Community
Participate
Working Groups
To reproduce, follow the instructions below: - Download AC on a Linux IA32 machine - Unzip AC in a directory with a very long path. Example: /home/navid/TPTP_Drivers/TPTP-4.4.0-200704191232/eclipse/plugins/org.eclipse.tptp.platform.ac.linux_ia32_4.4.0.v200704191232/agent_controller - Run SetConfig.sh - Try to start AC with the RAStat.sh command. When I do this on my Debian box, I get a Java core dump: JVMDUMP006I Processing Dump Event "abort", detail "" - Please Wait. JVMDUMP007I JVM Requesting System Dump using '/home/navid/TPTP_Drivers/TPTP-4.4.0-200704191232/eclipse/plugins/org.eclipse.tptp.platform.ac.linux_ia32_4.4.0.v200704191232/agent_controller/bin/core.20070423.110054.5154.dmp' JVMDUMP010I System Dump written to /home/navid/TPTP_Drivers/TPTP-4.4.0-200704191232/eclipse/plugins/org.eclipse.tptp.platform.ac.linux_ia32_4.4.0.v200704191232/agent_controller/bin/core.20070423.110054.5154.dmp JVMDUMP007I JVM Requesting Snap Dump using '/home/navid/TPTP_Drivers/TPTP-4.4.0-200704191232/eclipse/plugins/org.eclipse.tptp.platform.ac.linux_ia32_4.4.0.v200704191232/agent_controller/bin/Snap0001.20070423.110054.5154.trc' JVMDUMP010I Snap Dump written to /home/navid/TPTP_Drivers/TPTP-4.4.0-200704191232/eclipse/plugins/org.eclipse.tptp.platform.ac.linux_ia32_4.4.0.v200704191232/agent_controller/bin/Snap0001.20070423.110054.5154.trc JVMDUMP007I JVM Requesting Java Dump using '/home/navid/TPTP_Drivers/TPTP-4.4.0-200704191232/eclipse/plugins/org.eclipse.tptp.platform.ac.linux_ia32_4.4.0.v200704191232/agent_controller/bin/javacore.20070423.110054.5154.txt' ACServer started successfully. navidm:/home/navid/TPTP_Drivers/TPTP-4.4.0-200704191232/eclipse/plugins/org.eclipse.tptp.platform.ac.linux_ia32_4.4.0.v200704191232/agent_controller/bin# JVMDUMP010I Java Dump written to /home/navid/TPTP_Drivers/TPTP-4.4.0-200704191232/eclipse/plugins/org.eclipse.tptp.platform.ac.linux_ia32_4.4.0.v200704191232/agent_controller/bin/javacore.20070423.110054.5154.txt JVMDUMP013I Processed Dump Event "abort", detail "". I'll attach the core dump file for your analysis. When I do the same on paradox-rh21as (Redhat box), I get the following: ./RAStart.sh: line 56: 2495 Segmentation fault ACServer $* ACServer failed to start. Note that everything works fine if I use a shorter path. This defect is blocking further work on the new IAC for Linux IA32 (bugzilla 173330).
Java core dump file is too large to be attached (more than 460MB). I don't think it'll be of any help anyways.
Hi Navid, What jvm did you use? I successfuly started AC with long path using RHEL AS 4 upd 3 and Sun jvm 1.5.0_06. But I had the same problem on Windows. In any case this defect is caused by jvm in FindClass method of JNI entry point and can't be fixed in TPTP code. Might it be resolved by adding some notes to the documentation?
I'm using the IBM JVM 1.5.0. That's strange, I didn't experience this problem at all on Windows. Which Windows distribution did you experience this in? Unfortunately this can't be resolved by just documenting it. The problem is that for the new IAC, the agent controller is shipped as an Eclipse plug-in and as a result, it will have a long path associated with it. Because of this defect, I'm unable to start the Agent Contorller that's shipped as an Eclipse plug-in. Can you try reproducing this with the IBM JVM and let me know if it still can't be resolved in the TPTP code?
(In reply to comment #3) > Can you try reproducing this with the IBM JVM and let me know if it still can't > be resolved in the TPTP code? Navid, I successfuly started AC and launched simple smoke tests with IBM jvm - IBM J9 VM (build 2.3, J2RE 1.5.0 IBM J9 2.3 Linux x86-32 j9vmxi3223-20061001 (JIT enabled). I used path you said.
I believe this is related to bug 141098. Having a set of long CLASSPATH entries along with a long PLUGIN_HOME value in AC's config file causes a segmentation fault when AC is started.
I have found out that there are too many plugins adding jars to the global CLASSPATH environment. This will cause the AC to crash because the length of the CLASSPATH exceed the system limit. I have updated the config files for the following plugins to localize the classpath to the "java.exe" application instead. - org.eclipse.tptp.monitoring.instrumentation - org.eclipse.tptp.trace.arm I have tested it myself and the AC starts up fine after this change. More testing is required to make sure this classpath localization will not break the plugins specified above.
Created attachment 65949 [details] Config generator patch for "org.eclipse.tptp.trace.arm"
Created attachment 65950 [details] Config generator patch for "org.eclipse.tptp.monitoring.instrumentation"
The attached patches have been checked in. Marking this as fixed awaiting for the next AC build for verification.
As of TPTP 4.6.0, TPTP is in maintenance mode and focusing on improving quality by resolving relevant enhancements/defects and increasing test coverage through test creation, automation, Build Verification Tests (BVTs), and expanded run-time execution. As part of the TPTP Bugzilla housecleaning process (see http://wiki.eclipse.org/Bugzilla_Housecleaning_Processes), this enhancement/defect is verified/closed by the Project Lead since this enhancement/defect has been resolved and unverified for more than 1 year and considered to be fixed. If this enhancement/defect is still unresolved and reproducible in the latest TPTP release (http://www.eclipse.org/tptp/home/downloads/), please re-open.
