[boost] Re: Rational Numbers - abs() again

From: Boris Fomitchev (fbp_at_[hidden])
Date: 1999-12-14 20:01:44

I see. Looks like full-scale solution is needed then :

template <class _Tp>
struct _STL_math_proxy {
  static inline _Tp _do_abs(const _Tp& __x)     { return __STL_MATH_STD::abs(__x); }
  static inline _Tp _do_acos(const _Tp& __x)    { return __STL_MATH_STD::acos(__x); }
  static inline _Tp _do_asin(const _Tp& __x)    { return __STL_MATH_STD::asin(__x); }
  static inline _Tp _do_atan(const _Tp& __x)    { return __STL_MATH_STD::atan(__x); }
  static inline _Tp _do_atan2(const _Tp& __x, const _Tp& __y)  { return __STL_MATH_STD::atan2(__x,
__y); }
  static inline _Tp _do_cos(const _Tp& __x)     { return __STL_MATH_STD::cos(__x); }
  static inline _Tp _do_cosh(const _Tp& __x)    { return __STL_MATH_STD::cosh(__x); }
  static inline _Tp _do_log(const _Tp& __x)     { return __STL_MATH_STD::log(__x); }
  static inline _Tp _do_log10(const _Tp& __x)   { return __STL_MATH_STD::log10(__x); }
  static inline _Tp _do_pow(const _Tp& __x, const _Tp& __y)    { return __STL_MATH_STD::pow(__x, __y);
}
  static inline _Tp _do_sin(const _Tp& __x)     { return __STL_MATH_STD::sin(__x); }
  static inline _Tp _do_sinh(const _Tp& __x)    { return __STL_MATH_STD::sinh(__x); }

  static inline _Tp _do_sqrt(const _Tp& __x)    { return __STL_MATH_STD::sqrt(__x); }
  static inline _Tp _do_tan(const _Tp& __x)     { return __STL_MATH_STD::tan(__x); }
  static inline _Tp _do_tanh(const _Tp& __x)    { return __STL_MATH_STD::tanh(__x); }
  static inline _Tp _do_exp(const _Tp& __x)     { return __STL_MATH_STD::exp(__x); }
  static inline _Tp _do_hypot(const _Tp& __x, const _Tp& __y)   { return __STL_MATH_STD::hypot(__x,
__y); }
};

where __STL_MATH_STD is empty for VC++.
Then you have to write specialization for complex<> (you cannot do partial with VC++, so the best you
can do
is to provide two full specializations for complex<float> and complex<double>), which would use
std::abs instead of ::abs.

From rational<Tp>::abs you call __STL_math_proxy<Tp>::abs()

Looks like a crap, doesn't it ?  ;).

BTW, if you intend to use rational on complex<>, you also have to change default in rational
constructor to IntType().

-Boris.

jsiek_at_[hidden] wrote:

