|
|
Boost-Commit : |
Subject: [Boost-commit] svn:boost r78104 - sandbox/fixed_point/boost/fixed_point
From: vicente.botet_at_[hidden]
Date: 2012-04-20 17:51:35
Author: viboes
Date: 2012-04-20 17:51:34 EDT (Fri, 20 Apr 2012)
New Revision: 78104
URL: http://svn.boost.org/trac/boost/changeset/78104
Log:
FixedPoint: document the interface
Text files modified:
sandbox/fixed_point/boost/fixed_point/number.hpp | 455 ++++++++++++++++++++++++++++-----------
1 files changed, 323 insertions(+), 132 deletions(-)
Modified: sandbox/fixed_point/boost/fixed_point/number.hpp
==============================================================================
--- sandbox/fixed_point/boost/fixed_point/number.hpp (original)
+++ sandbox/fixed_point/boost/fixed_point/number.hpp 2012-04-20 17:51:34 EDT (Fri, 20 Apr 2012)
@@ -147,15 +147,22 @@
*/
namespace round
{
+ /**
+ * Use the fastest rounding.
+ */
struct fastest
{
BOOST_STATIC_CONSTEXPR
std::float_round_style round_style = std::round_indeterminate;
};
+
+ /**
+ * Rounds toward negative infinity.
+ */
struct negative
{
BOOST_STATIC_CONSTEXPR
- std::float_round_style round_style = std::round_toward_infinity;
+ std::float_round_style round_style = std::round_toward_neg_infinity;
template <typename From, typename To>
static typename To::underlying_type round(From const& rhs)
{
@@ -198,6 +205,9 @@
}
}
};
+ /**
+ * Rounds toward zero.
+ */
struct truncated
{
BOOST_STATIC_CONSTEXPR
@@ -227,10 +237,13 @@
return ci;
}
};
+ /**
+ * Rounds toward positive infinity.
+ */
struct positive
{
BOOST_STATIC_CONSTEXPR
- std::float_round_style round_style = std::round_toward_neg_infinity;
+ std::float_round_style round_style = std::round_toward_infinity;
template <typename From, typename To>
static typename To::underlying_type round(From const& rhs)
{
@@ -277,21 +290,33 @@
}
}
};
+ /**
+ * Rounds to nearest half up.
+ */
struct nearest_half_up
{
BOOST_STATIC_CONSTEXPR
std::float_round_style round_style = std::round_to_nearest;
};
+ /**
+ * Rounds to nearest half down.
+ */
struct nearest_half_down
{
BOOST_STATIC_CONSTEXPR
std::float_round_style round_style = std::round_to_nearest;
};
+ /**
+ * Rounds to nearest even.
+ */
struct nearest_even
{
BOOST_STATIC_CONSTEXPR
std::float_round_style round_style = std::round_to_nearest;
};
+ /**
+ * Rounds to nearest odd.
+ */
struct nearest_odd
{
BOOST_STATIC_CONSTEXPR
@@ -304,6 +329,11 @@
*/
namespace overflow
{
+ /**
+ * Overflow is impossible and ensured by the user.
+ *
+ * An assertion is raised on debug mode.
+ */
struct impossible
{
BOOST_STATIC_CONSTEXPR
@@ -327,6 +357,11 @@
return value;
}
};
+ /**
+ * Overflow is undefined.
+ *
+ * Usually uses the fastest overflow approach.
+ */
struct undefined
{
BOOST_STATIC_CONSTEXPR
@@ -387,6 +422,9 @@
}
};
}
+ /**
+ * Overflow results are wrapped.
+ */
struct modulus
{
BOOST_STATIC_CONSTEXPR
@@ -404,6 +442,9 @@
return detail::modulus_on_negative_overflow<T,U>::value(val);
}
};
+ /**
+ * On overflow the results is the nearest.
+ */
struct saturate
{
BOOST_STATIC_CONSTEXPR
@@ -422,6 +463,9 @@
}
};
+ /**
+ * On overflow an exception is thrown.
+ */
struct exception
{
BOOST_STATIC_CONSTEXPR
@@ -446,6 +490,9 @@
namespace storage
{
+ /**
+ * The storage is undefined.
+ */
struct undefined
{
/**
@@ -468,6 +515,9 @@
typedef typename ::boost::uint_t<Range - Resolution>::least type;
};
};
+ /**
+ * The storage is chosen to be least wide possible.
+ */
struct space
{
template <int Range, int Resolution>
@@ -481,6 +531,9 @@
typedef typename ::boost::uint_t<Range - Resolution>::least type;
};
};
+ /**
+ * The storage is chosen to be fastest possible.
+ */
struct speed
{
template <int Range, int Resolution>
@@ -648,7 +701,8 @@
};
/**
- * helper function to make easier the use of index_tag.
+ * Helper function to make easier the use of @c index_tag.
+ * @returns @c index_tag<T>(v).
*/
template <typename T>
BOOST_CONSTEXPR index_tag<T> index(T v)
@@ -670,17 +724,29 @@
};
/**
- * helper function to make easier the use of convert_tag.
