Hi, does exist a way in Codea to handle large ‘integer’ values with right precision?
for example, the following code will print 0 instead of 6
i1 = 2147483688
i2 = 2147483682
print(i1-i2)
and that’s because both are considered as 2.14748e9
Codea’s Lua uses floats (32 bit) for numbers, that gives you only 23 bits and a sign bit for the integer range, i.e. 8388607 is the biggest possible exact integer.
There’s a put Lua BigNum library here:
http://oss.digirati.com.br/luabignum/index.htm
I haven’t looked into it, but maybe it’s what you need.
Yes, thanks @Codeslinger, I’ ve also read another post that refers to a similar problem:
My real need is to correctly parse a file written by another app where some integers are used in a mixed way: 3 most significant bit as flags, the remaining as int value. I’ve a solution for this specific problem, but it’s not something I really like 
Anyway is a real pity having numbers as single float precision (even if probably having them as double float precision would slow down codea performance, or am I wrong?)