Loss of integer precision

[samstronghammer]

This library works cleanly and well for decimal-based arithmetic, but I was having some issues with getting precise integral answers when the inputs were large, due to the conversion from integer -> decimal that happens with the division operation. If this sounds like helpful behavior, I'd be interested in contributing this functionality, but wanted to get @SebastianCallh's input first. Would it be best to make a new set of functions that attempt to do lossless division (by checking if the remainder would be zero and doing integer division) or modify the existing functions to support this? Modifying the existing functions would likely cause some amount of a performance hit.