[ http://jira.jboss.com/jira/browse/JASSIST-28?page=comments#action_12345473 ] Armin Haaf commented on JASSIST-28: ----------------------------------- Scenario is: - business logic on jboss with hibernate as or mapper. EJB3 Beans, but not EJB3 persistence, instead use hibernate directly. - client code (swing application) which executes a business method, return value is a hibernate entity with maybe uninitialized proxies (eg a many-to-one mapping with lazy fetching) Give me some time (1 or 2 days), I will come up with a test case. -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: http://jira.jboss.com/jira/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira

[ http://jira.jboss.com/jira/browse/JASSIST-28?page=comments#action_12345512 ] Shigeru Chiba commented on JASSIST-28: -------------------------------------- Thank you for the test case. We (Muga Nishizawa and I) found that the problem is due to a difference in their caching scheme between Cglib and Javassist. Cglib records all proxy classes that have been generated but Javassist does not for saving memory consumption (or avoiding memory leak). Since your test code repeatedly creates the same proxy class, Javassist creates the same class more than once and attempts to load them. The loading is rejected by the JVM and an exception is thrown. We will fix this problem soon (when we find the best caching scheme for proxy classes). Thank you for your bug report and help! -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: http://jira.jboss.com/jira/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira

[ http://jira.jboss.com/jira/browse/JASSIST-28?page=comments#action_12346068 ] Armin Haaf commented on JASSIST-28: ----------------------------------- I'll check it against our performance suite till the end of the week -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: http://jira.jboss.com/jira/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira

[ http://jira.jboss.com/jira/browse/JASSIST-28?page=comments#action_12346139 ] Armin Haaf commented on JASSIST-28: ----------------------------------- I've checked the current head and checked it against our performance suite. The previous is fixed, but on heavy concurrent load the system hangs up and is only killabled by "kill -9". Check it with the attached script, and replace testStartClient with the following code. public void testStartClients() throws Exception { startClients(100); startClients(100); } public void startClients(int pCount) throws Exception { Thread[] tClientThreads = new Thread[pCount]; On the second iteration it hangs (I check that only on a linux platform, witth 1.5.0_08-b03 Java HotSpot(TM) Client VM) -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: http://jira.jboss.com/jira/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira

[ http://jira.jboss.com/jira/browse/JASSIST-28?page=comments#action_12346151 ] Armin Haaf commented on JASSIST-28: ----------------------------------- It did not run better without caching. And it is no multithread problem. I've run the attached code in a single thread, but with a big iteration count 20000 instead of 100. After about 16000 iterations the process hangs (with default Xmx and with a reduced Xmx16M). So this number seems like a size constraint of some object inside javassist (a guess is a 4byte object/reference in a list/array, where the list exceeds the maximum size) -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: http://jira.jboss.com/jira/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira

[ http://jira.jboss.com/jira/browse/JASSIST-28?page=comments#action_12346182 ] Muga Nishizawa commented on JASSIST-28: --------------------------------------- Hi Armin, please, tell me your experimental platform, in particular, CPU and OS (which linux do you use?). -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: http://jira.jboss.com/jira/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira

[ http://jira.jboss.com/jira/browse/JASSIST-28?page=all ] Armin Haaf updated JASSIST-28: ------------------------------ Attachment: TestJavassist.tgz a complete test environment -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: http://jira.jboss.com/jira/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira

[ http://jira.jboss.com/jira/browse/JASSIST-28?page=comments#action_12346218 ] Muga Nishizawa commented on JASSIST-28: --------------------------------------- Great, I will use the new program and test the behavior of javassist or cglib soon. Thanks for your help, -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: http://jira.jboss.com/jira/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira

[ http://jira.jboss.com/jira/browse/JASSIST-28?page=comments#action_12346303 ] Armin Haaf commented on JASSIST-28: ----------------------------------- Hi Muga, you version of javassist did not work at all (see below). I'm not sure, If this is a problem with java 5 [java] SCHWERWIEGEND: Javassist Enhancement failed: Server$TestEntity [java] java.lang.Error: Unresolved compilation problems: [java] The import javassist.CannotCompileException cannot be resolved [java] The import javassist.bytecode cannot be resolved [java] ClassFile cannot be resolved to a type [java] The method make() is undefined for the type ProxyFactory [java] CannotCompileException cannot be resolved to a type [java] e cannot be resolved [java] e cannot be resolved [java] ClassFile cannot be resolved to a type [java] CannotCompileException cannot be resolved to a type [java] CannotCompileException cannot be resolved to a type [java] ClassFile cannot be resolved to a type [java] ClassFile cannot be resolved to a type [java] AccessFlag cannot be resolved [java] ConstPool cannot be resolved to a type [java] FieldInfo cannot be resolved to a type [java] FieldInfo cannot be resolved to a type [java] AccessFlag cannot be resolved [java] AccessFlag cannot be resolved [java] FieldInfo cannot be resolved to a type [java] FieldInfo cannot be resolved to a type [java] AccessFlag cannot be resolved [java] ClassFile cannot be resolved to a type [java] ClassFile cannot be resolved to a type [java] ConstPool cannot be resolved to a type [java] CannotCompileException cannot be resolved to a type [java] FieldInfo cannot be resolved to a type [java] FieldInfo cannot be resolved to a type [java] AccessFlag cannot be resolved [java] AccessFlag cannot be resolved [java] MethodInfo cannot be resolved to a type [java] MethodInfo cannot be resolved to a type [java] Bytecode cannot be resolved to a type [java] Bytecode cannot be resolved to a type [java] Bytecode cannot be resolved [java] ClassFile cannot be resolved to a type [java] ConstPool cannot be resolved to a type [java] CannotCompileException cannot be resolved to a type [java] MethodInfo cannot be resolved to a type [java] MethodInfo cannot be resolved to a type [java] AccessFlag cannot be resolved [java] Bytecode cannot be resolved to a type [java] Bytecode cannot be resolved to a type [java] Bytecode cannot be resolved [java] ClassFile cannot be resolved to a type [java] ConstPool cannot be resolved to a type [java] CannotCompileException cannot be resolved to a type [java] CannotCompileException cannot be resolved to a type [java] CannotCompileException cannot be resolved to a type [java] MethodInfo cannot be resolved to a type [java] ConstPool cannot be resolved to a type [java] MethodInfo cannot be resolved to a type [java] MethodInfo cannot be resolved to a type [java] Bytecode cannot be resolved to a type [java] Bytecode cannot be resolved to a type [java] Opcode cannot be resolved [java] Opcode cannot be resolved [java] Opcode cannot be resolved [java] Opcode cannot be resolved [java] MethodInfo cannot be resolved to a type [java] ConstPool cannot be resolved to a type [java] MethodInfo cannot be resolved to a type [java] MethodInfo cannot be resolved to a type [java] Bytecode cannot be resolved to a type [java] Bytecode cannot be resolved to a type [java] MethodInfo cannot be resolved to a type [java] ConstPool cannot be resolved to a type [java] MethodInfo cannot be resolved to a type [java] MethodInfo cannot be resolved to a type [java] Descriptor cannot be resolved [java] Bytecode cannot be resolved to a type [java] Bytecode cannot be resolved to a type [java] Opcode cannot be resolved [java] Opcode cannot be resolved [java] Opcode cannot be resolved [java] Opcode cannot be resolved [java] MethodInfo cannot be resolved to a type [java] ConstPool cannot be resolved to a type [java] MethodInfo cannot be resolved to a type [java] ConstPool cannot be resolved to a type [java] ExceptionsAttribute cannot be resolved to a type [java] ExceptionsAttribute cannot be resolved to a type [java] Bytecode cannot be resolved to a type [java] Bytecode cannot be resolved to a type [java] Bytecode cannot be resolved to a type [java] Opcode cannot be resolved [java] Opcode cannot be resolved [java] Opcode cannot be resolved [java] Opcode cannot be resolved [java] Opcode cannot be resolved [java] Opcode cannot be resolved [java] Bytecode cannot be resolved to a type [java] Opcode cannot be resolved [java] Opcode cannot be resolved [java] Bytecode cannot be resolved to a type [java] Opcode cannot be resolved [java] Bytecode cannot be resolved to a type [java] Opcode cannot be resolved [java] Opcode cannot be resolved [java] Bytecode cannot be resolved to a type [java] Opcode cannot be resolved [java] at javassist.util.proxy.ProxyFactory.<init>(ProxyFactory.java:32) [java] at org.hibernate.proxy.pojo.javassist.JavassistLazyInitializer.getProxyFactory(JavassistLazyInitializer.java:137) [java] at org.hibernate.proxy.pojo.javassist.JavassistProxyFactory.postInstantiate(JavassistProxyFactory.java:42) [java] at org.hibernate.tuple.entity.PojoEntityTuplizer.buildProxyFactory(PojoEntityTuplizer.java:162) [java] at org.hibernate.tuple.entity.AbstractEntityTuplizer.<init>(AbstractEntityTuplizer.java:135) [java] at org.hibernate.tuple.entity.PojoEntityTuplizer.<init>(PojoEntityTuplizer.java:55) [java] at org.hibernate.tuple.entity.EntityEntityModeToTuplizerMapping.<init>(EntityEntityModeToTuplizerMapping.java:56) [java] at org.hibernate.tuple.entity.EntityMetamodel.<init>(EntityMetamodel.java:269) [java] at org.hibernate.persister.entity.AbstractEntityPersister.<init>(AbstractEntityPersister.java:425) [java] at org.hibernate.persister.entity.SingleTableEntityPersister.<init>(SingleTableEntityPersister.java:109) [java] at org.hibernate.persister.PersisterFactory.createClassPersister(PersisterFactory.java:55) [java] at org.hibernate.impl.SessionFactoryImpl.<init>(SessionFactoryImpl.java:226) [java] at org.hibernate.cfg.Configuration.buildSessionFactory(Configuration.java:1218) [java] at Server.initHibernate(Server.java:54) [java] at Server.startServer(Server.java:58) [java] at Server.main(Server.java:126) [java] at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method) [java] at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:39) [java] at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:25) [java] at java.lang.reflect.Method.invoke(Method.java:585) [java] at org.apache.tools.ant.taskdefs.ExecuteJava.run(ExecuteJava.java:193) [java] at org.apache.tools.ant.taskdefs.ExecuteJava.execute(ExecuteJava.java:130) [java] at org.apache.tools.ant.taskdefs.Java.run(Java.java:705) [java] at org.apache.tools.ant.taskdefs.Java.executeJava(Java.java:177) [java] at org.apache.tools.ant.taskdefs.Java.execute(Java.java:83) [java] at org.apache.tools.ant.UnknownElement.execute(UnknownElement.java:275) [java] at org.apache.tools.ant.Task.perform(Task.java:364) [java] at org.apache.tools.ant.Target.execute(Target.java:341) [java] at org.apache.tools.ant.Target.performTasks(Target.java:369) [java] at org.apache.tools.ant.Project.executeTarget(Project.java:1214) [java] at org.apache.tools.ant.Project.executeTargets(Project.java:1062) [java] at org.apache.tools.ant.Main.runBuild(Main.java:673) [java] at org.apache.tools.ant.Main.startAnt(Main.java:188) [java] at org.apache.tools.ant.launch.Launcher.run(Launcher.java:196) [java] at org.apache.tools.ant.launch.Launcher.main(Launcher.java:55) -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: http://jira.jboss.com/jira/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira

