Re: [boost] [Review Request] Multiprecision Arithmetic Library

Subject: Re: [boost] [Review Request] Multiprecision Arithmetic Library
From: Vicente J. Botet Escriba (vicente.botet_at_[hidden])
Date: 2012-04-04 01:30:11

Le 04/04/12 00:49, Andrii Sydorchuk a écrit :
>> What is the result type of operator+ on this upper bound range? Currently
>> the compiler fails.
>
> Following code compiles for me:
>
> template<int N>
> struct big_int {
>    big_int() { val = N; }
>    template<int M>
>    big_int<(N>M?N:M)>  operator+(const big_int<M>  &that) const {
>      return big_int<(N>M?N:M)>();
>    }
>    int val;
> };
>
> int main() {
>    big_int<1>  a;
>    big_int<2>  b;
>    std::cout<<  (a + b).val<<  std::endl;
>    std::cout<<  (b + a).val<<  std::endl;
> }
>
My comment was respect to my library ;-)
>> I hope this answers to your argument against arithmetic operators growing
>> automatically the range of the result.
>
> And do you think this is worth the effort?
Yes. At least for my needs.
> Also I wouldn't like the idea
> of  fixed_point<range,0>  structures. It is going to be too complicated from
> the user perspective. All I need from fixed integer library is efficient
> fixed integer type (behaving like C++ integral types) and no complex
> structures to compute addition, multiplication, etc.
>
I've not yet defined an specific class for fixed point integers. Note 
that the resolution parameter defaults to 0, so that fixed_point<range> 
defines an integer. I'll include it once the generic class is stable enough.

Best,
Vicente