that you're experiencing. > > This will need to be straightened out. I haven't looked at how to > straighten this out. You want to look? Or Gianny should be back soon... > > --kevan > > -- ~Jason Warner
