Yes, I should most definitely fix this. Would you mind opening an issue?
On Thu, May 1, 2014 at 11:34 AM, Alain Bouchard <[email protected]> wrote: > I undestand that for BigInt there is no hard limit like for Int64 for > example but the lack of a value seems to cause a problem with rationalize() > at line 69 where it is not possible to get a Rational with BigInt num and > den. It seems to me that line 69 of rational.jl shall special case the > BigInt or that BigInt shall return something valid for typemin/typemax. > > at line 69: > > if x < typemin(T); return -one(T)//zero(T); end > > >