+ * Helper function to make easier the use of @c convert_tag.
+ * @returns @c convert_tag<T>(v).
*/
template <typename T>
BOOST_CONSTEXPR convert_tag<T> convert(T v)
{ return convert_tag<T>(v);}
/**
- * explicit conversion between fixed_point numbers.
+ * Explicit conversion between fixed_point numbers.
+ *
+ * @TParams
+ * @Param{From,the source type of the conversion}
+ * @Param{To,the target type of the conversion}
+ *
+ * @Params
+ * @Param{from,the number from which the conversion is done}
+ *
+ * @Requires @c From is a fixed_point number or a builtin type and @c To is a fixed_point number.
+ * @Returns the conversion with possible reduced range and loss of resolution.
+ *
*/
template <class From, class To>
- To number_cast(From const&);
+ To number_cast(From const& from);
namespace detail
{
@@ -1461,6 +1527,11 @@
} // namespace detail
+ /**
+ * @c common_type alike allowing to have a different type as result.
+ *
+ *
+ */
template <typename T, typename U >
struct default_type
{
@@ -1650,14 +1721,21 @@
struct is_convertible : is_more_precisse<T,S>
{};
- template <typename S, typename T>
- struct is_explicitly_convertible :
+ template <typename S, typename T, bool IsArithmetic=is_arithmetic<S>::value >
+ struct is_explicitly_convertible;
+ template <typename S, typename T >
+ struct is_explicitly_convertible<S,T,false> :
mpl::and_<
mpl::not_<allows_implicit_conversion_from_fp<T> >,
mpl::not_<is_more_precisse<T,S> >
>
{};
+ template <typename S, typename T>
+ struct is_explicitly_convertible<S,T,true> :
+ allows_explicit_conversion_from_builtin<T>
+ {};
+
template <typename S, typename T, bool IsArithmetic=is_arithmetic<S>::value >
struct is_implicitly_convertible;
template <typename S, typename T >
@@ -1678,11 +1756,11 @@
* @brief Signed fixed point number.
*
* @TParams
- * @Param{Range,Range specified by an integer. The range of a signed number x is 2^Range < x < 2^Range. Note that the range interval is open for signed numbers.}
- * @Param{Resolution,resolution specified by an integer. The resolution of a fractional number x is 2^Resolution.}
- * @Param{Rounding,The rounding policy.}
- * @Param{Overflow,The overflow policy.}
- * @Param{Family,The family traits.}
+ * @Param{Range,Range specified by an integer. The range of a signed number x is 2^Range < x < 2^Range. Note that the range interval is open for signed numbers}
+ * @Param{Resolution,resolution specified by an integer. The resolution of a fractional number x is 2^Resolution}
+ * @Param{Rounding,The rounding policy}
+ * @Param{Overflow,The overflow policy}
+ * @Param{Family,The family traits}
*
* @Example For example, real_t<8,-4> has values n such that -256 < n < 256 in increments of 2^(-4) = 1/16.
*/
@@ -1746,33 +1824,67 @@
*/
BOOST_CONSTEXPR real_t()
{} // = default;
+
/**
* Copy constructor.
*/
BOOST_CONSTEXPR real_t(real_t const& rhs) : value_(rhs.value_)
{} // = default;
- //! Implicit constructor from a real_t with no larger range and no better resolution
+ /**
+ * Implicit constructor from a real_t no larger range and no better resolution.
+ *
+ * @Params
+ * @Param{rhs,a real_t with no larger range and no better resolution}
+ *
+ * @Effects Adapt the resolution of @c rhs to this resolution.
+ * @Throws Nothing.
+ * @Notes This overload participates in overload resolution only if the source @c is_convertible to the target.
+ */
template <int R, int P, typename RP, typename OP, typename F>
real_t(real_t<R,P,RP,OP,F> const& rhs
+#if !defined(BOOST_FIXED_POINT_DOXYGEN_INVOKED)
, typename boost::enable_if<is_convertible<real_t<R,P,RP,OP,F>, real_t > >::type* = 0
+#endif
)
: value_(fixed_point::detail::fxp_number_cast<real_t<R,P,RP,OP,F>, real_t, true, true>()(rhs).count())
{
}
- //! Implicit constructor from a ureal_t with no larger range and no better resolution
+ /**
+ * Implicit constructor from a ureal_t with no larger range and no better resolution.
+ *
+ * @Params
+ * @Param{rhs,a ureal_t with no larger range and no better resolution}
+ *
+ * @Effects Adapt the resolution of @c rhs to this resolution.
+ * @Throws Nothing.
+ * @Notes This overload participates in overload resolution only if the source @c is_convertible to the target.
+ */
template <int R, int P, typename RP, typename OP, typename F>
real_t(ureal_t<R,P,RP,OP,F> const& rhs
+#if !defined(BOOST_FIXED_POINT_DOXYGEN_INVOKED)
, typename boost::enable_if<is_convertible<ureal_t<R,P,RP,OP,F>, real_t > >::type* = 0
+#endif
)
: value_(fixed_point::detail::fxp_number_cast<ureal_t<R,P,RP,OP,F>, real_t, true, true>()(rhs).count())
{
}
- //! explicit constructor from a real_t with larger range or better resolution
+ /**
+ * Explicit constructor from a real_t with larger range or better resolution.
