|
|
Boost-Commit : |
Subject: [Boost-commit] svn:boost r85516 - sandbox/multiprecision.cpp_bin_float/libs/multiprecision/doc
From: pbristow_at_[hidden]
Date: 2013-08-29 12:44:18
Author: pbristow
Date: 2013-08-29 12:44:18 EDT (Thu, 29 Aug 2013)
New Revision: 85516
URL: http://svn.boost.org/trac/boost/changeset/85516
Log:
1st commit for generator program for numeric_limits tables.
Added:
sandbox/multiprecision.cpp_bin_float/libs/multiprecision/doc/numeric_limits_qbk.cpp (contents, props changed)
Added: sandbox/multiprecision.cpp_bin_float/libs/multiprecision/doc/numeric_limits_qbk.cpp
==============================================================================
--- /dev/null 00:00:00 1970 (empty, because file is newly added)
+++ sandbox/multiprecision.cpp_bin_float/libs/multiprecision/doc/numeric_limits_qbk.cpp 2013-08-29 12:44:18 EDT (Thu, 29 Aug 2013) (r85516)
@@ -0,0 +1,721 @@
+// Copyright Paul A. Bristow 2013
+
+// Use, modification and distribution are subject to the
+// Boost Software License, Version 1.0.
+// (See accompanying file LICENSE_1_0.txt
+// or copy at http://www.boost.org/LICENSE_1_0.txt)
+
+// Program to list all numeric_limits items for any type to a file in Quickbook format.
+
+// C standard http://www.open-std.org/jtc1/sc22/wg11/docs/n507.pdf
+// SC22/WG11 N507 DRAFT INTERNATIONAL ISO/IEC STANDARD WD 10967-1
+// Information technology Language independent arithmetic Part 1: Integer and Floating point arithmetic
+
+/* E.3 C++
+The programming language C++ is defined by ISO/IEC 14882:1998, Programming languages C++ [18].
+An implementation should follow all the requirements of LIA-1 unless otherwise specified by
+this language binding.
+*/
+
+// http://754r.ucbtest.org/standards/754.pdf
+
+// http://www.cesura17.net/~will/Professional/Research/Papers/retrospective.pdf
+
+// http://www.exploringbinary.com/using-integers-to-check-a-floating-point-approximation/
+
+// http://stackoverflow.com/questions/12466745/how-to-convert-float-to-doubleboth-stored-in-ieee-754-representation-without-loss
+
+
+#ifdef _MSC_VER
+# pragma warning (disable : 4127) // conditional expression is constant.
+# pragma warning (disable : 4100) // unreferenced formal parameter.
+#endif
+
+
+#include <iostream>
+#include <iomanip>
+#include <string>
+#include <sstream>
+#include <fstream>
+#include <limits> // numeric_limits
+
+#include <typeinfo>
+
+#include <boost/version.hpp>
+#include <boost/config.hpp>
+
+// May need extra includes for other types, for example:
+#include <boost/multiprecision/cpp_dec_float.hpp> // is decimal.
+#include <boost/multiprecision/cpp_bin_float.hpp> // is binary.
+
+
+// Assume that this will be run on MSVC to get the 32 or 64 bit info.
+#ifdef _WIN32
+ std::string bits32_64 = "32";
+std::string filename = "I:/boost-sandbox/multiprecision.cpp_bin_float/libs/multiprecision/doc/numeric_limits_32_tables.qbk";
+#else
+ std::string bits32_64 = "64";
+std::string filename = "I:/boost-sandbox/multiprecision.cpp_bin_float/libs/multiprecision/doc/numeric_limits_64_tables.qbk";
+#endif
+
+#ifdef INT32_T_MAX
+ int i = INT256_T_MAX;
+#endif
+
+std::array<std::string, 16> integer_type_names =
+{
+"bool",
+"char",
+"unsigned char",
+"char16_t",
+"char32_t",
+"short",
+"unsigned short",
+"int",
+"unsigned int",
+"long",
+"unsigned long",
+"long long",
+"unsigned long long",
+"int32_t",
+//"uint32_t",
+"int64_t",
+//"uint64_t",
+"int128_t",
+//"uint128_t" // Too big?
+//"int256_t",
+//"uint256_t"
+//"int512_t"
+};
+
+std::array<std::string, 6> float_type_names =
+{
+ "function", "float", "double", "long double", "cpp_dec_50", "cpp_bin_128"
+};
+
+// Table headings for integer constants.
+std::array<std::string, 8> integer_constant_heads =
+{
+ "type", "signed", "bound", "modulo", "round", "radix", "digits", "digits10", // "max_digits10"
+};
+
+// Table headings for integer functions.
+std::array<std::string, 2> integer_function_heads =
+{
+ "max", // "lowest", assumes is same for all integer types, so not worth listing.
+ "min"
+};
+
+// Table headings for float constants.
+std::array<std::string, 12> float_constant_heads =
+{
+ "type", // "signed", "exact", "bound", // "modulo",
+ "round", "radix", "digits", "digits10", "max_digits10", "min_exp", "min_exp10", "max_exp", "max_exp10", "tiny", "trap"
+};
+
+// Table headings for float functions.
+std::array<std::string, 10> float_function_heads =
+{
+ "function", "max", "lowest", "min", "eps", "round", "infinity", "NaN", "sig_NaN", "denorm_min"
+};
+
+std::string versions()
+{ // Build a string of info about Boost, platform, STL, etc.
+ std::stringstream mess;
+ //mess << "\n" << "Program:\n\" " __FILE__ << "\"\n"; // \n is mis-interpreted!
+ mess << "\n" << "Program:\n numeric_limits_qbk.cpp \n";
+#ifdef __TIMESTAMP__
+ mess << __TIMESTAMP__;
+#endif
+ mess << "\nBuildInfo:\n" " Platform " << BOOST_PLATFORM;
+ mess << "\n Compiler " BOOST_COMPILER;
+#ifdef _MSC_FULL_VER
+ mess << "\n MSVC version "<< BOOST_STRINGIZE(_MSC_FULL_VER) << ".";
+#endif
+ mess << "\n STL " BOOST_STDLIB;
+ mess << "\n Boost version " << BOOST_VERSION/100000 << "." << BOOST_VERSION/100 % 1000 << "." << BOOST_VERSION % 100 << std::endl;
+ return mess.str();
+} // std::string versions()
+
+template <typename T>
+void out_round_style(std::ostream& os)
+{ //! Send short string describing STD::round_style to stream os.
+ // os << "Round style is ";
+ if (std::numeric_limits<T>::round_style == std::round_indeterminate)
+ {
+ os << "indeterminate" ;
+ }
+ else if (std::numeric_limits<T>::round_style == std::round_toward_zero)
+ {
+ os << "to zero" ;
+ }
+ else if (std::numeric_limits<T>::round_style == std::round_to_nearest)
+ {
+ os << "to nearest" ;
+ }
+ else if (std::numeric_limits<T>::round_style == std::round_toward_infinity)
+ {
+ os << "to infin[]"; // Or << "to \u221E" << "to infinity" ;
+ }
+ else if (std::numeric_limits<T>::round_style == std::round_toward_neg_infinity)
+ {
+ os << "to -infin[]" ;
+ }
+ else
+ {
+ os << "undefined!";
+ std::cout << "std::numeric_limits<T>::round_style is undefined value!" << std::endl;
+ }
+ return;
+} // out_round_style(std::ostream& os);
+
+template<typename T>
+void integer_constants(std::string type_name, std::ostream& os)
+{ //! Output a line of table integer constant values to ostream os.
+ os << "\n["
+ "[" << type_name << "]" ;
+ os << "[" << (std::numeric_limits<T>::is_signed ? "signed" : "unsigned") << "]" ;
+ // Is always exact for integer types, so removed:
+ // os << "[" << (std::numeric_limits<T>::is_exact ? "exact" : "inexact") << "]" ;
+ os << "[" << (std::numeric_limits<T>::is_bounded ? "bound" : "unbounded") << "]" ;
+ os << "[" << (std::numeric_limits<T>::is_modulo ? "modulo" : "no") << "]" ;
+ os << "[" ; out_round_style<T>(os); os << "]" ;
+ os << "[" << std::numeric_limits<T>::radix << "]" ;
+ os << "[" << std::numeric_limits<T>::digits << "]" ;
+ os << "[" << std::numeric_limits<T>::digits10 << "]" ;
+ // Undefined for integers, so removed:
+ // os << "[" << std::numeric_limits<T>::max_digits10 << "]"
+ os << "]";
+} // void integer_constants
+
+
+template<typename T>
+void float_constants(std::string type_name, std::ostream& os)
+{ //! Output a row of table values to `ostream` os.
+ os << "\n["
+ "[" << type_name << "]" ;
+ //os << "[" << (std::numeric_limits<T>::is_signed ? "signed" : "unsigned") << "]" ;
+ //os << "[" << (std::numeric_limits<T>::is_exact ? "exact" : "inexact") << "]" ;
+ //os << "[" << (std::numeric_limits<T>::is_bounded ? "bound" : "no") << "]" ;
+ // os << "[" << (std::numeric_limits<T>::is_modulo ? "modulo" : "no") << "]" ;
+ os << "[" ; out_round_style<T>(os); os << "]" ;
+ os << "[" << std::numeric_limits<T>::radix << "]" ;
+ os << "[" << std::numeric_limits<T>::digits << "]" ;
+ os << "[" << std::numeric_limits<T>::digits10 << "]" ;
+ os << "[" << std::numeric_limits<T>::max_digits10 << "]";
+ os << "[" << std::numeric_limits<T>::min_exponent << "]" ;
+ os << "[" << std::numeric_limits<T>::min_exponent10 << "]" ;
+ os << "[" << std::numeric_limits<T>::max_exponent << "]" ;
+ os << "[" << std::numeric_limits<T>::max_exponent10 << "]" ;
+ os << "[" << (std::numeric_limits<T>::tinyness_before ? "tiny" : "no") << "]" ;
+ os << "[" << (std::numeric_limits<T>::traps ? "traps" : "no") << "]" ;
+ os << "]" "\n"; // end of row.
+} // void float_constants
+
+/* Types across and two functions down.
+
+template<typename T>
+void integer_functions(std::string type_name, std::ostream& os)
+{ //! Output a line of table integer function values to `ostream` os.
+ os << "\n["
+ "[" << type_name << "]" ;
+ os << "[" << std::numeric_limits<T>::max() << "]" ;
+ //os << "[" << std::numeric_limits<T>::lowest() << "]" ; always == min for integer types,
+ // so removed to save space.
+ os << "[" << std::numeric_limits<T>::min() << "]"
+ "]";
+} // void integer_functions
+
+*/
+
+// Types down and two (or three) functions across.
+template<typename T>
+void integer_functions(std::string type_name, std::ostream& os)
+{ //! Output a line of table integer function values to `ostream` os.
+ os << "\n[" // start of row.
+ "[" << type_name << "]" ;
+ os << "[" << std::numeric_limits<T>::max() << "]" ;
+ // os << "[" << std::numeric_limits<T>::lowest() << "]" ;
+ os << "[" << std::numeric_limits<T>::min() << "]" ;
+ os << "]"; // end of row.
+} // void integer_functions
+
+
+template<typename T>
+void float_functions(std::string type_name, std::ostream& os)
+{ //! Output a line of table float-point function values to `ostream` os.
+ os << "\n[" // start of row.
+ "[" << type_name << "]" ;
+ os << "[" << std::numeric_limits<T>::max() << "]" ;
+ os << "[" << std::numeric_limits<T>::lowest() << "]" ;
+ os << "[" << std::numeric_limits<T>::min() << "]"
+ os << "[" << std::numeric_limits<T>::epsilon() << "]"
+ os << "[" << std::numeric_limits<T>::round_error() << "]"
+ os << "[" << std::numeric_limits<T>::infinity() << "]"
+ os << "[" << std::numeric_limits<T>::quiet_NaN() << "]"
+ os << "[" << std::numeric_limits<T>::signaling_NaN() << "]"
+ os << "[" << std::numeric_limits<T>::denorm_min() << "]"
+ "]"; // end of row.
+} // void float_functions
+
+template<typename T>
+int numeric_limits_list(std::string description)
+{//! Output numeric_limits for numeric_limits<T>, for example `numeric_limits_list<bool>()`.
+ // This is not used for Quickbook format.
+ // std::cout << versions() << std::endl;
+
+ std::cout << "\n" << description << "\n" << std::endl; // int, int64_t rather than full long typeid(T).name().
+
+ std::cout << "Type " << typeid(T).name() << " std::numeric_limits::<" << typeid(T).name() << ">\n" << std::endl;
+ // ull long typeid(T).name()
+
+ if (std::numeric_limits<T>::is_specialized == false)
+ {
+ std::cout << "type " << typeid(T).name() << " is not specialized for std::numeric_limits!" << std::endl;
+ //return -1;
+ }
+
+ // Member constants.
+
+ std::cout << (std::numeric_limits<T>::is_signed ? "is signed." : "unsigned.") << std::endl;
+ std::cout << (std::numeric_limits<T>::is_integer ? "is integer." : "not integer (fixed or float-point).") << std::endl;
+ std::cout << (std::numeric_limits<T>::is_exact ? "is exact." : "not exact.") << std::endl;
+ std::cout << (std::numeric_limits<T>::has_infinity ? "has infinity." : "no infinity.") << std::endl;
+ std::cout << (std::numeric_limits<T>::has_quiet_NaN ? "has quiet NaN." : "no quiet NaN.") << std::endl;
+ std::cout << (std::numeric_limits<T>::has_signaling_NaN ? "has signaling NaN." : "no signaling NaN.") << std::endl;
+ if (!std::numeric_limits<T>::is_integer)
+ { // is floating_point
+ std::cout << "Denorm style is " ;
+ if (std::numeric_limits<T>::has_denorm == std::denorm_absent)
+ {
+ std::cout << "denorm_absent." << std::endl;
+ }
+ else if (std::numeric_limits<T>::has_denorm == std::denorm_present)
+ {
+ std::cout << "denorm_present." << std::endl;
+ }
+ else if (std::numeric_limits<T>::has_denorm == std::denorm_indeterminate)
+ {
+ std::cout << "denorm_indeterminate." << std::endl;
+ }
+ else
+ {
+ std::cout << "std::numeric_limits<T>::has_denorm unexpected value!" << std::endl;
+ }
+
+ std::cout << (std::numeric_limits<T>::has_denorm_loss ? "has denorm loss." : "no denorm loss.") << std::endl;
+ // true if a loss of accuracy is detected as a denormalization loss, rather than an inexact result.
+
+ std::cout << "Round style is ";
+ if (std::numeric_limits<T>::round_style == std::round_indeterminate)
+ {
+ std::cout << "round_indeterminate!" << std::endl;
+ }
+ else if (std::numeric_limits<T>::round_style == std::round_toward_zero)
+ {
+ std::cout << "round_toward_zero." << std::endl;
+ }
+ else if (std::numeric_limits<T>::round_style == std::round_to_nearest)
+ {
+ std::cout << "round_to_nearest." << std::endl;
+ }
+ else if (std::numeric_limits<T>::round_style == std::round_toward_infinity)
+ {
+ std::cout << "round_toward_infinity." << std::endl;
+ }
+ else if (std::numeric_limits<T>::round_style == std::round_toward_neg_infinity)
+ {
+ std::cout << "round_toward_neg_infinity." << std::endl;
+ }
+ else
+ {
+ std::cout << "undefined value!" << std::endl;
+ }
+
+ } // is floating_point
+
+ std::cout << (std::numeric_limits<T>::is_iec559 ? "is IEC599." : "not IEC599.") << std::endl;
+ std::cout << (std::numeric_limits<T>::is_bounded ? "is bound." : "unbounded.") << std::endl;
+ std::cout << (std::numeric_limits<T>::is_modulo ? "is modulo." : "no modulo.") << std::endl;
+ std::cout << std::dec << "radix " << std::numeric_limits<T>::radix << std::endl;
+ std::cout << "digits " << std::numeric_limits<T>::digits << std::endl;
+ std::cout << "digits10 " << std::numeric_limits<T>::digits10 << std::endl;
+
+ std::cout.precision(std::numeric_limits<T>::max_digits10); // Full precision for floating-point values like max, min ...
+
+ std::cout << "max_digits10 " << std::numeric_limits<T>::max_digits10 << std::endl;
+ std::cout << "min_exponent " << std::numeric_limits<T>::min_exponent << std::endl;
+ std::cout << "min_exponent10 " << std::numeric_limits<T>::min_exponent10 << std::endl;
+ std::cout << "max_exponent " << std::numeric_limits<T>::max_exponent << std::endl;
+ std::cout << "max_exponent10 " << std::numeric_limits<T>::max_exponent10 << std::endl;
+
+ std::cout << (std::numeric_limits<T>::tinyness_before ? "Can tiny values before rounding." : "no tinyness_before.") << std::endl;
+ // true if the type can detect tiny values before rounding; false if it cannot.
+ std::cout << (std::numeric_limits<T>::traps ? "traps" : "no trapping.") << std::endl;
+ // Whether trapping that reports on arithmetic exceptions is implemented for a type.
+
+ std::cout << "Member functions." << std::endl;
+ // (assumes operator<< for type T is available).
+ // If floating-point then hex format may not be available.
+
+ std::cout << "max = " << std::numeric_limits<T>::max() << std::endl;
+ //if (std::numeric_limits<T>::is_integer)
+ //{
+ // std::cout << " = " << std::hex << std::numeric_limits<T>::max() << std::endl;
+ //}
+
+ std::cout << "lowest = " << std::dec << std::numeric_limits<T>::lowest() << std::endl;
+ //if (std::numeric_limits<T>::is_integer)
+ //{
+ // std::cout << " = " << std::hex << std::numeric_limits<T>::lowest() << std::endl;
+ //}
+
+ std::cout << "min = " << std::dec << std::numeric_limits<T>::min() << std::endl;
+ //if (std::numeric_limits<T>::is_integer)
+ //{
+ // std::cout << " = " << std::hex << std::numeric_limits<T>::min() << std::endl;
+ //}
+
+ std::cout << "epsilon = " << std::dec << std::numeric_limits<T>::epsilon() << std::endl;
+ //if (std::numeric_limits<T>::is_integer)
+ //{
+ // std::cout << " = " << std::hex << std::numeric_limits<T>::epsilon() << std::endl;
+ //}
+ // round_error is always zero for integer types.
+ // round_error is usually 1/2 = (T)0.5 for floating-point types.
+ // round_error is ? for fixed-point.
+ std::cout << "round_error = " << std::numeric_limits<T>::round_error() << " ULP." << std::endl;
+
+ std::cout << "infinity = " << std::dec << std::numeric_limits<T>::infinity() << std::endl;
+ std::cout << " = " << std::hex << std::numeric_limits<T>::infinity() << std::endl;
+
+ std::cout << "quiet_NaN = " << std::dec << std::numeric_limits<T>::quiet_NaN() << std::endl;
+ std::cout << " = " << std::hex << std::numeric_limits<T>::quiet_NaN() << std::endl;
+
+ std::cout << "signaling_NaN = " << std::dec << std::numeric_limits<T>::signaling_NaN() << std::endl;
+ std::cout << " = " << std::hex << std::numeric_limits<T>::signaling_NaN() << std::endl;
+
+ // Only meaningful if (std::numeric_limits<T>::has_denorm == std::denorm_present)
+ // so might not bother to show if absent?
+ std::cout << "denorm_min = " << std::numeric_limits<T>::denorm_min() << std::endl;
+ std::cout << " = " << std::numeric_limits<T>::denorm_min() << std::endl;
+ return 0;
+}
+
+int main()
+{
+
+
+
+ try
+ {
+ using namespace boost::multiprecision;
+
+ std::cout << versions() << std::endl;
+
+ std::ofstream fout(filename, std::ios_base::out);
+ if (!fout.is_open())
+ {
+ std::cout << "Unable to open file " << filename << " for output.\n" << std::endl;
+ return -1; // boost::EXIT_FAILURE;
+ }
+ fout <<
+ "[/""\n"
+ "Copyright 2013 Paul A. Bristow.""\n"
+ "Copyright 2013 John Maddock.""\n"
+ "Distributed under the Boost Software License, Version 1.0.""\n"
+ "(See accompanying file LICENSE_1_0.txt or copy at""\n"
+ "http://www.boost.org/LICENSE_1_0.txt).""\n"
+ "]""\n"
+ << std::endl;
+
+ // Output platform version info (32 or 64).
+ fout << "[section:version_" << bits32_64 << " "
+ << bits32_64 << "-bit version information used for tables below." "]\n"
+ "[pre""\n"
+ << versions()
+ << "]""\n"
+ "[endsect] [/section:version_32]" "\n"
+ << std::endl;
+
+ fout << "[section:limits"<< bits32_64 << " Numeric limits for " << bits32_64 << "-bit platform]" << std::endl;
+
+ fout << "[table:integral_constants Integer types constants (`std::numeric_limits<T>::is_integer == true` && is_exact == true)" "\n"
+ "[";
+
+ for (size_t i = 0; i < integer_constant_heads.size(); i++)
+ { // signed, bound, modulo ...
+ fout << "[" << integer_constant_heads[i] << "]" ;
+ }
+ fout << "]";
+
+ integer_constants<bool>("bool", fout);
+ integer_constants<char>("char", fout);
+ integer_constants<unsigned char>("unsigned char", fout);
+ integer_constants<char16_t>("char16_t", fout);
+ integer_constants<char32_t>("char32_t", fout);
+ integer_constants<short>("short", fout);
+ integer_constants<unsigned short>("unsigned short", fout);
+ integer_constants<int>("int", fout);
+ integer_constants<unsigned int>("unsigned", fout);
+ integer_constants<long>("long", fout);
+ integer_constants<unsigned long>("unsigned long", fout);
+ integer_constants<long long>("long long", fout);
+ integer_constants<unsigned long long>("unsigned long long", fout);
+ integer_constants<int32_t>("int32_t", fout);
+ integer_constants<uint32_t>("uint32_t", fout);
+ integer_constants<int64_t>("int64_t", fout);
+ integer_constants<uint64_t>("uint64_t", fout);
+ integer_constants<int128_t>("int128_t", fout);
+ integer_constants<uint128_t>("uint128_t", fout);
+ integer_constants<int256_t>("int256_t", fout);
+ integer_constants<uint256_t>("uint256_t", fout);
+ // integer_constants<int512_t>("int512_t", fout);
+ //integer_constants<uint512_t>("uint512_t", fout); // Too big?
+ integer_constants<cpp_int>("cpp_int", fout);
+
+ fout << "\n]\n\n";
+
+
+ fout << "[table:integral_functions Integer types functions (`std::numeric_limits<T>::is_integer == true && std::numeric_limits<T>::min() == std::numeric_limits<T>::lowest()` )" "\n"
+ "[";
+ // Display types across the page, and 2 (or 3) functions as rows.
+
+ fout << "[function]";
+ for (size_t i = 0; i < integer_function_heads.size(); i++)
+ {
+ fout << "[" << integer_function_heads[i] << "]" ;
+ }
+ fout << "]"; // end of headings.
+ integer_functions<bool>("bool", fout);
+ //integer_functions<char>("char", fout); // Need int value not char.
+ fout << "\n[" // start of row.
+ "[" << "char"<< "]" ;
+ fout << "[" << static_cast<int>((std::numeric_limits<char>::max)()) << "]" ;
+ // fout << "[" << (std::numeric_limits<T>::lowest)() << "]" ;
+ fout << "[" << static_cast<int>((std::numeric_limits<char>::min)()) << "]" ;
+ fout << "]"; // end of row.
+ //integer_functions<unsigned char>("unsigned char", fout); // Need int value not char.
+ fout << "\n[" // start of row.
+ "[" << "unsigned char"<< "]" ;
+ fout << "[" << static_cast<int>((std::numeric_limits<unsigned char>::max)()) << "]" ;
+ // fout << "[" << std::numeric_limits<unsigned char>::lowest() << "]" ;
+ fout << "[" << static_cast<int>(std::numeric_limits<unsigned char>::min()) << "]" ;
+ fout << "]"; // end of row.
+
+ integer_functions<char16_t>("char16_t", fout);
+ integer_functions<char32_t>("char32_t", fout);
+ integer_functions<short>("short", fout);
+ integer_functions<unsigned short>("unsigned short", fout);
+ integer_functions<int>("int", fout);
+ integer_functions<unsigned int>("unsigned int", fout);
+ integer_functions<long>("long", fout);
+ integer_functions<unsigned long>("unsigned long", fout);
+ integer_functions<long long>("long long", fout);
+ integer_functions<unsigned long long>("unsigned long long", fout);
+ integer_functions<int32_t>("int32_t", fout);
+ integer_functions<int64_t>("int64_t", fout);
+ integer_functions<int128_t>("int128_t", fout);
+ fout << "]" "\n"; // end of table;
+
+
+ //fout << "[[max]"
+ // << "[" << std::numeric_limits<bool>::max() << "]"
+ // << "[" << static_cast<int>(std::numeric_limits<char>::max()) << "]"
+ // << "[" << static_cast<int>(std::numeric_limits<unsigned char>::max()) << "]"
+ // << "[" << static_cast<int>(std::numeric_limits<char16_t>::max()) << "]"
+ // << "[" << static_cast<int>(std::numeric_limits<char32_t>::max()) << "]"
+ // << "[" << std::numeric_limits<short>::max() << "]"
+ // << "[" << std::numeric_limits<unsigned short>::max() << "]"
+ // << "[" << std::numeric_limits<int>::max() << "]"
+ // << "[" << std::numeric_limits<unsigned int>::max() << "]"
+ // << "[" << std::numeric_limits<long>::max() << "]"
+ // << "[" << std::numeric_limits<unsigned long>::max() << "]"
+ // << "[" << std::numeric_limits<long long>::max() << "]"
+ // << "[" << std::numeric_limits<unsigned long long>::max() << "]"
+ // << "[" << std::numeric_limits<int32_t>::max() << "]"
+ // << "[" << std::numeric_limits<int64_t>::max() << "]"
+ // << "[" << std::numeric_limits<int128_t>::max() << "]"
+ // //<< "[" << std::numeric_limits<int256_t>::max() << "]" // too big?
+ // //<< "[" << std::numeric_limits<int512_t>::max() << "]" // too big?
+ // << "]" "\n";
+ ///* Assume lowest() is not useful as == min for all integer types.
