detail::dtoa_impl Namespace Reference

implements the Grisu2 algorithm for binary to decimal floating-point conversion. More...

Classes
struct	boundaries
struct	cached_power
struct	diyfp

Functions
template<typename Target , typename Source >
Target	reinterpret_bits (const Source source)
template<typename FloatType >
boundaries	compute_boundaries (FloatType value)
cached_power	get_cached_power_for_binary_exponent (int e)
int	find_largest_pow10 (const std::uint32_t n, std::uint32_t &pow10)
void	grisu2_round (char *buf, int len, std::uint64_t dist, std::uint64_t delta, std::uint64_t rest, std::uint64_t ten_k)
void	grisu2_digit_gen (char *buffer, int &length, int &decimal_exponent, diyfp M_minus, diyfp w, diyfp M_plus)
void	grisu2 (char *buf, int &len, int &decimal_exponent, diyfp m_minus, diyfp v, diyfp m_plus)
template<typename FloatType >
void	grisu2 (char *buf, int &len, int &decimal_exponent, FloatType value)
JSON_HEDLEY_RETURNS_NON_NULL char *	append_exponent (char *buf, int e)
	appends a decimal representation of e to buf
JSON_HEDLEY_RETURNS_NON_NULL char *	format_buffer (char *buf, int len, int decimal_exponent, int min_exp, int max_exp)
	prettify v = buf * 10^decimal_exponent

Variables
constexpr int	kAlpha = -60
constexpr int	kGamma = -32

Detailed Description

implements the Grisu2 algorithm for binary to decimal floating-point conversion.

This implementation is a slightly modified version of the reference implementation which may be obtained from http://florian.loitsch.com/publications (bench.tar.gz).

The code is distributed under the MIT license, Copyright (c) 2009 Florian Loitsch.

For a detailed description of the algorithm see:

[1] Loitsch, "Printing Floating-Point Numbers Quickly and Accurately with Integers", Proceedings of the ACM SIGPLAN 2010 Conference on Programming Language Design and Implementation, PLDI 2010 [2] Burger, Dybvig, "Printing Floating-Point Numbers Quickly and Accurately", Proceedings of the ACM SIGPLAN 1996 Conference on Programming Language Design and Implementation, PLDI 1996

Function Documentation

append_exponent()

JSON_HEDLEY_RETURNS_NON_NULL char * detail::dtoa_impl::append_exponent ( char * buf, int e )

inline

appends a decimal representation of e to buf

Returns

a pointer to the element following the exponent.

Precondition

-1000 < e < 1000

compute_boundaries()

template<typename FloatType >

boundaries detail::dtoa_impl::compute_boundaries

(

FloatType

value

)

Compute the (normalized) diyfp representing the input number 'value' and its boundaries.

Precondition

value must be finite and positive

find_largest_pow10()

int detail::dtoa_impl::find_largest_pow10 ( const std::uint32_t n, std::uint32_t & pow10 )

inline

For n != 0, returns k, such that pow10 := 10^(k-1) <= n < 10^k. For n == 0, returns 1 and sets pow10 := 1.

format_buffer()

JSON_HEDLEY_RETURNS_NON_NULL char * detail::dtoa_impl::format_buffer ( char * buf, int len, int decimal_exponent, int min_exp, int max_exp )

inline

prettify v = buf * 10^decimal_exponent

If v is in the range [10^min_exp, 10^max_exp) it will be printed in fixed-point notation. Otherwise it will be printed in exponential notation.

Precondition

min_exp < 0

max_exp > 0

get_cached_power_for_binary_exponent()

cached_power detail::dtoa_impl::get_cached_power_for_binary_exponent ( int e )

inline

For a normalized diyfp w = f * 2^e, this function returns a (normalized) cached power-of-ten c = f_c * 2^e_c, such that the exponent of the product w * c satisfies (Definition 3.2 from [1])

 alpha <= e_c + e + q <= gamma.

grisu2() [1/2]

void detail::dtoa_impl::grisu2 ( char * buf, int & len, int & decimal_exponent, diyfp m_minus, diyfp v, diyfp m_plus )

inline

v = buf * 10^decimal_exponent len is the length of the buffer (number of decimal digits) The buffer must be large enough, i.e. >= max_digits10.

grisu2() [2/2]

template<typename FloatType >

void detail::dtoa_impl::grisu2	(	char *	buf,
		int &	len,
		int &	decimal_exponent,
		FloatType	value )

v = buf * 10^decimal_exponent len is the length of the buffer (number of decimal digits) The buffer must be large enough, i.e. >= max_digits10.

grisu2_digit_gen()

void detail::dtoa_impl::grisu2_digit_gen ( char * buffer, int & length, int & decimal_exponent, diyfp M_minus, diyfp w, diyfp M_plus )

inline

Generates V = buffer * 10^decimal_exponent, such that M- <= V <= M+. M- and M+ must be normalized and share the same exponent -60 <= e <= -32.