> that option also sets lua floating point numbers to 32bit, which > is unacceptable for my uses (I need IEEE 754 double precision arithmetic). Try defining LUA_C89_NUMBERS instead of LUA_32BITS. This will give you long integers and double floats.