+ *
+ * @Params
+ * @Param{rhs,a real_t with larger range and better resolution}
+ *
+ * @Effects Rounds and check overflow if needed.
+ * @Throws Whatever the target overflow policy can throw.
+ * @Notes This overload participates in overload resolution only if the source @c is_explicitly_convertible to the target.
+ */
template <int R, int P, typename RP, typename OP, typename F>
explicit real_t(real_t<R,P,RP,OP,F> const& rhs
+#if !defined(BOOST_FIXED_POINT_DOXYGEN_INVOKED)
, typename boost::enable_if <is_explicitly_convertible<real_t<R,P,RP,OP,F>,real_t> >::type* = 0
+#endif
)
: value_(fixed_point::detail::fxp_number_cast<real_t<R,P,RP,OP,F>, real_t,
mpl::less_equal < mpl::int_<R>, mpl::int_<Range> >::value,
@@ -1780,10 +1892,22 @@
>()(rhs).count())
{
}
- //! constructor from a real_t with larger range or better resolution
+
+ /**
+ * Constructor from a real_t with larger range or better resolution.
+ *
+ * @Params
+ * @Param{rhs,a real_t with larger range and better resolution}
+ *
+ * @Effects Rounds and check overflow if needed.
+ * @Throws Whatever the target overflow policy can throw.
+ * @Notes This overload participates in overload resolution only if the source @c is_implicitly_convertible to the target.
+ */
template <int R, int P, typename RP, typename OP, typename F>
real_t(real_t<R,P,RP,OP,F> const& rhs
+#if !defined(BOOST_FIXED_POINT_DOXYGEN_INVOKED)
, typename boost::enable_if <is_implicitly_convertible<real_t<R,P,RP,OP,F>,real_t> >::type* = 0
+#endif
)
: value_(fixed_point::detail::fxp_number_cast<real_t<R,P,RP,OP,F>, real_t,
mpl::less_equal < mpl::int_<R>, mpl::int_<Range> >::value,
@@ -1791,10 +1915,22 @@
>()(rhs).count())
{
}
- //! explicit constructor from a real_t with larger range or better resolution
+
+ /**
+ * Explicit constructor from a ureal_t with larger range or better resolution.
+ *
+ * @Params
+ * @Param{rhs,a ureal_t with larger range and better resolution}
+ *
+ * @Throws Whatever the target overflow policy can throw.
+ * @Effects Rounds and check overflow if needed.
+ * @Notes This overload participates in overload resolution only if the source @c is_explicitly_convertible to the target.
+ */
template <int R, int P, typename RP, typename OP, typename F>
explicit real_t(ureal_t<R,P,RP,OP,F> const& rhs
+#if !defined(BOOST_FIXED_POINT_DOXYGEN_INVOKED)
, typename boost::enable_if <is_explicitly_convertible<ureal_t<R,P,RP,OP,F>,real_t> >::type* = 0
+#endif
)
: value_(fixed_point::detail::fxp_number_cast<ureal_t<R,P,RP,OP,F>, real_t,
mpl::less_equal < mpl::int_<R>, mpl::int_<Range> >::value,
@@ -1803,10 +1939,21 @@
{
}
- //! constructor from a real_t with larger range or better resolution
+ /**
+ * Constructor from a ureal_t with larger range or better resolution.
+ *
+ * @Params
+ * @Param{rhs,a ureal_t with larger range and better resolution}
+ *
+ * @Effects Rounds and check overflow if needed.
+ * @Throws Whatever the target overflow policy can throw.
+ * @Notes This overload participates in overload resolution only if the source @c is_implicitly_convertible to the target.
+ */
template <int R, int P, typename RP, typename OP, typename F>
real_t(ureal_t<R,P,RP,OP,F> const& rhs
+#if !defined(BOOST_FIXED_POINT_DOXYGEN_INVOKED)
, typename boost::enable_if <is_implicitly_convertible<ureal_t<R,P,RP,OP,F>,real_t> >::type* = 0
+#endif
)
: value_(fixed_point::detail::fxp_number_cast<ureal_t<R,P,RP,OP,F>, real_t,
mpl::less_equal < mpl::int_<R>, mpl::int_<Range> >::value,
@@ -1815,7 +1962,17 @@
{
}
- //! implicit constructor from a real_t with larger range or better resolution when wrapped by a convert_tag.
+ /**
+ * Implicit constructor from a real_t wrapped by a convert_tag.
+ *
+ * @Params
+ * @Param{rhs,a real_t wrapped by @c convert_tag }
+ *
+ * @Effects Rounds and check overflow if needed.
+ * @Throws Whatever the target overflow policy can throw.
+ * @Notes This overload participates in overload resolution only if the source @c is_implicitly_convertible to the target.