> Does this work if the type is complex? (perhaps this
> doesn't apply here, but it does to many of my examples)
>
> Thanks,
>
> Jeremy
>
> Boris Fomitchev writes:
>  > Well,  abs:: already contains the conditional you need, just make it explicit (example below
>  > works):
>  >
>  >
>  > template <typename IntType>
>  > inline rational<IntType> abs(const rational<IntType>& r)
>  > {
>  > #ifndef BOOST_NO_STDC_NAMESPACE
>  >     // We want to use abs() unadorned below, so that if IntType is a
>  >     // user-defined type, the name lookup rules will work to find an
>  >     // appropriate abs() defined for IntType.
>  >     //
>  >     // We protect this with BOOST_NO_STDC_NAMESPACE, as some compilers
>  >     // don't put abs into std:: (notably MS Visual C++) and so we can't
>  >     // "use" it from there.
>  >     // fbp : changed to explicit qualifier
>  >     //    using std::abs;
>  >     // Note that with normalized fractions, the denominator is always positive.
>  >     return rational<IntType>(std::abs(r.num), r.den);
>  > #else
>  >     // Note that with normalized fractions, the denominator is always positive.
>  >     return rational<IntType>(::abs(r.num), r.den);
>  > #endif
>  > }
>  >
>  >
>  > If you need more of suct tricks, take a look at STLport (www.stlport.org).
>  > There, I often had to resolve this problem in same namespace, which even required
>  > proxy functions to be deployed.
>  >
>  >
>  > Best regards,
>  > -Boris.
>  >
>  > Paul Moore wrote:
>  >
>  > > I am still struggling with implementing abs() for rationals.
>  > >
>  > > To summarise, I have
>  > >
>  > >     template <typename IntType> class rational {
>  > >         ...
>  > >         friend rational abs (const rational &);
>  > >     }
>  > >
>  > >     template <typename IntType>
>  > >     inline rational<IntType> abs (const rational<IntType> &r)
>  > >     {
>  > >         // point (1)
>  > >         return rational<IntType>(abs(r.num), r.den);
>  > >     }
>  > >
>  > > Now, Koenig Lookup should find the friend abs when abs(r) is called with a
>  > > rational. But what about the *definition* of abs()? I use abs() here, as using
>  > > std::abs precludes IntType being a user-defined type. But to do this, I need to
>  > > add "using std::abs;" at point (1).
>  > >
>  > > Now we have an issue - the using declaration needs to be protected by
>  > > BOOST_NO_STDC_NAMESPACE from <boost/config.hpp>.
>  > >
>  > > But even with this, MSVC *still* has a problem - it fails to look for ::abs
>  > > [[ie, std::abs]] when compiling abs(rational). The error is
>  > >
>  > > ..\..\boost/rational.hpp(248) : error C2784: 'class boost::rational<IntType>
>  > > __cdecl boost::abs (const class boost::rational<IntType> &)' : could not deduce
>  > > template argument for 'const class boost::rational<IntType> &' from 'const int'
>  > >         rational_example.cpp(61) : see reference to function template
>  > > instantiation 'class boost::rational<int> __cdecl boost::abs(const class
>  > > boost::rational<int> &)' being compiled
>  > >
>  > > which seems to imply that MSVC isn't looking for the "normal" abs(int).
>  > >
>  > > Can anybody suggest a workaround for this??? I ought to say that MSVC
>  > > compatibility is an absolute requirement - I use MSVC, so I'm not going to write
>  > > code which doesn't work with it (even allowing for the fact that I'm not sure
>  > > exactly what the standard would mandate on this issue!).
>  > >
>  > > I'm thinking that I might be reduced to replacing abs(n) with (n < 0 ? -n : n).
>  > > This is OK, as we are requiring IntType to have "integer-like" semantics, but it
>  > > blows any hope of optimisation for a user-defuined IntType right out of the
>  > > water. (If someone can offer me a suitable implementation for
>  > > standard-conforming compilers, I'd be willing to put the hack in an ifdef
>  > > _MSC_VER).
>  > >
>  > > All assistance gratefully accepted :-) I attach the header and test program.
>  > >
>  > > Paul
>  > >
>  > > ------------------------------------------------------------------------
>  > > Check out Media Grok and 14 other FREE newsletters from
>  > > The Industry Standard. http://clickhere.egroups.com/click/1762
>  > >
>  > > -- Talk to your group with your own voice!
>  > > -- http://www.egroups.com/VoiceChatPage?listName=boost&m=1
>  > >
>  > >   ------------------------------------------------------------------------
>  > >                    Name: rational.hpp
>  > >    rational.hpp    Type: C++ Header File (application/x-unknown-content-type-hppfile)
>  > >                Encoding: quoted-printable
>  > >
>  > >                            Name: rational_example.cpp
>  > >    rational_example.cpp    Type: C++ Source File (application/x-unknown-content-type-cppfile)
>  > >                        Encoding: 7bit
>  >
>  >
>  > ------------------------------------------------------------------------
>  > GRAB THE GATOR! FREE SOFTWARE DOES ALL THE TYPING FOR YOU!
>  > Gator fills in forms and remembers passwords with NO TYPING at over
>  > 100,000 web sites! Get $100 in coupons for trying Gator!
>  > http://clickhere.egroups.com/click/2092
>  >
>  > -- Create a poll/survey for your group!
>  > -- http://www.egroups.com/vote?listname=boost&m=1
>  >
>  >
>  >
>
> ------------------------------------------------------------------------
> GET $100 IN COUPONS FOR TRYING GATOR!
> Grab the Gator! Free software does all the typing for you!
> Gator fills in forms and remembers passwords with NO TYPING at over
> 100,000 web sites! http://clickhere.egroups.com/click/2093
>
> -- Check out your group's private Chat room
> -- http://www.egroups.com/ChatPage?listName=boost&m=1