Re: [boost] libraries affected by common factor gcd and lcm namespace change moving from math to integer library

Subject: Re: [boost] libraries affected by common factor gcd and lcm namespace change moving from math to integer library
From: Paul A. Bristow (pbristow_at_[hidden])
Date: 2015-01-30 04:49:15

> -----Original Message-----
> From: Boost [mailto:boost-bounces_at_[hidden]] On Behalf Of Vicente J. Botet Escriba
> Sent: 29 January 2015 17:58
> To: boost_at_[hidden]
> Subject: Re: [boost] libraries affected by common factor gcd and lcm namespace change moving from
> math to integer library
> 
> Le 29/01/15 15:34, Paul A. Bristow a écrit :
> > A quick grep (for gcd_ and lcm_ in *.*pp) suggests that the following
> > libraries are affected by gcd and lcm namespace change moving from math to integer library.
> >
> > At present, math_fwd.hpp provides some typedefs and declarations
> > needed for these,  and a copy of the common factor codes, so nothing
> > may fail immediately, but these duplicate will be removed after at least one release.
> >
> >
> > Chrono
> >
> > Interprocess
> >
> > Math
> >
> > Multiprecision
> >
> > Ratio
> >
> > Rational
> >
> > Units
> >
> > All libraries should use the declarations in integer/integer_fwd.hpp
> > and /integer/common_factor_ct.hpp and /integer/common_factor_rt.hpp.
> >
> > (And there may be documentation changes too?)
> >
> > Something needs to be done ;-)
> >
> > (If you want *me* to resolve anything, then I need write access to your library).
> >
> >
> Paul you should refine your grep. Lastly there were some libraries that changed already from
math:: to
> integer.
> 
> IMO, the single library I see is using math::gcd/lcm is multiprecision, and I'm not sure removing
this
> dependency is really needed.
> Please provide the result of your grep.
> 
> I don't think we should remove math::gcd/lcm, just forward to the integer::gcd/lcm.

I'm not suggesting removal (yet) and we will provide a forwarding,

but wasn't the Big Idea that we reduced interdependence between libraries?

I was just suggesting authors of libraries that use gcd or lcm might like to make a very small
change to use the Boost.Integer library instead of Boost.Math (if they haven't already)?

Or am misunderstanding?

Paul

---
Paul A. Bristow
Prizet Farmhouse
Kendal UK LA8 8AB
[phone]