+ */
+
template <int R, int P, typename RP, typename OP, typename F>
real_t(convert_tag<real_t<R,P,RP,OP,F> > rhs)
: value_(fixed_point::detail::fxp_number_cast<real_t<R,P,RP,OP,F>, real_t,
@@ -1824,7 +1981,16 @@
>()(rhs.get()).count())
{
}
- //! implicit constructor from a ureal_t with larger range or better resolution when wrapped by a convert_tag.
+ /**
+ * Implicit constructor from a ureal_t wrapped by a convert_tag.
+ *
+ * @Params
+ * @Param{rhs,a ureal_t wrapped by @c convert_tag }
+ *
+ * @Effects Rounds and check overflow if needed.
+ * @Throws Whatever the target overflow policy can throw.
+ * @Notes This overload participates in overload resolution only if the source @c is_implicitly_convertible to the target.
+ */
template <int R, int P, typename RP, typename OP, typename F>
real_t(convert_tag<ureal_t<R,P,RP,OP,F> > rhs)
: value_(fixed_point::detail::fxp_number_cast<ureal_t<R,P,RP,OP,F>, real_t,
@@ -1842,7 +2008,7 @@
* Explicit construction from an index.
*
* @Params
- * @Param{i,the index.}
+ * @Param{i,the index}
*
* @Requires <c>min_index<=i<=max_index</c>.
*/
@@ -1896,7 +2062,7 @@
return count() >> resolution_exp;
}
- //! conversion factor.
+ //! @Returns the conversion factor.
template <typename FP>
static FP factor()
{
@@ -1904,6 +2070,10 @@
else return FP(1)/(detail::shift_left<-Resolution>(1));
}
+
+ /**
+ * Reconstructs a floating point type from the underlying type.
+ */
template <typename FP>
static FP reconstruct(underlying_type k)
{
@@ -1913,6 +2083,7 @@
}
//! explicit conversion to FP.
+ //! @Returns the @c FP represented by @c *this.
template <typename FP>
FP as() const
{
@@ -1994,16 +2165,22 @@
return integer_part(x/factor<FP>());
}
- //! implicit conversion from int
+ /**
+ * Implicit conversion from arithmetic @T.
+ */
template <typename T>
- real_t(T x,
- typename boost::enable_if<is_implicitly_convertible<T, real_t> >::type* = 0
+ real_t(T x
+#if !defined(BOOST_FIXED_POINT_DOXYGEN_INVOKED)
+ , typename boost::enable_if<is_implicitly_convertible<T, real_t> >::type* = 0
+#endif
) : value_(fixed_point::detail::arthm_number_cast<T, real_t>()(x).count())
{}
- //! explicit conversion from int
+ //! explicit conversion from arithmetic @T.
template <typename T>
explicit real_t(T x
+#if !defined(BOOST_FIXED_POINT_DOXYGEN_INVOKED)
, typename boost::disable_if<is_implicitly_convertible<T ,real_t> >::type* = 0
+#endif
) : value_(fixed_point::detail::arthm_number_cast<T, real_t>()(x).count())
{}
@@ -2190,10 +2367,10 @@
*
* @TParams
* @Param{Range,Range specified by an integer. The range of a signed number x is 0<=x<2^Range},
- * @Param{Resolution,resolution specified by an integer. The resolution of a fractional number x is 2^Resolution.}
- * @Param{Rounding,The rounding policy.}
- * @Param{Overflow,The overflow policy.}
- * @Param{Family,The family policy.}
+ * @Param{Resolution,resolution specified by an integer. The resolution of a fractional number x is 2^Resolution}
+ * @Param{Rounding,The rounding policy}
+ * @Param{Overflow,The overflow policy}
+ * @Param{Family,The family policy}
*/
template <int Range, int Resolution, typename Rounding, typename Overflow, typename Family>
class ureal_t
@@ -2312,7 +2489,7 @@
* Explicit construction from an index.
*
* @Params
- * @Param{i,the index.}
+ * @Param{i,the index}
*
* @Requires <c>0<=i<=max_index</c>.
*/
@@ -2636,6 +2813,9 @@
/**
* ureal_t compile time factory.
+ * @TParams
+ * @Param{Times,an @c int representing the number of times}
+ * @Param{Resolution,the binary resolution}
*
* @Returns an @c ureal_t enough large to represent <c>Times*(2^Resolution)</c>.
*/
@@ -2653,8 +2833,12 @@
static_log2<mpl::abs<mpl::int_<Times+1> >::type::value>::value+Resolution, Resolution
>(index(Times));
}
+
/**
* real_t compile time factory.
+ * @TParams
+ * @Param{Times,an @c int representing the number of times}
+ * @Param{Resolution,the binary resolution}
*
* @Returns a @c real_t enough large to represent <c>Times*(2^Resolution)</c>.
*/
@@ -2673,7 +2857,7 @@
> (index(Times));
}
- // /**
+ // /*
// * ureal_t compile time factory from integer and fractional parts
// *
// * @Returns an @c ureal_t enough large to represent <c>Integral.Fractional</c>.
