I suspect everyone who plays around with numerical problems on Perl 6 has been as frustrated as I am not to have arbitrary precision integers in Rakudo. As a sort of first step toward getting them, I’ve thrown together the beginnings of a Math::BigInt module using the BIgDigit multiple-precision arithmetic library and Zavolaj.

(Why BigDigit instead of GMP? Two simple reasons. First, building GMP failed on my main development machine (OS X 10.5). GMP’s suggestion for fixing the problem involved getting a new main development environment for the machine, which seemed like a major pain in the neck, and didn’t impress me with GMP’s portability. Second, GMP’s main multiple-precision type is a C struct you store on the stack. That means using it with Zavolaj would have required creating wrappers for every GMP function I wanted to use. Yuck.)

Without further ado, here’s a few examples:

> my @fib := Math::BigInt.new(1), Math::BigInt.new(1), *+* ... *; > say @fib[200]; 453973694165307953197296969697410619233826 > my @powers-of-two := Math::BigInt.new(1), * * 2 ... *; > say @powers-of-two[80]; 1208925819614629174706176

From which you can quickly conclude two things: Math::BigInts easily worked with the sequence operator. And creating a Math::BigInt is ugly and awkward. Once again, blogging has forced me to confront the limitations of my code.

So… if I had an operator to create a BigInt, the code would look a lot nicer. What’s more, you could avoid the term Math::BigInt actually appearing in your code, making it very easy to port to Perl 6s which actually support arbitrary precision Ints. Clearly then, I need to figure out a symbol to use…