Interesting. If the changes don't hurt the original use case of the type, they may be reasonable to make.
On Fri, Feb 12, 2016 at 4:59 PM, Fengyang Wang <fengyangwa...@gmail.com> wrote: > You are right; it did not work. I made some modifications to fix it, but > they required a significant rethinking of many of the methods, so it would > be a different type entirely. Nemo supports rational functions in any case, > and I think that's a better idea. > > On Sunday, February 7, 2016 at 11:21:17 AM UTC-5, Stefan Karpinski wrote: >> >> There are various assumptions baked into the rational code that may or >> may not be satisfied by non-integer numeric types. I would suggest taking >> the code from Base and trying it out without that restriction and seeing >> how it goes. >> >> On Saturday, February 6, 2016, Fengyang Wang <fengya...@gmail.com> wrote: >> >>> I was looking for a Julia package to handle rational functions, when I >>> noticed that the `Polynomials` package implements `gcd`, `div`, and `rem`. >>> So it would be possible to simply use `Rational{Poly}`... or so I thought. >>> Unfortunately, the type `Rational` prevents this use, since it requires its >>> type parameter to derive from `Integer`. >>> >>> I think it would be more in line with Julia's goal of polymorphism if >>> `Rational` "just worked" with any Euclidean domain. Is there some >>> justification for the current behaviour, or should I file this as a issue >>> (or make a pull request)? >>> >>