> (The bootclasspath hackery which the jaco script normally does doesn't work
> with gij, at least not for me.) At least the name server appears to start up
> that way, I haven't done any further testing yet.
Unfortunately, value types don't work when doing things this way due to some
type conflict between the libgcj and the JacORB classes:
Exception in thread "main" java.lang.NoClassDefFoundError:
gnu.javax.rmi.CORBA.RmiUtilities
at java.lang.Class.initializeClass(libgcj.so.7rh)
at java.lang.Class.initializeClass(libgcj.so.7rh)
at java.lang.Class.forName(libgcj.so.7rh)
at gnu.CORBA.ObjectCreator.forName(libgcj.so.7rh)
at gnu.javax.rmi.CORBA.DelegateFactory.getInstance(libgcj.so.7rh)
at javax.rmi.CORBA.Util.<clinit>(libgcj.so.7rh)
at java.lang.Class.initializeClass(libgcj.so.7rh)
at org.jacorb.util.ValueHandler.getRMIRepositoryID(ValueHandler.java:10)
at org.jacorb.orb.CDROutputStream.write_value(CDROutputStream.java:2099)
Caused by: java.lang.VerifyError: Incompatible types: In class
gnu.javax.rmi.CORBA.RmiUtilities: org.omg.CORBA.WStringValueHelper is not
assignable to org.omg.CORBA.portable.BoxedValueHelper
at java.lang.Class.initializeClass(libgcj.so.7rh)
...10 more
Something somewhere probably needs to be rebuilt, but if I rebuild
gnu.javax.rmi.CORBA.RmiUtilities and put it in a jar in the gcj-endorsed
directory, I get "java.lang.VerifyError: verification failed at PC 92 in
gnu.javax.rmi.CORBA.RmiUtilities:writeFields((Lorg.omg.CORBA.portable.
OutputStream;Ljava.io.Serializable;)V):
incompatible type on stack" instead, and I couldn't get that to go away.
I don't feel like hunting this mess any further right now. It works on the
university machine using Sun Java.
Kevin Kofler
--
fedora-devel-java-list mailing list
fedora-devel-java-list@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/fedora-devel-java-list
