I am eagerly waiting for solid data on this. We need many datapoints: from no parameter and void in return, up to 10,000k in parameter 10,000k in return and many points in between, average time on byte[] vs MO there. mo.get() seems to be a very very expensive operation (for some reason).
marcf |-----Original Message----- |From: Francisco Reverbel [mailto:[EMAIL PROTECTED]] |Sent: Thursday, November 22, 2001 9:26 AM |To: marc fleury |Cc: David Jencks; [EMAIL PROTECTED] |Subject: RE: [JBoss-dev] Classloader puzzle | | |Hi, | |On Wed, 21 Nov 2001, marc fleury wrote: | |> |It sure looks like the system (Service libraries) rather than |application |> |classloader is being used, despite your code that clearly sets the |> |contextClassLoader to the container's. |> |> And some people wonder why I loved having Rickard Oberg on the |project back |> in EJBoss 0.9 |> |> attached is a ObjectinputStreamWithClassLoader that rickard and monson |> haefel put together one night when I ran into the same problem you did |> Francisco :) |> |> Rickard Oberg was a fucking genius make no mistakes about it ... |> |> Francisco to use this you want to create the ObjectInputStream with this |> puppy and pass the context cl as the parameter to the constructor. Then |> call readObject as you did. | |This is great! | |> |> BTw don't commit this, MarshalledObject operates pretty much the |same way. | |I might want to discuss this... ObjectInputStreamWithClassLoader might the |way of cutting down bean IOR sizes. I'll get back after I try it. | |Best, | |Francisco | |> |> marcf |> |> | |> |Have you checked that poaCurrent doesn't change the context |classloader and |> |the classloader is really what you expect when you call toObject? |> | |> |david jencks |> | |> | |> |On 2001.11.21 13:04:25 -0500 Francisco Reverbel wrote: |> |> Hi, |> |> |> |> I have a classloader problem in the IIOP container invoker |code and would |> |> really appreciate some helpful hints... Classloader wizards: to |> |> understand |> |> the problem you do not have to know anything about |CORBA/IIOP, so please |> |> read on! |> |> |> |> The problem only shows up with entity beans. When the IIOP container |> |> invoker creates a CORBA reference for such a bean, it embeds |the beanīs |> |> primary key into the CORBA reference. When handling a bean invocation, |> |> the |> |> container invoker extracts the PK from the CORBA reference and passes |> |> this |> |> PK to the container as the first argument of a MethodInvocation. |> |> |> |> A ClassNotFoundException is thrown when the container invoker |attempts to |> |> extract the PK from the CORBA reference (got the stack trace below |> |> running |> |> RMH's well-known cabin bean): |> |> |> |> java.lang.ClassNotFoundException: com.titan.cabin.CabinPK |> |> at |> |org.jboss.system.ServiceLibraries.loadClass(ServiceLibraries.java:372) |> |> at |> |org.jboss.system.URLClassLoader.loadClass(URLClassLoader.java:111) |> |> at java.lang.ClassLoader.loadClass(ClassLoader.java:253) |> |> at java.lang.ClassLoader.loadClassInternal(ClassLoader.java:313) |> |> at java.lang.Class.forName0(Native Method) |> |> at java.lang.Class.forName(Class.java:195) |> |> at |> |java.io.ObjectInputStream.resolveClass(ObjectInputStream.java:654) |> |> at |> ||java.io.ObjectInputStream.inputClassDescriptor(ObjectInputStream.java:918) |> |> at java.io.ObjectInputStream.readObject(ObjectInputStream.java:366) |> |> at java.io.ObjectInputStream.readObject(ObjectInputStream.java:236) |> |> at |> |java.io.ObjectInputStream.inputObject(ObjectInputStream.java:1186) |> |> at java.io.ObjectInputStream.readObject(ObjectInputStream.java:386) |> |> at java.io.ObjectInputStream.readObject(ObjectInputStream.java:236) |> |> at org.jboss.ejb.CacheKey.readExternal(CacheKey.java:118) |> |> at |> |java.io.ObjectInputStream.inputObject(ObjectInputStream.java:1212) |> |> at java.io.ObjectInputStream.readObject(ObjectInputStream.java:386) |> |> at java.io.ObjectInputStream.readObject(ObjectInputStream.java:236) |> |> at |> |org.jboss.ejb.plugins.iiop.server.IIOPContainerInvoker.toObject(IIO |> |PContainerInvoker.java:896) |> |> at |> |org.jboss.ejb.plugins.iiop.server.IIOPContainerInvoker._invoke(IIOP |> |ContainerInvoker.java:665) |> |> at |> ||org.jacorb.poa.RequestProcessor.invokeOperation(RequestProcessor.java:239) |> |> at |> |org.jacorb.poa.RequestProcessor.process(RequestProcessor.java:456) |> |> at org.jacorb.poa.RequestProcessor.run(RequestProcessor.java:581) |> |> |> |> I've found a way of circumventing the problem, but am not |satisfied with |> |> it. Things work if I wrap the PK into a MarshalledObject |before embedding |> |> it into the CORBA reference, and do the opposite thing when extracting |> |> the |> |> PK from the reference. Runs fine, but is very slow... The |bank test takes |> |> forever to run. |> |> |> |> This code is in |org.jboss.ejb.plugins.iiop.server.IIOPContainerInvoker, |> |> under contrib/iiop. Insertion/extraction of the PK is done through the |> |> methods toByteArray() and toObject(). Each of these methods has two |> |> implementations, one of them commented out. The commented out |> |> implementation "solves" the problem in the way described above. |> |> |> |> Suggestions are very welcome! |> |> |> |> Cheers, |> |> |> |> Francisco |> |> |> |> |> |> |> |> |> |> |> |> |> |> |> |> |> |> _______________________________________________ |> |> Jboss-development mailing list |> |> [EMAIL PROTECTED] |> |> https://lists.sourceforge.net/lists/listinfo/jboss-development |> |> |> |> |> | |> |_______________________________________________ |> |Jboss-development mailing list |> |[EMAIL PROTECTED] |> |https://lists.sourceforge.net/lists/listinfo/jboss-development |> | _______________________________________________ Jboss-development mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/jboss-development
