3 ms·

I remember this bug. Our workaround was to parse the output of toString() for 0x0000000, and simply drop the frame when it happens.

• threeseed 12y ago
  I'm curious. Was it something that existing across many JVM updates ? Given the stability of the language if I have ever seen an issue (rare) there's always been a choice of 10-20 JVM versions to pick from.
  • gizmo686 12y ago
    The competition gives us six weeks to build the robot, before we have one or two three day competitions. Apart from possibly attending non-official off-season competitions, that is the only time we would use the code before the next years competition, which is completely different, and for which we are required to right entirely new code[1]. Once we got a hack working, we ran with it and rarely tried to clean up. In this case, we didn't bother trying a different JVM. [1] Unless the code is open source (as ours was), in which case we were explicitly allowed to use it.
  • maaaats 12y ago
    It's not really a JVM bug, but an artifact of dealing with native code through Java without thinking about how the memory of the native.