>>>>>ShortDescription The jmx ObjectName.checkComponents fails off the standard sun reference implementation (jmxri). I'm not sure which of the validation checks is causing the failure. Have been using the jmxri for several years in a variety of projects with no problems with it. >>>HowToReproduce The following lines are enough to produce the failure (1.2.1 jmxri.jar dependency http://java.sun.com/javase/technologies/core/mntr-mgmt/javamanagement/) >>>Code System.out.print("__________________\n"); System.out.print("System.class.forName(\"com.sun.jmx.mbeanserver.RepositorySupport\");" + "\n"); System.class.forName("com.sun.jmx.mbeanserver.RepositorySupport"); >>>Output __________________ System.class.forName("com.sun.jmx.mbeanserver.RepositorySupport"); Exception in thread "main" java.lang.ExceptionInInitializerError at cli.System.Runtime.CompilerServices.RuntimeHelpers.RunClassConstructor(Unknown Source) at IKVM.NativeCode.java.lang.VMClass.forName0(Unknown Source) at java.lang.VMClass.forName0(VMClass.java) at java.lang.VMClass.forName(VMClass.java:240) at java.lang.Class.forName(Class.java:189) at com.teamcenter.fms.servercache.FMSServerCache.forceClassLoad(FMSServerCache.java:202) at com.teamcenter.fms.servercache.FMSServerCache.main(FMSServerCache.java:160) Caused by: java.lang.NullPointerException at javax.management.ObjectName.checkComponents(ObjectName.java:246) at javax.management.ObjectName.<init>(ObjectName.java:178) at com.sun.jmx.mbeanserver.RepositorySupport.<clinit>(RepositorySupport.java:49) >>>>>Environment D:\ikvm>ikvm -version IKVM.NET Launcher version 0.32.0.0 Copyright (C) 2002-2006 Jeroen Frijters http://www.ikvm.net/ CLR version: 1.1.4322.573 (32 bit) mscorlib: 1.0.5000.0 ikvm: 0.32.0.0 IKVM.GNU.Classpath: 0.32.0.0 IKVM.Runtime: 0.32.0.0 GNU Classpath version: 0.93
I can confirm this bug on 0.93, but it appears to be fixed in cvs already. Could you please try this snapshot: http://www.frijters.net/ikvmbin-20070321.zip
Thanx, very helpful This compiles with many warnings (see attachment). It also starts up all the classes (see attachment). -------- Question #1 It appears that you fixed this by a class loader change. It appears that the base classpath gnu java code is still bad, and that you have not fixed that (please confirm).
Created attachment 13253 [details] Warnings from the compile jmxri vs classpath
Created attachment 13254 [details] Good start up / load for FSC process via ikvm
I did not change the class loading, I just added the warning. Does this means that this issue is fixed?
Yes, appears to work now on my box.
Thanks. Closing the bug.