Does D support types like cpp_int or cpp_rational in
boost::multiprecision? D is designed directly mapped to hardware
for speed reason. But in some cases, we do need types like
cpp_int or cpp_rational support.

Is there a plan to support those kind of types? If support, is it
a core language feature or library support?

SteveGuo:

> Is there a plan to support those kind of types? If support,
> is it a core language feature or library support?

That will not have a core language support, beside perhaps cent/ucent. At the moment there are no plans for library support, but eventually in some years fixed-size multi-precision integers will be good to have in Phobos, just like several other things, like nD packed arrays (http://pempek.net/articles/2013/08/03/packedarray-random-access-array-tightly-packed-unsigned-integers/ ) etc.

Bye,
bearophile

> That will not have a core language support, beside perhaps cent/ucent. At the moment there are no plans for library support, but eventually in some years fixed-size multi-precision integers will be good to have in Phobos, just like several other things, like nD packed arrays (http://pempek.net/articles/2013/08/03/packedarray-random-access-array-tightly-packed-unsigned-integers/ ) etc.
>
> Bye,
> bearophile

Thanks for replying:) I have to choose c++ on my project. Though I love D.

SteveGuo:

> I have to choose c++ on my project.

Are std.bigint not efficient enough?

Bye,
bearophile

> Are std.bigint not efficient enough?
>
> Bye,
> bearophile

std.bigint, I think it's an equevelent of cpp_int, but My project has to use cpp_rational

SteveGuo:

> std.bigint, I think it's an equevelent of cpp_int, but My project has to use cpp_rational

A rational was implemented in D by several persons, it's easy. Of the few problems is implement an efficient GCD among bigints (that sooner or later will be needed in std.bigint).

Bye,
bearophile

> A rational was implemented in D by several persons, it's easy. Of the few problems is implement an efficient GCD among bigints (that sooner or later will be needed in std.bigint).
>
> Bye,
> bearophile

That's great!

On Tue, Aug 06, 2013 at 07:43:29AM +0200, SteveGuo wrote:
> Does D support types like cpp_int or cpp_rational in boost::multiprecision? D is designed directly mapped to hardware for speed reason. But in some cases, we do need types like cpp_int or cpp_rational support.

There's std.bigint, but I don't know whether there's a rational library yet. But you can probably use BigInt to implement one.


T

-- 
Help a man when he is in trouble and he will remember you when he is in trouble again.