Re: [boost] Compile-time power of a run-time base

From: Bruno Lalande (bruno.lalande_at_[hidden])
Date: 2008-02-15 04:49:20

Hello,

> "These rules are chosen to be compatible with the behaviour of
> > ISO/IEC 9899:1999 Programming languages - C and with the Draft
> > Technical Report on C++ Library Extensions, 2005-06-24, section
> > 5.2.1, paragraph 5.
> >
> > http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2005/n1836.pdf
> > "
>
> Right the idea is to be compatible with what C99 and C++0x do when an
> integer argument is passed to a math function.
>

OK I have read this section of the draft and now I understand better the
purpose of promote_args, so I will use it in order to to be C99 and C++0x
compliant.

> The other issue is that if:
>
> pow<-3>(2)
>
> returned an integer, then presumably it can only return 0, which isn't
> exactly right :-)

Just to see if it was even possible, I've implemented what I had talked
about before. It takes this issue into account. Basically, here is the
output of my test program with the result of pow on several types and their
corresponding typeids:
pow<2>(2) =     (i)4
pow<2>(2.0) =   (d)4
pow<2>(2.0f) =  (f)4
pow<-2>(2) =    (d)0.25
pow<-2>(2.0) =  (d)0.25
pow<-2>(2.0f) = (d)0.25
And regarding the way in which I did it, I "guess" a use with
Boost.Unitswould give the expected result. Something like:
pow<3>((length)5) = (volume)125;

But the most important is to implement the promote_args version. I will
provide this soon. After that we can always propose, as Paul said, a special
version with a different name.

Regards
Bruno