@@ -2709,7 +2893,11 @@
* Add type metafunction.
*
* The result type depends on whether the types are open/closed.
- * The specializations must define a nested typedef type.
+ *
+ * - Both are closed: The nested typedef type is only defined if @c is_same<T1,T2> and is @c T1.
+ * - One of them is Open:
+ * - if one of them is signed : real_t<MAX(R1,R2)+1, MIN(P1,P2), DT(RP1,RP2), DT(OP1,OP2), DT(F1,F2)>.
+ * - if both are unsigned : ureal_t<MAX(R1,R2)+1, MIN(P1,P2), DT(RP1,RP2), DT(OP1,OP2), DT(F1,F2)>.
*/
template <typename T1, typename T2=T1, bool B1=is_open<T1>::value, bool B2=is_open<T2>::value >
struct add_result
@@ -2733,8 +2921,7 @@
int R2, int P2, typename RP2, typename OP2, typename F2, bool B1, bool B2>
struct add_result<real_t<R1,P1,RP1,OP1,F1>, real_t<R2,P2,RP2,OP2,F2>, B1, B2 >
{
- typedef
- real_t<
+ typedef real_t<
mpl::max<mpl::int_<R1>,mpl::int_<R2> >::type::value+1,
mpl::min<mpl::int_<P1>,mpl::int_<P2> >::type::value,
typename default_type<RP1,RP2>::type,
@@ -2771,8 +2958,7 @@
int R2, int P2, typename RP2, typename OP2, typename F2, bool B1, bool B2>
struct add_result<ureal_t<R1,P1,RP1,OP1,F1>, ureal_t<R2,P2,RP2,OP2,F2>, B1, B2 >
{
- typedef
- ureal_t<
+ typedef ureal_t<
mpl::max<mpl::int_<R1>,mpl::int_<R2> >::type::value+1,
mpl::min<mpl::int_<P1>,mpl::int_<P2> >::type::value,
typename default_type<RP1,RP2>::type,
@@ -2784,7 +2970,10 @@
#endif
/**
- * signed + int -> signed.
+ * @Params
+ * @Param{lhs,a @c ureal_t}
+ * @Param{rhs,an @c int}
+ *
* @Returns <c>lhs+AT(rhs)</c>.
*/
template <int R1, int P1, typename RP1, typename OP1, typename F1>
@@ -2797,7 +2986,10 @@
return lhs + arg_type(rhs);
}
/**
- * unsigned + unsigned int -> signed.
+ * @Params
+ * @Param{lhs,a @c ureal_t}
+ * @Param{rhs,an <c>unsigned int</c>}
+ *
* @Returns <c>lhs+AT(rhs)</c>.
*/
template <int R1, int P1, typename RP1, typename OP1, typename F1>
@@ -2826,9 +3018,13 @@
return result_type(index(result_type(lhs).count()+result_type(rhs).count()));
}
#endif
+
/**
- * signed + signed -> signed.
- * @Returns <c>RT(index(RT(lhs).count()+RT(rhs).count())</c>.
+ * @Params
+ * @Param{lhs,a @c real_t}
+ * @Param{rhs,a @c real_t}
+ *
+ * @Returns <c>RT(index(RT(lhs).count() + RT(rhs).count())</c>.
*/
template <int R1, int P1, typename RP1, typename OP1, typename F1,
int R2, int P2, typename RP2, typename OP2, typename F2>
@@ -2842,8 +3038,11 @@
}
/**
- * unsigned + signed -> signed.
- * @Returns a signed fixed point enough large to avoid overflow.
+ * @Params
+ * @Param{lhs,a @c ureal_t}
+ * @Param{rhs,a @c real_t}
+ *
+ * @Returns <c>RT(index(RT(lhs).count() + RT(rhs).count())</c>.
*/
template <int R1, int P1, typename RP1, typename OP1, typename F1,
int R2, int P2, typename RP2, typename OP2, typename F2>
@@ -2856,8 +3055,11 @@
return result_type(index(result_type(lhs).count()+result_type(rhs).count()));
}
/**
- * signed + unsigned -> signed.
- * @Returns a signed fixed point enough large to avoid overflow.
+ * @Params
+ * @Param{lhs,a @c real_t}
+ * @Param{rhs,a @c ureal_t}
+ *
+ * @Returns <c>rhs + lhs</c>.
*/
template <int R1, int P1, typename RP1, typename OP1, typename F1,
int R2, int P2, typename RP2, typename OP2, typename F2>
@@ -2868,8 +3070,11 @@
return rhs+lhs;
}
/**
- * unsigned + unsigned -> unsigned.
- * @Returns a unsigned fixed point enough large to avoid overflow.
+ * @Params
+ * @Param{lhs,a @c ureal_t}
+ * @Param{rhs,a @c ureal_t}
+ *
+ * @Returns <c>RT(index(RT(lhs).count() + RT(rhs).count())</c>.