[ http://jira.jboss.com/jira/browse/JASSIST-28?page=comments#action_12346358 ] Armin Haaf commented on JASSIST-28: ----------------------------------- It did not work! Same problem, client process hangs - even ealier - -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: http://jira.jboss.com/jira/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira

[ http://jira.jboss.com/jira/browse/JASSIST-28?page=all ] Shigeru Chiba updated JASSIST-28: --------------------------------- Attachment: JavassistLazyInitializer.java Hibernate fix. -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: http://jira.jboss.com/jira/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira

[ http://jira.jboss.com/jira/browse/JASSIST-28?page=comments#action_12346367 ] Shigeru Chiba commented on JASSIST-28: -------------------------------------- It turned out that we must modify a source file of Hibernate. JavassistLazyInitializer.java is the new source file that has to be substituted. Hibernate with that source file and the latest snapshot of Javassist shows as good performance as Hibernate+cglib. However, both cglib and Javassist cannot finish your test code; both of them hang up. -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: http://jira.jboss.com/jira/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira

[ http://jira.jboss.com/jira/browse/JASSIST-28?page=comments#action_12346369 ] Armin Haaf commented on JASSIST-28: ----------------------------------- I checked the supplied LazyInitializer with hibernate-3.2.0GA and the performance is good now, even a little bit faster than cglib. And the test cases succeed now without any problem. Thats good news. -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: http://jira.jboss.com/jira/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira

[ http://jira.jboss.com/jira/browse/JASSIST-28?page=comments#action_12346370 ] Shigeru Chiba commented on JASSIST-28: -------------------------------------- My previous comment was wrong. Both Javassist and cglib finsih the test code under your work load (loop count: 200) although they could not finish under much heavier work load (e.g. 10,000 on Apple Xserve G5). -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: http://jira.jboss.com/jira/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira

[ http://jira.jboss.com/jira/browse/JASSIST-28?page=comments#action_12346373 ] Armin Haaf commented on JASSIST-28: ----------------------------------- This might be a problem in the attached test environment, there is a memory leak. I'm checking it and will comment in a few minutes -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: http://jira.jboss.com/jira/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira

[ http://jira.jboss.com/jira/browse/JASSIST-28?page=all ] Shigeru Chiba resolved JASSIST-28. ---------------------------------- Resolution: Done To really solve this problem, JavassistLazyInitializer.java must be substituted for org/hibernate/proxy/pojo/javassist/JavassistLazyInitializer.java and Javassist 3.4 must be released. -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: http://jira.jboss.com/jira/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira