boost/math
2
0
Fork 0
mirror of https://github.com/boostorg/math.git synced 2026-01-30 08:02:11 +00:00
Code Issues Projects Releases Wiki Activity

Improve fix, no pch, test spec fun locally

Browse Source
This commit is contained in:
Christopher Kormanyos
2020-12-14 07:21:52 +01:00
parent 90ee0013e7
commit 1e61cea5df
3 changed files with 95 additions and 85 deletions
Download Patch File Download Diff File Expand all files Collapse all files

19
include/boost/math/special_functions/gamma.hpp
View File

@@ -366,9 +366,9 @@ std::size_t highest_bernoulli_index()
template<class T>
int minimum_argument_for_bernoulli_recursion()
{
const float digits10_of_type = (std::numeric_limits<T>::is_specialized
? static_cast<float>(std::numeric_limits<T>::digits10)
: static_cast<float>(boost::math::tools::digits<T>() * 0.301F));
BOOST_CONSTEXPR_OR_CONST float digits10_of_type =
(std::numeric_limits<T>::is_specialized ? (float) std::numeric_limits<T>::digits10
: (float) boost::math::tools::digits<T>() * 0.301F);
float min_arg_as_float = (float) digits10_of_type * 1.7F;
@@ -377,14 +377,17 @@ int minimum_argument_for_bernoulli_recursion()
// The following code sequence has been modified
// within the context of issue 396.
// The calculation of the test-variable limit has now been
// made more clearly safe regarding overflow/underflow dangers.
// The calculation of the test-variable limit has now
// been protected against overflow/underflow dangers.
// The previous line looked like this and did, in fact,
// underflow ldexp when using certain multiprecision types.
// The previous line looked like this (and did underflow ldexp for multiprecision types):
// const float limit = std::ceil(std::pow(1.0f / std::ldexp(1.0f, 1-boost::math::tools::digits<T>()), 1.0f / 20.0f));
// The new safe versoin of the limit check is now here.
const float limit = std::exp(((digits10_of_type / 0.301F) * std::log(2.0F)) / 20.0F);
// The new safe version of the limit check is now here.
const float digits2_of_type = (digits10_of_type + 1.0F) / 0.301F;
const float limit = std::exp((digits2_of_type * std::log(2.0F)) / 20.0F);
min_arg_as_float = (std::min)(min_arg_as_float, limit);
}