The following generates integers natively.
The range is all possible integers between -2^32 and +2^31-1
infinity cannot be generated with this setup, as infinity is not part of the integer system.
To generate integers outside of the range indicated above, you need to switch to double(), but if you do that, you only have 53 bits of precision available, and all values outside the range +/- 2^53 are integers that are multiplied by a power of 2 -- you could not, for example, generate 2^53 + 17 . By 1E50, adjacent representable values are 2^114 apart. Is that what you want to deal with?