*/
template <int R1, int P1, typename RP1, typename OP1, typename F1,
int R2, int P2, typename RP2, typename OP2, typename F2>
@@ -2883,8 +3088,11 @@
}
/**
- * signed - signed -> signed.
- * @Returns a signed fixed point enough large to avoid overflow.
+ * @Params
+ * @Param{lhs,a @c real_t}
+ * @Param{rhs,a @c real_t}
+ *
+ * @Returns <c>RT(index(RT(lhs).count() - RT(rhs).count())</c>.
*/
template <int R1, int P1, typename RP1, typename OP1, typename F1,
int R2, int P2, typename RP2, typename OP2, typename F2>
@@ -2898,8 +3106,11 @@
}
/**
- * unsigned - signed -> signed.
- * @Returns a signed fixed point enough large to avoid overflow.
+ * @Params
+ * @Param{lhs,a @c ureal_t}
+ * @Param{rhs,a @c real_t}
+ *
+ * @Returns <c>RT(index(RT(lhs).count() - RT(rhs).count())</c>.
*/
template <int R1, int P1, typename RP1, typename OP1, typename F1,
int R2, int P2, typename RP2, typename OP2, typename F2>
@@ -2913,8 +3124,11 @@
}
/**
- * signed - unsigned -> signed.
- * @Returns a signed fixed point enough large to avoid overflow.
+ * @Params
+ * @Param{lhs,a @c real_t}
+ * @Param{rhs,a @c ureal_t}
+ *
+ * @Returns <c>RT(index(RT(lhs).count() - RT(rhs).count())</c>.
*/
template <int R1, int P1, typename RP1, typename OP1, typename F1,
int R2, int P2, typename RP2, typename OP2, typename F2>
@@ -2928,8 +3142,11 @@
}
/**
- * unsigned - unsigned -> signed.
- * @Returns a signed fixed point enough large to avoid overflow.
+ * @Params
+ * @Param{lhs,a @c ureal_t}
+ * @Param{rhs,a @c ureal_t}
+ *
+ * @Returns <c>RT(index(RT(lhs).count() - RT(rhs).count())</c>.
*/
template <int R1, int P1, typename RP1, typename OP1, typename F1,
int R2, int P2, typename RP2, typename OP2, typename F2>
@@ -2942,11 +3159,14 @@
return result_type(index(result_type(lhs).count()-result_type(rhs).count()));
}
- /*****************************************************************
+ /**
* Multiply type metafunction.
*
- * The result type depends on whether the types are open/closed.
- * The specializations must define a nested typedef type.
+ * The result type depends on whether the types are open/closed:
+ * - Both are closed: The nested typedef type is only defined if @c is_same<T1,T2> and is @c T1.
+ * - One of them is Open:
+ * - if one of them is signed : real_t<R1+R2, P1+P2, DT(RP1,RP2), DT(OP1,OP2), DT(F1,F2)>
+ * - if both are unsigned : ureal_t<R1+R2, P1+P2, DT(RP1,RP2), DT(OP1,OP2), DT(F1,F2)>
*/
template <typename T1, typename T2=T1, bool B1=is_open<T1>::value, bool B2=is_open<T2>::value >
struct multiply_result
@@ -2970,10 +3190,7 @@
int R2, int P2, typename RP2, typename OP2, typename F2, bool B1, bool B2>
struct multiply_result<real_t<R1,P1,RP1,OP1,F1>, real_t<R2,P2,RP2,OP2,F2>, B1, B2 >
{
- typedef
- real_t<
- R1+R2,
- P1+P2,
+ typedef real_t< R1+R2, P1+P2,
typename default_type<RP1,RP2>::type,
typename default_type<OP1,OP2>::type,
typename default_type<F1,F2>::type
@@ -2987,16 +3204,8 @@
template <int R1, int P1, typename RP1, typename OP1, typename F1,
int R2, int P2, typename RP2, typename OP2, typename F2, bool B1, bool B2>
struct multiply_result<real_t<R1,P1,RP1,OP1,F1>, ureal_t<R2,P2,RP2,OP2,F2>, B1, B2 >
+ : multiply_result<real_t<R1,P1,RP1,OP1,F1>, real_t<R2,P2,RP2,OP2,F2> >
{
- typedef
- real_t<
- R1+R2,
- P1+P2,
- typename default_type<RP1,RP2>::type,
- typename default_type<OP1,OP2>::type,
- typename default_type<F1,F2>::type
- >
- type;
};
/**
* When one of them is open, the result is open
@@ -3005,16 +3214,8 @@
template <int R1, int P1, typename RP1, typename OP1, typename F1,
int R2, int P2, typename RP2, typename OP2, typename F2, bool B1, bool B2>
struct multiply_result<ureal_t<R1,P1,RP1,OP1,F1>, real_t<R2,P2,RP2,OP2,F2>, B1, B2 >
+ : multiply_result<real_t<R1,P1,RP1,OP1,F1>, real_t<R2,P2,RP2,OP2,F2> >
{
- typedef
- real_t<
- R1+R2,
- P1+P2,
- typename default_type<RP1,RP2>::type,
- typename default_type<OP1,OP2>::type,
- typename default_type<F1,F2>::type
- >
- type;
};
/**
* When one of them is open, the result is open
@@ -3024,10 +3225,7 @@
int R2, int P2, typename RP2, typename OP2, typename F2, bool B1, bool B2>
struct multiply_result<ureal_t<R1,P1,RP1,OP1,F1>, ureal_t<R2,P2,RP2,OP2,F2>, B1, B2 >
{
- typedef
- ureal_t<
- R1+R2,
- P1+P2,
+ typedef ureal_t< R1+R2, P1+P2,
typename default_type<RP1,RP2>::type,
typename default_type<OP1,OP2>::type,
typename default_type<F1,F2>::type
@@ -3035,9 +3233,13 @@
type;
};
#endif
+
/**
- * signed * signed -> signed.