+ // */
+
+ //fout << "[[min]"
+ // << "[" << std::numeric_limits<bool>::min() << "]"
+ // << "[" << static_cast<int>(std::numeric_limits<char>::min()) << "]"
+ // << "[" << static_cast<int>(std::numeric_limits<unsigned char>::min()) << "]"
+ // << "[" << static_cast<int>(std::numeric_limits<char16_t>::min()) << "]"
+ // << "[" << static_cast<int>(std::numeric_limits<char32_t>::min()) << "]"
+ // << "[" << std::numeric_limits<short>::min() << "]"
+ // << "[" << std::numeric_limits<unsigned short>::min() << "]"
+ // << "[" << std::numeric_limits<int>::min() << "]"
+ // << "[" << std::numeric_limits<unsigned int>::min() << "]"
+ // << "[" << std::numeric_limits<long>::min() << "]"
+ // << "[" << std::numeric_limits<unsigned long>::min() << "]"
+ // << "[" << std::numeric_limits<long long>::min() << "]"
+ // << "[" << std::numeric_limits<unsigned long long>::min() << "]"
+ // << "[" << std::numeric_limits<int32_t>::min() << "]"
+ // << "[" << std::numeric_limits<int64_t>::min() << "]"
+ // << "[" << std::numeric_limits<int128_t>::min() << "]"
+ // // << "[" << std::numeric_limits<int256_t>::min() << "]" // too big?
+ // // << "[" << std::numeric_limits<int512_t>::min() << "]" // too big?
+ // << "]""\n";
+
+
+
+ // Floating-point
+
+ typedef number<cpp_dec_float<50> > cpp_dec_float_50; // 50 decimal digits.
+ typedef number<cpp_bin_float<113> > bin_128bit_double_type; // == Binary rare long double.
+
+ fout <<
+ //"[table:float_functions Floating-point types constants (`std::numeric_limits<T>::is_integer == false && std::numeric_limits<T>::is_modulo == false` )" "\n"
+ "[table:float_functions Floating-point types constants (`std::numeric_limits<T>::is_integer==false && is_signed==true && is_modulo==false && is_exact==false && is_bound==true`)" "\n"
+ "[";
+ for (size_t i = 0; i < float_constant_heads.size(); i++)
+ {
+ fout << "[" << float_constant_heads[i] << "]" ;
+ }
+ fout << "]"; // end of headings.
+
+ float_constants<float>("float", fout);
+ float_constants<double>("double", fout);
+ float_constants<long double>("long double", fout);
+ float_constants<cpp_dec_float_50>("cpp_dec_float_50", fout);
+ float_constants<bin_128bit_double_type>("bin_128bit_double_type", fout);
+ fout << "]" "\n"; // end of table;
+
+ fout <<
+ "[table:float_functions Floating-point types functions (`std::numeric_limits<T>::is_integer == false`)" "\n"
+ "[";
+
+ for (size_t i = 0; i < float_type_names.size(); i++)
+ {
+ fout << "[" << float_type_names[i] << "]" ;
+ }
+ fout << "]"; // end of headings.
+
+ fout << "[[max]"
+ << "[" << std::numeric_limits<float>::max() << "]"
+ << "[" << std::numeric_limits<double>::max() << "]"
+//#if LDBL_MANT_DIG > DBL_MANT_DIG
+ // Perhaps to test Long double is not just a duplication of double (but need change is headings too).
+ << "[" << std::numeric_limits<long double>::max() << "]"
+//#endif
+ << "[" << std::numeric_limits<cpp_dec_float_50>::max() << "]"
+ << "[" << std::numeric_limits<bin_128bit_double_type >::max() << "]"
+ << "]" "\n"; // end of row.
+
+ fout << "[[min]"
+ << "[" << std::numeric_limits<float>::min() << "]"
+ << "[" << std::numeric_limits<double>::min() << "]"
+//#if LDBL_MANT_DIG > DBL_MANT_DIG
+ // Long double is not just a duplication of double.
+ << "[" << std::numeric_limits<long double>::min() << "]"
+//#endif
+ << "[" << std::numeric_limits<cpp_dec_float_50 >::min() << "]"
+ << "[" << std::numeric_limits<bin_128bit_double_type >::min() << "]"
+ << "]" "\n"; // end of row.
+
+ fout << "[[epsilon]"
+ << "[" << std::numeric_limits<float>::epsilon() << "]"
+ << "[" << std::numeric_limits<double>::epsilon() << "]"
+//#if LDBL_MANT_DIG > DBL_MANT_DIG
+ // Long double is not just a duplication of double.
