Math::BigInt is a really wonderful module.

- Yes, I reinvent wheels.
- Spam: Visit eurotraQ.

BTW, you should relize that though the floats on your system can represent up to 10**300 (say), they still only have a strongly limited number of significant digits -- not all of the digits actualy get traced. So 10**5+1 == 10**5.

Also, consider if you want to use Math::BigInt, or rearange your arithmetic to produce smaller numbers. If you don't need all the accuracy, scaling things will make your program much faster. If you do need the accuracy, BigInt and suchlike can be a lifesaver.

Have you looked at Math::BigInt and Math::BigFloat? I don't know if there's a limit to the size of number they can store, but a quick scan of the docs didn't show one.

Actually, the limit is a lot smaller: as perl uses floating points to store its numbers, what you end up with is a big inaccurate number. the exact limits and accuracy are dependand.

If you want to know how perl stores numbers type:

perl -V
[download]

Find the line saying nvtype=' ... ' which is your system representation for perl numbers (usually double I guess). Also uselongdouble and uselongfloat if set to define might make your numbers more accurate.

You will have to decide for yourself if you care about the inaccuracies, if you do, see Math::Bigint and Math::Bigfloat as mentioned in the posts above.

Note: If you have a number stored in a string, it will only get converted when use them as numbers. So things like $big_number = '10231273879128749879857'; stay exact if you do not do arithmetics on them (and don't do things like $big_number == $other_number, use $big_number eq $other_number if you really need to).

-- 
Joost       downtime n. The period during which a system
            is error-free and immune from user input.
[download]