- * @Returns a signed fixed point enough large to avoid overflow.
+ * @Params
+ * @Param{lhs,a @c real_t}
+ * @Param{rhs,a @c real_t}
+ *
+ * @Returns <c>RT(index(RT(lhs).count() * RT(rhs).count())</c>.
*/
template <int R1, int P1, typename RP1, typename OP1, typename F1,
int R2, int P2, typename RP2, typename OP2, typename F2>
@@ -3052,8 +3254,11 @@
}
/**
- * signed * unsigned -> signed.
- * @Returns a signed fixed point enough large to avoid overflow.
+ * @Params
+ * @Param{lhs,a @c real_t}
+ * @Param{rhs,a @c ureal_t}
+ *
+ * @Returns <c>RT(index(RT(lhs).count() * RT(rhs).count())</c>.
*/
template <int R1, int P1, typename RP1, typename OP1, typename F1,
int R2, int P2, typename RP2, typename OP2, typename F2>
@@ -3068,8 +3273,11 @@
}
/**
- * unsigned * signed -> signed.
- * @Returns a signed fixed point enough large to avoid overflow.
+ * @Params
+ * @Param{lhs,a @c ureal_t}
+ * @Param{rhs,a @c real_t}
+ *
+ * @Returns <c>RT(index(RT(lhs).count() * RT(rhs).count())</c>.
*/
template <int R1, int P1, typename RP1, typename OP1, typename F1,
int R2, int P2, typename RP2, typename OP2, typename F2>
@@ -3084,8 +3292,11 @@
}
/**
- * unsigned * unsigned -> unsigned.
- * @Returns a unsigned fixed point enough large to avoid overflow.
+ * @Params
+ * @Param{lhs,a @c ureal_t}
+ * @Param{rhs,a @c ureal_t}
+ *
+ * @Returns <c>RT(index(RT(lhs).count() * RT(rhs).count())</c>.
*/
template <int R1, int P1, typename RP1, typename OP1, typename F1,
int R2, int P2, typename RP2, typename OP2, typename F2>
@@ -3100,8 +3311,8 @@
}
/**
- * fixed point division giving the expected result type.
- * @Returns DT(lhs) / DT(rhs) taking in account the rounding policy of the expected result.
+ * Fixed point division giving the expected result type.
+ * @Returns <c>DT(lhs) / DT(rhs)</c> taking in account the rounding policy of the result type @c Res.
*/
template <
typename Res,
@@ -3126,8 +3337,8 @@
}
/**
- * fixed point division giving the expected result type.
- * @Returns DT(lhs) / DT(rhs) taking in account the rounding policy of the expected result.
+ * Fixed point division giving the expected result type.
+ * @Returns <c>DT(lhs) / DT(rhs)</c> taking in account the rounding policy of the result type @c Res.
*/
template <
typename Res,
@@ -3152,7 +3363,7 @@
/**
* fixed point division giving the expected result type.
- * @Returns DT(lhs) / DT(rhs) taking in account the rounding policy of the expected result.
+ * @Returns <c>DT(lhs) / DT(rhs)</c> taking in account the rounding policy of the result type @c Res.
*/
template <
typename Res,
@@ -3177,7 +3388,7 @@
/**
* fixed point division giving the expected result type.
- * @Returns DT(lhs) / DT(rhs) taking in account the rounding policy of the expected result.
+ * @Returns <c>DT(lhs) / DT(rhs)</c> taking in account the rounding policy of the result type @c Res.
*/
template <
typename Res,
@@ -3200,11 +3411,14 @@
return result_type(index(rounding_type::template round_divide<Res>(DT(lhs), DT(rhs))));
}
- /*****************************************************************
+ /**
* Divide type metafunction.
*
* The result type depends on whether the types are open/closed.
- * The specializations must define a nested typedef type.
+ * - Both are closed: The nested typedef type is only defined if @c is_same<T1,T2> and is @c T1.