+ << "[" << std::numeric_limits<long double>::epsilon() << "]"
+//#endif
+ << "[" << std::numeric_limits<cpp_dec_float_50 >::epsilon() << "]"
+ << "[" << std::numeric_limits<bin_128bit_double_type >::epsilon() << "]"
+ << "]" "\n"; // end of row.
+
+ fout << "[[round_error]"
+ << "[" << std::numeric_limits<float>::round_error() << "]"
+ << "[" << std::numeric_limits<double>::round_error() << "]"
+//#if LDBL_MANT_DIG > DBL_MANT_DIG
+ // Long double is not just a duplication of double.
+ << "[" << std::numeric_limits<long double>::round_error() << "]"
+//#endif
+ << "[" << std::numeric_limits<cpp_dec_float_50 >::round_error() << "]"
+ << "[" << std::numeric_limits<bin_128bit_double_type >::round_error() << "]"
+ << "]" "\n"; // end of row.
+
+ fout << "[[infinity]"
+ << "[" << std::numeric_limits<float>::infinity() << "]"
+ << "[" << std::numeric_limits<double>::infinity() << "]"
+//#if LDBL_MANT_DIG > DBL_MANT_DIG
+ // Long double is not just a duplication of double.
+ << "[" << std::numeric_limits<long double>::infinity() << "]"
+//#endif
+ << "[" << std::numeric_limits<cpp_dec_float_50 >::infinity() << "]"
+ << "[" << std::numeric_limits<bin_128bit_double_type >::infinity() << "]"
+ << "]" "\n"; // end of row.
+
+ fout << "[[quiet_NaN]"
+ << "[" << std::numeric_limits<float>::quiet_NaN() << "]"
+ << "[" << std::numeric_limits<double>::quiet_NaN() << "]"
+//#if LDBL_MANT_DIG > DBL_MANT_DIG
+ // Long double is not just a duplication of double.
+ << "[" << std::numeric_limits<long double>::quiet_NaN() << "]"
+//#endif
+ << "[" << std::numeric_limits<cpp_dec_float_50 >::quiet_NaN() << "]"
+ << "[" << std::numeric_limits<bin_128bit_double_type >::quiet_NaN() << "]"
+ << "]" "\n"; // end of row.
+
+ fout << "[[signaling_NaN]"
+ << "[" << std::numeric_limits<float>::signaling_NaN() << "]"
+ << "[" << std::numeric_limits<double>::signaling_NaN() << "]"
+//#if LDBL_MANT_DIG > DBL_MANT_DIG
+ // Long double is not just a duplication of double.
+ << "[" << std::numeric_limits<long double>::signaling_NaN() << "]"
+//#endif
+ << "[" << std::numeric_limits<cpp_dec_float_50 >::signaling_NaN() << "]"
+ << "[" << std::numeric_limits<bin_128bit_double_type >::signaling_NaN() << "]"
+ << "]" "\n"; // end of row.
+
+ fout << "[[denorm_min]"
+ << "[" << std::numeric_limits<float>::denorm_min() << "]"
+ << "[" << std::numeric_limits<double>::denorm_min() << "]"
+//#if LDBL_MANT_DIG > DBL_MANT_DIG
+ // Long double is not just a duplication of double.
+ << "[" << std::numeric_limits<long double>::denorm_min() << "]"
+//#endif
+ << "[" << std::numeric_limits<cpp_dec_float_50 >::denorm_min() << "]"
+ << "[" << std::numeric_limits<bin_128bit_double_type >::denorm_min() << "]"
+ << "]" "\n"; // end of row.
+
+
+
+ fout << "]" "\n"; // end of table;
+
+
+ fout << "\n\n"
+ "[endsect] [/section:limits32 Numeric limits for 32-bit platform]" "\n" << std::endl;
+
+
+
+ fout.close();
+ }
+ catch(std::exception ex)
+ {
+ std::cout << "exception thrown: " << ex.what() << std::endl;
+ }
+
+
+} // int main()
+
+/*
+ Description: Autorun "J:\Cpp\Misc\Debug\numeric_limits_qbk.exe"
+
+ Program: I:\boost-sandbox\multiprecision.cpp_bin_float\libs\multiprecision\doc\numeric_limits_qbk.cpp
+ Wed Aug 28 14:17:21 2013
+ BuildInfo:
+ Platform Win32
+ Compiler Microsoft Visual C++ version 10.0
+ MSVC version 160040219.
+ STL Dinkumware standard library version 520
+ Boost version 1.55.0
+
+ */
+
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