| View Issue Details [ Jump to Notes ] | [ Issue History ] [ Print ] | ||||||||
| ID | Project | Category | View Status | Date Submitted | Last Update | ||||
|---|---|---|---|---|---|---|---|---|---|
| 0001890 | Spring engine | Linux | public | 2010-04-08 20:22 | 2011-01-01 20:20 | ||||
| Reporter | jandd | ||||||||
| Assigned To | abma | ||||||||
| Priority | normal | Severity | crash | Reproducibility | always | ||||
| Status | resolved | Resolution | unable to reproduce | ||||||
| Product Version | 0.81.2 | ||||||||
| Target Version | Fixed in Version | ||||||||
| Summary | 0001890: Java AI JNA crash on Debian | ||||||||
| Description | After trying to get the Java AIs running in the Debian packaged version of spring together with hoijui we got stuck last night. I fixed a bug in Debian's version of libjna-java to properly load the platform specific .so file and got to another bug that produces the following stacktrace (with HughAI): <SkirmishAI: HughAI 0.1 (team 1)>: Creating new cachefile for mod Balanced Annihilation V7.12 0 ms INFO: Generating indexes mod Balanced Annihilation V7.12 Segmentation fault (SIGSEGV) in spring 0.81.2.1 Stacktrace: spring [0x8f7407] /usr/lib/jvm/java-6-sun/jre/lib/amd64/server/libjvm.so [0x7f4ead3b8acd] /usr/lib/jvm/java-6-sun/jre/lib/amd64/server/libjvm.so [0x7f4ead3b608b] /usr/lib/jvm/java-6-sun/jre/lib/amd64/server/libjvm.so(JVM_handle_linux_signal+0x170) [0x7f4ead3b9660] /usr/lib/jvm/java-6-sun/jre/lib/amd64/server/libjvm.so [0x7f4ead3b5f9e] /lib/libpthread.so.0 [0x7f4ebb446990] spring(skirmishAiCallback_UnitDef_MoveData_getMaxSlope+0x10) [0x993f10] /usr/lib/libffi.so.5(ffi_call_unix64+0x4c) [0x7f4e90790f18] /usr/lib/libffi.so.5(ffi_call+0x204) [0x7f4e90790cb4] /usr/lib/jni/libjnidispatch.so [0x7f4e9099a47e] /usr/lib/jni/libjnidispatch.so(Java_com_sun_jna_Function_invokeFloat+0x18) [0x7f4e9099a8b8] [0x7f4e888d6f50] # NOTE: no path -> not translating The bug is reproducible with both Sun JRE and OpenJDK's JRE. The Java AI Interface is compiled using OpenJDK (which is the default JDK on Debian AMD64). I experimented with the JNA jar file which is bundled with the spring tarball but got the same result. | ||||||||
| Additional Information | The current spring packaging code is available from http://git.debian.org/?p=pkg-games/spring.git | ||||||||
| Tags | No tags attached. | ||||||||
| Checked infolog.txt for Errors | |||||||||
| Attached Files |
| ||||||||
 Relationships |
|
 Relationships |
| Notes |
|
|
hoijui (reporter) 2010-04-11 10:51 |
HughAI is long unmaintained, and even before that, it was hardly ever tested, so it might well be a problem with HughAI. One particular limitation of the JavaAI Interface is, that only a single thread is allowed to call back to the engine, and it has to be the same thread that receives the events from the engine. Could be that HughAI violates this.. in theory (though i have seen it working already). The bad thing is.. as said HughAI is not really reliable.. nobody still active in the community knows much about it, but it is the only Java AI with sources publicly available that really does something. |
|
abma (administrator) 2011-01-01 20:19 |
please retry with curren version, as i couldn't reproduce this. |
| Notes |
| Issue History |
|||
| Date Modified | Username | Field | Change |
|---|---|---|---|
| 2010-04-08 20:22 | jandd | New Issue | |
| 2010-04-11 10:51 | hoijui | Note Added: 0004819 | |
| 2011-01-01 20:19 | abma | Severity | minor => crash |
| 2011-01-01 20:19 | abma | Note Added: 0006181 | |
| 2011-01-01 20:19 | abma | Status | new => resolved |
| 2011-01-01 20:19 | abma | Resolution | open => fixed |
| 2011-01-01 20:19 | abma | Assigned To | => abma |
| 2011-01-01 20:20 | abma | Resolution | fixed => unable to reproduce |
| Issue History |