Apache OpenOffice (AOO) Bugzilla – Issue 103877
reportbuilder: NULL pointer exceptions and crashes
Last modified: 2013-02-07 22:41:08 UTC
I get Null pointer exceptions and crashes with the Sun Report Builder. OS: openSUSE-11.1-x86_64 OOo: OOO310m11 (Build:9399) from OOo_3.1.0_LinuxX86-64_install_en-US.tar.gz from http://download.openoffice.org/other.html JDK1: java version "1.6.0_14" Java(TM) SE Runtime Environment (build 1.6.0_14-b08) Java HotSpot(TM) 64-Bit Server VM (build 14.0-b16, mixed mode) JDK2: java version "1.6.0_0" IcedTea6 1.4 (suse-24.8.12-x86_64) Runtime Environment (build 1.6.0_0-b14) OpenJDK 64-Bit Server VM (build 14.0-b08, mixed mode) Sun Report Builder: version 1.1.0 from http://extensions.services.openoffice.org/project/reportdesign Steps to reproduce: 1. start /opt/openoffice.org3/program/sbase 2. select "Open an existing database file"; choos ~/.openoffice.org/3/user/database/biblio.odb; press the button "Finish" 3. select "Reports" in the "Databases" window (left side) 4. select "Use Wizard to Create Report..." 5. press the button ">>" to use all available fields 6. press the "Finish" button Result with JDK1 (Sun build 1.6.0_14-b08): It shows the report but it prints some error messages about Null pointer exceptions on the standard output, see srb.sun.jdk.console.log Result with JDK2 (OpenJDK build 14.0-b08) It ends with segmentation fault, see srb.openjdk.console.log and hs_err_pid28650.log Expected result: It should work without any exceptions or crashes Note 1: I am not sure if the Null pointer exception is really related to the crash but it is suspicious. Note 2: We have similar bug also for the openSUSE build, see https://bugzilla.novell.com/show_bug.cgi?id=510795. I tried to debug it but without big success. It would be nice to get some hints from someone who is more familiar with the code.
Created attachment 63839 [details] Console log with JDK1 (Sun build 1.6.0_14-b08) - shows Null pointer exceptions
Created attachment 63840 [details] Console log with JDK2 (OpenJDK build 14.0-b08) - crash
Created attachment 63841 [details] Backtrace from the crashed JDK (hs_err_pid28650.log)
putting into DBA pool, assigning it directly to me is not the way to go (the more since I'd not be the right developer for this).
okay, finally found the time to look into this in more detail :) I somehow doubt the both problems are related. That's a gut feeling only, admittedly. Nonetheless, I am not really happy with having those two in one issue. Unfortunately, the backtrace from the crash is not really helpful due to the "error occurred during error reporting (printing native stack), id 0xb" part - not sure what important information are missing here. In any case, the fact that the crash happens with one JVM only, but not with the other, seems to indicate bug in the JVM, not in OOo. Finally, we're talking about Java code being executed, and neither a NullPointerException in this code (though certainly worth separate investigations) nor anything else justifies the JVM crashing. So, I'm tempted to see the crash as "not an OOo bug". Did you try to debug the JDK during the crash?
Thanks for feedback. I'll try to debug the problematic JVM. Well, it might take some time until I find time to do it. Heh, these things are really hard to debug.