+ * - One of them is open:
+ * - if one of them is signed : <c>real_t<R1-P2, P1-R2, DT(RP1,RP2), DT(OP1,OP2), DT(F1,F2)></c>
+ * - if both are unsigned : <c>ureal_t<R1-P2, P1-R2, DT(RP1,RP2), DT(OP1,OP2), DT(F1,F2)></c>
*/
template <typename T1, typename T2=T1, bool B1=is_open<T1>::value, bool B2=is_open<T2>::value >
struct divide_result
@@ -3228,9 +3442,7 @@
int R2, int P2, typename RP2, typename OP2, typename F2, bool B1, bool B2>
struct divide_result<real_t<R1,P1,RP1,OP1,F1>, real_t<R2,P2,RP2,OP2,F2>, B1, B2 >
{
- typedef real_t<
- R1-P2,
- P1-R2,
+ typedef real_t< R1-P2, P1-R2,
typename default_type<RP1,RP2>::type,
typename default_type<OP1,OP2>::type,
typename default_type<F1,F2>::type
@@ -3244,15 +3456,8 @@
template <int R1, int P1, typename RP1, typename OP1, typename F1,
int R2, int P2, typename RP2, typename OP2, typename F2, bool B1, bool B2>
struct divide_result<ureal_t<R1,P1,RP1,OP1,F1>, real_t<R2,P2,RP2,OP2,F2>, B1, B2 >
+ : divide_result<real_t<R1,P1,RP1,OP1,F1>, real_t<R2,P2,RP2,OP2,F2> >
{
- typedef real_t<
- R1-P2,
- P1-R2,
- typename default_type<RP1,RP2>::type,
- typename default_type<OP1,OP2>::type,
- typename default_type<F1,F2>::type
- >
- type;
};
/**
* When one of them is open, the result is open
@@ -3261,15 +3466,8 @@
template <int R1, int P1, typename RP1, typename OP1, typename F1,
int R2, int P2, typename RP2, typename OP2, typename F2, bool B1, bool B2>
struct divide_result<real_t<R1,P1,RP1,OP1,F1>, ureal_t<R2,P2,RP2,OP2,F2>, B1, B2 >
+ : divide_result<real_t<R1,P1,RP1,OP1,F1>, real_t<R2,P2,RP2,OP2,F2> >
{
- typedef real_t<
- R1-P2,
- P1-R2,
- typename default_type<RP1,RP2>::type,
- typename default_type<OP1,OP2>::type,
- typename default_type<F1,F2>::type
- >
- type;
};
/**
* When one of them is open, the result is open
@@ -3279,9 +3477,7 @@
int R2, int P2, typename RP2, typename OP2, typename F2, bool B1, bool B2>
struct divide_result<ureal_t<R1,P1,RP1,OP1,F1>, ureal_t<R2,P2,RP2,OP2,F2>, B1, B2 >
{
- typedef ureal_t<
- R1-P2,
- P1-R2,
+ typedef ureal_t< R1-P2, P1-R2,
typename default_type<RP1,RP2>::type,
typename default_type<OP1,OP2>::type,
typename default_type<F1,F2>::type
@@ -3292,7 +3488,7 @@
/**
* fixed point division deducing the result type as <R1-P2, P1,R2>.
- * @Returns DT(lhs) / DT(rhs) taking in account the rounding policy of the result type.
+ * @Returns <c>divide<RT>(lhs,rhs)</c> taking in account the rounding policy of the result type.
*/
template <int R1, int P1, typename RP1, typename OP1, typename F1,
int R2, int P2, typename RP2, typename OP2, typename F2>
@@ -3307,7 +3503,7 @@
/**
* fixed point division deducing the result type as <R1-P2, P1,R2>.
- * @Returns DT(lhs) / DT(rhs) taking in account the rounding policy of the result type.
+ * @Returns <c>divide<RT>(lhs,rhs)</c> taking in account the rounding policy of the result type.
*/
template <int R1, int P1, typename RP1, typename OP1, typename F1,
int R2, int P2, typename RP2, typename OP2, typename F2>
@@ -3321,7 +3517,7 @@
/**
* fixed point division deducing the result type as <R1-P2, P1,R2>.
- * @Returns DT(lhs) / DT(rhs) taking in account the rounding policy of the result type.
+ * @Returns <c>divide<RT>(lhs,rhs)</c> taking in account the rounding policy of the result type.
*/
template <int R1, int P1, typename RP1, typename OP1, typename F1,
int R2, int P2, typename RP2, typename OP2, typename F2>
@@ -3335,7 +3531,7 @@
/**
* fixed point division deducing the result type as <R1-P2, P1,R2>.
- * @Returns DT(lhs) / DT(rhs) taking in account the rounding policy of the result type.
+ * @Returns <c>divide<RT>(lhs,rhs)</c> taking in account the rounding policy of the result type.
*/
template <int R1, int P1, typename RP1, typename OP1, typename F1,
int R2, int P2, typename RP2, typename OP2, typename F2>
@@ -3495,11 +3691,6 @@
return !(lhs < rhs);
}
- /**
- * fixed point number cast.
- *
- * @Returns the conversion with possible reduced range and loss of resolution.
- */
template <typename To, typename From>
BOOST_CONSTEXPR
To number_cast(From const& f)
Boost-Commit list run by bdawes at acm.org, david.abrahams at rcn.com, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk