The configuration is to make code be compatible with the target OS/CPU, this includes big-endian/little-endian differences, pointer size, etc. If it is not compatible with the target OS/CPU it may not compile properly. Actually the bytecode depends mostly on the CPU architecture rather than on the OS, so as long as the CPU architecture is the same in regards to pointer size and endianess on both machines, the scripts ought to be compatible.

In a future version I'll make sure the saved precompiled bytecode is platform independent, but it is currently rather low priority. If you feel up to doing this adaption yourself, the necessary changes should be solely in the asCRestore class.

I've made a few improvements to the saved bytecode in the WIP. The problem you're reporting may have actually been fixed already. I changed especially how the global variables are found when loading the bytecode.