Closed mojjj closed 8 years ago
That error log doesn't suggest anything went wrong with kgb. If kgb has a memory allocation error then java throws an OOM exception, but that was all C++ and looked like spring itself had some sort of error. It could possibly be related to circuit but I don't really know.
well now, when i only use one zkgb as oponent, the result is the same. does not happen with CAI or circuit (nor any multiplayer)...
http://i.imgur.com/NkUVL7g.png
tell me what to do to get more intel to fix this issue if you have an idea...
Do you have the same problem with all JavaAIs? Try for example http://zero-k.info/Forum/Thread/22533
Have you tried monitoring memory usage with something like htop?
The only thing notable I could see is that you're running linux. It seems to be using the correct jvm though, so at least that shouldn't be a problem. As DF said, either test CSI or the java nullOOAI and see if that gives the same result. There's not enough info to tell if it's a spring issue or if it's your system or what.
i have 16g memory i doubt that is an issue there... infologs, using 0.8.2 of zkgb CSI 1337 crashes the same way => : http://pastebin.com/A8PzWRaZ NullOOJavaAI", Version:"0.1 too => http://pastebin.com/aLtHsyEt
This looks like an engine issue, then. I can replicate this, so maybe i'll find some clues.
@mojjj i have succeeded in launching zkgbai (also nullai and csi) by passing Spring a proper JAVA_HOME environment variable.
It seems that Spring fails at detecting JRE on newer Ubuntu.
This is an engine bug.
nice. is this limited to ubuntu or other lunix-systems too? why not windows? do you report this to engine tracker?
could you share how to pass the env var?
JAVA_HOME=/usr/share/whatever/ ~/.spring/engine/103.0.1-abracadabra/spring ~/somewhere/startscript.spring
You can also export
, of course.
The method i used to find the current java home was either update-alternatives --config java
or which java
, can't be sure. Also i removed last two path elements (/bin/java
) to get to proper JAVA_HOME and not just binary location.
I have not reported it on mantis yet.
i reported it on mantis https://springrts.com/mantis/view.php?id=5354
might be unrelated, but 3 humans vs 2 circuit + 2 zkgb => window freeze when com plopp should be, infolog attached.
http://pastebin.com/afNXHbgn