This Bugzilla instance is a read-only archive of historic NetBeans bug reports. To report a bug in NetBeans please follow the project's instructions for reporting issues.
dev build 200505171800 Mustang b36 When I run the memory profiler I get tons (every 5 seconds or so) of these errors: *** JFluid: redefinition failed for classes: <some class name> *** with message: Redefinition failed with error 62 Any ideas?
Gili, can you please provide more info about the application being profiled? J2SE, J2EE, what appserver, etc. Also, if possible, attach the IDE log file. Thanks.
I no longer have an IDE log because it's been a while since I originally reported this issue. To my recollection, I was trying to profile code running inside Tomcat 5.5.9. Let me know whether you can reproduce the problem or not.
We cannot reproduce this issue. Let's see if M7 due out next week still has this problem.
Created attachment 23994 [details] NetBeans message.log
I also experienced this problem, Windows XP, NetBeans 4.1, NB Profiler m8v2, profiling a J2EE application running on Caucho Resin 3.0.14 Open Source app server. I'm running memory liveness, with record stack trace, track every 10 allocations. After about 5 hours of profiling, while having the telemetry window / heap tab active, the IDE started to flood me with error popups: "Redefinition failed with error 62 Check JVMTI documentation for this error code." I attached the relevant section from messages.log
Thanks for providing details. Is this on JDK 5.0 update 4?
Yes, JDK 1.5.0_04 for the IDE and the profiled VM.
I get this error immediately after starting memory profiling on an external Tomcat 5.5 server under JDK 1.5.0_04 and Netbeans dev build 200508171800. I get flooded with tons of these warnings and eventually the IDE goes into some sort of deadlock and has to be killed. Anyway, I've yet to get the profiler working on my machine :) and it's been months since I've begun trying.
I'm attaching an extracted IDE log (NB5.0 beta, profiler M9, Java 1.5.0_04-b05) showing the same problem. I thought that running NetBeans with bytecode verification off should help, but it didn't. My netbeans.conf: netbeans_default_userdir="${HOME}/.netbeans/5.0beta" netbeans_default_options="-J-Xms256m -J-Xmx256m -J-XX:PermSize=64m -J-XX:MaxPermSize=64m -J-XX:NewRatio=3 -J-Xverify:none -J-ea" netbeans_jdkhome="D:\java5" The method I was trying to profile contains ILOG JViews method calls. JViews' jar file is obfuscated bytecode, by the way.
Created attachment 26303 [details] IDE messages log featuring profiler JVMTI error 62
We've fixed the memory profiling problems in M10, the issue was caused by the fact that the bytecode generated by the JSP compiler in Tomcat/JBoss is slightly different than that produced by javac. In Milestone 10 this will work correctly. For the last error added by bruckner, I will file a separate issue.
Verification of old issues.
Closing old issues.