Last updated
Was this helpful?
Last updated
Was this helpful?
The simple solution to doing calculations where floating point numbers exhibit pathological behavior is: don't do floating point calculations. :-) Raku is just as susceptible to floating point error as any other C based language, however, it offers built-in rational Types; where numbers are represented as a ratio of two integers. For normal precision it uses Rats - accurate to 1/2^64, and for arbitrary precision, FatRats, whose denominators can grow as large as available memory. Rats don't require any special setup to use. Any decimal number within its limits of precision is automatically stored as a Rat. FatRats require explicit coercion and are "sticky". Any FatRat operand in a calculation will cause all further results to be stored as FatRats.