nlohmann::detail::dtoa_impl Namespace Reference

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


struct  boundaries
struct  cached_power
struct  diyfp


JSON_HEDLEY_RETURNS_NON_NULL char * append_exponent (char *buf, int e)
 appends a decimal representation of e to buf More...
template<typename FloatType >
boundaries compute_boundaries (FloatType value)
int find_largest_pow10 (const std::uint32_t n, std::uint32_t &pow10)
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 More...
cached_power get_cached_power_for_binary_exponent (int e)
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)
void grisu2_digit_gen (char *buffer, int &length, int &decimal_exponent, diyfp M_minus, diyfp w, diyfp M_plus)
void grisu2_round (char *buf, int len, std::uint64_t dist, std::uint64_t delta, std::uint64_t rest, std::uint64_t ten_k)
template<typename Target , typename Source >
Target reinterpret_bits (const Source source)


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 (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* nlohmann::detail::dtoa_impl::append_exponent ( char *  buf,
int  e 

appends a decimal representation of e to buf

a pointer to the element following the exponent.
-1000 < e < 1000

Referenced by format_buffer().

◆ compute_boundaries()

template<typename FloatType >
boundaries nlohmann::detail::dtoa_impl::compute_boundaries ( FloatType  value)

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

value must be finite and positive

References nlohmann::detail::dtoa_impl::diyfp::e, nlohmann::detail::dtoa_impl::diyfp::normalize(), and nlohmann::detail::dtoa_impl::diyfp::normalize_to().

Referenced by grisu2().

◆ find_largest_pow10()

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

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

Referenced by grisu2_digit_gen().

◆ format_buffer()

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

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.

min_exp < 0
max_exp > 0

References append_exponent().

Referenced by nlohmann::detail::to_chars().

◆ get_cached_power_for_binary_exponent()

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

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.

References nlohmann::detail::dtoa_impl::cached_power::e, kAlpha, and kGamma.

Referenced by grisu2().

◆ grisu2() [1/2]

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

◆ grisu2() [2/2]

template<typename FloatType >
void nlohmann::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.

References compute_boundaries(), grisu2(), and nlohmann::detail::dtoa_impl::diyfp::kPrecision.

◆ grisu2_digit_gen()

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

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.

References DraftVecUtils::dist(), nlohmann::detail::dtoa_impl::diyfp::e, nlohmann::detail::dtoa_impl::diyfp::f, find_largest_pow10(), grisu2_round(), kAlpha, kGamma, and nlohmann::detail::dtoa_impl::diyfp::sub().

Referenced by grisu2().

◆ grisu2_round()

void nlohmann::detail::dtoa_impl::grisu2_round ( char *  buf,
int  len,
std::uint64_t  dist,
std::uint64_t  delta,
std::uint64_t  rest,
std::uint64_t  ten_k 

References DraftVecUtils::dist().

Referenced by grisu2_digit_gen().

◆ reinterpret_bits()

template<typename Target , typename Source >
Target nlohmann::detail::dtoa_impl::reinterpret_bits ( const Source  source)

Variable Documentation

◆ kAlpha

constexpr int nlohmann::detail::dtoa_impl::kAlpha = -60

◆ kGamma

constexpr int nlohmann::detail::dtoa_impl::kGamma = -32