Matt Borland
a0360d8f6d
Use charconv in convert_from_string for arithmetic types
2023-05-16 15:40:01 +02:00
Matt Borland
0c1920b7fb
Move macro definitions to config
2023-05-16 15:39:42 +02:00
Matt Borland
5045047bcd
Use charconv instead of streaming operator for std::float128_t
2023-05-16 15:02:31 +02:00
jzmaddock
843363650f
Fix remaining warnings from float32.cpp.
2023-05-16 13:14:08 +01:00
Matt Borland
d217813b03
Fix stack overflow on test_finite_singular_boundary with _Float64
2023-05-16 11:49:30 +02:00
jzmaddock
c5b4d28e05
Suppress lots of warnings for std::float32_t.
...
Better configure promote_args test case.
2023-05-15 17:08:41 +01:00
jzmaddock
b3f361c123
Fix the promote_args<float32_t, float32_t> case.
...
Also fixes a lot of warnings.
[CI SKIP]
2023-05-14 12:23:06 +01:00
jzmaddock
ac4605b12d
Fix Minw-64 C++23 compile failures.
2023-05-13 19:13:11 +01:00
Matt Borland
5d044d52f5
Remove cast for output iterator
2023-05-09 13:43:55 +02:00
Matt Borland
be37704674
Explicit casting of promoted types
2023-05-02 13:50:31 +02:00
Matt Borland
2b1d5b613a
Cast assignment of promoted type
2023-05-02 13:17:56 +02:00
Matt Borland
4c753c20f7
Reverse 128bits and prefer F64 over long double when it's 64bits
2023-04-25 12:06:16 +02:00
Matt Borland
5dcf5a9fdb
Test beta dist for F64, F32, and F16
2023-04-24 14:53:05 +02:00
Matt Borland
7a6efabba5
First draft of promotion
2023-04-24 14:25:13 +02:00
Matt Borland
298a243ccd
Fix for scipy issue 18302 ( #977 )
2023-04-21 12:31:23 +02:00
Madiyar
109a814e89
Fix fisher_f mode ( #976 )
...
* Fix fisher_f mode
The mode for F-distribution is defined when `df1 > 2` according to https://en.wikipedia.org/wiki/F-distribution . It is also reasonable since `df2 * (df1 - 2) / (df1 * (df2 + 2))` becomes zero or negative when `df1 <= 2`
Co-authored-by: Matt Borland <matt@mattborland.com >
2023-04-14 12:30:53 +02:00
Matt Borland
93448acea7
Merge pull request #946 from mborland/logcdf
...
Add logcdf to distributions
2023-04-13 14:10:35 +02:00
Matt Borland
3a81f17b14
Replace language version check with feature check
2023-04-05 14:15:59 +02:00
Matt Borland
2c141fe268
Disable __float128
2023-04-03 17:43:50 +02:00
Matt Borland
03f4498b16
Change handling for non-literal types
2023-04-03 17:04:06 +02:00
Matt Borland
93b31d1e19
Use boost::math fpclassify functions
2023-04-03 15:40:28 +02:00
Matt Borland
c597cb1112
Fix ccmath fpclassify for poison
2023-04-03 15:03:11 +02:00
Matt Borland
e5b4f3f0c3
Improve trunc handling
2023-04-03 14:50:22 +02:00
Matt Borland
1de39c69ef
Add coverage for lround and iround
2023-04-03 14:06:29 +02:00
Matt Borland
3200ce108d
Change comparison to max
2023-03-21 08:11:00 -07:00
Matt Borland
6989ad19ff
Don't use constexpr calculation with concept tests
2023-03-20 17:50:04 -07:00
Matt Borland
b1a95fa2c1
Add BOOST_PREVENT_MACRO_SUBSTITUTION
2023-03-20 17:03:47 -07:00
Matt Borland
5a50ec7868
Simplify and constexperize
2023-03-20 15:22:44 -07:00
Matt Borland
5284fb7fa8
Explicitly cast the result
2023-03-16 21:49:43 -07:00
Matt Borland
b8fb413405
Compensate for signedness
2023-03-16 21:23:52 -07:00
Matt Borland
f349ddbb5a
Fix for long doubles
2023-03-16 20:11:05 -07:00
Matt Borland
73b8ffc13c
Fix llround for non-representable numbers
2023-03-16 19:26:23 -07:00
Matt Borland
1f94d9806b
Fix -Wmaybe-uninitialized in bessel jy series
2023-03-10 12:31:51 -08:00
jzmaddock
cf6cd75477
Improve error handling in non-central distributions. ( #963 )
...
* Improve error handling in non-central distributions.
* Try turning debug symbols off for msvc-14.0.
So we don't run out of disk space on CI.
2023-03-07 12:11:29 +00:00
jzmaddock
bf3bc2e6c2
Fix ibeta_inv for very small p. ( #962 )
...
* Fix ibeta_inv for very small p.
Change assert's in temme_method_1_ibeta_inverse to corrections when guess goes out of range.
Change handling of non-convergence in second_order_root_finder to use bracketing when the end points are many orders of magnitude apart.
Fixes: https://github.com/boostorg/math/issues/961 .
* Add missing copyright.
[CI SKIP]
2023-03-05 13:18:27 +00:00
Matt Borland
8b52854cb5
Use control points to determine amount of scratch space needed
2023-03-02 09:47:23 -08:00
Matt Borland
91e98b89e3
Replace int with long long in nct pdf and cdf
2023-02-21 08:16:31 -08:00
Matt Borland
3886633231
Add complement logcdf for laplace distribution
2023-02-14 09:39:28 -08:00
Matt Borland
fcf05006cb
Add complement logcdf for extreme value distribution
2023-02-14 08:43:47 -08:00
jzmaddock
79bf64e23f
Merge pull request #949 from mborland/15101
...
Fix for scipy issue 15101
2023-02-13 15:36:56 +00:00
jzmaddock
879e26013a
Disable legendre_stieltjes.hpp inclusion when exceptions are off.
2023-02-12 18:37:43 +00:00
jzmaddock
d7b68a7fb5
Fix gigenbauer for non-eh environments.
2023-02-12 15:27:56 +00:00
jzmaddock
bf4c7fdd2c
Fix hypergeometric for use with -fno-exceptions.
...
Add test case.
Refs: https://github.com/boostorg/math/pull/947
2023-02-12 12:18:42 +00:00
Matt Borland
7b678a9001
Add logcdf specialization to rayleigh distribution
2023-02-11 18:07:27 -08:00
Matt Borland
614445505a
Add logcdf specialization to pareto distribution
2023-02-11 17:49:29 -08:00
Matt Borland
e387f9c429
Add logcdf specialization to geometric distribution
2023-02-11 17:28:27 -08:00
Matt Borland
a77f0fa196
Add logcdf specialization to exponential distribution
2023-02-11 16:42:20 -08:00
Matt Borland
1e47518075
Fix for scipy issue 15101
2023-02-11 09:16:32 -08:00
Matt Borland
6159c05a2b
Improve laplace logcdf
2023-02-07 10:13:22 -08:00
Matt Borland
c0f89544c3
Add logistic distribution logcdf for complement two type
2023-02-06 18:14:49 -08:00