boost/math
2
0
Fork 0
mirror of https://github.com/boostorg/math.git synced 2026-01-29 19:52:08 +00:00
Code Issues Projects Releases Wiki Activity
1,973 Commits 36 Branches 125 Tags
e1fc40cf110c80fcadac14f34a33c465e4e8efbe
Commit Graph

768 Commits

Author SHA1 Message Date
John Maddock
e1fc40cf11 Yet another try at getting constexp usage correct! 
Refs #6820.

[SVN r78160]
 2012-04-23 15:13:10 +00:00
John Maddock
73bd07a44d Fix constexp usage. 
[SVN r78133]
 2012-04-22 11:17:08 +00:00
John Maddock
ec742b4b80 Merge Owen's T function and Skew Normal code from sandbox and fix whatever failures show up. 
Regenerate docs.

[SVN r78130]
 2012-04-22 10:23:03 +00:00
John Maddock
ad520dbbcb Disable real_concept tests for IBM's compiler. 
[SVN r78108]
 2012-04-21 08:58:09 +00:00
John Maddock
f149d4cf53 Fix double-definition of float_t in tr1.hpp. 
Tweak behaviour of BOOST_MATH_BIG_CONSTANT.
Fix expected error rates for a bunch of tests (mostly on Mingw).
Fix paths in owens_t.qbk.

[SVN r78107]
 2012-04-21 08:43:06 +00:00
John Maddock
15baa10fac Add checks to discrete quantile finder to handle cases where there is no true answer. 
Fixes #6805.

[SVN r78090]
 2012-04-20 10:51:55 +00:00
John Maddock
85d55b2ea5 Update test_data.hpp to improve definition of SC_ so it's compatible with multiprecision testing. 
Update BOOST_MATH_BIG_CONSTANT to use a long double when the result is long double or smaller.

[SVN r78042]
 2012-04-17 16:05:52 +00:00
John Maddock
d94751ba33 Fix numeric error in bessel_jy.hpp. 
Remove some dead code from bessel_jy.hpp.
Add additional test cases for above bug.
Update and regenerate docs to match.

[SVN r77513]
 2012-03-24 11:59:21 +00:00
John Maddock
14b97d2dd4 Add missing include. 
[SVN r77350]
 2012-03-16 16:55:22 +00:00
John Maddock
940c3824c8 Change put_num_and_fill to work on CharType not char, so that formatting of signed zeroes is wide character safe even if '0' is not used for zeroes. 
[SVN r77286]
 2012-03-09 18:36:29 +00:00
John Maddock
b745e80ede Fix warning with a cast rather than a blanket disable. 
[SVN r77283]
 2012-03-09 09:35:36 +00:00
Paul A. Bristow
fec1f2407e Version that passes new tests for both default and signed_zero flag set based on modifications by KR Walker. 
BUT had to suppress MSVC warning "C4244 conversion wchar_t* to char* may cause data loss." 

[SVN r77268]
 2012-03-08 18:12:28 +00:00
John Maddock
6fc9eb72a9 Fix GCC some warnings, plus one failure. 
[SVN r76821]
 2012-02-01 13:06:17 +00:00
Paul A. Bristow
a94442f5de Added tests for new constants two_div_pi and root_two_div_pi 
[SVN r76820]
 2012-02-01 12:32:22 +00:00
John Maddock
20f26bd52c Use underscores in attribute names. 
Fixes #6411.

[SVN r76773]
 2012-01-29 12:27:53 +00:00
John Maddock
59eb0f44df Don't be too quick to disable long double functions if __NO_LONG_DOUBLE_MATH is set by GLIBC. 
Fixes #6459.

[SVN r76764]
 2012-01-28 19:13:44 +00:00
John Maddock
682d60c363 Merge new math constants code from the sandbox. 
Regenerate docs.
Fix build failures caused by merge.

[SVN r76736]
 2012-01-28 12:13:17 +00:00
John Maddock
887f590fcd Refactor the tests some more: allow tables of data to be of a different type to the type being tested - allows us to have tables of POD's (fast to compile) rather than tables of UDT's which compile very slowly on some compilers (usually Intel). 
[SVN r76544]
 2012-01-16 18:25:02 +00:00
John Maddock
c5bdd41539 Correct overflow logic in Bessel Y and add new test case. 
Fixes #6367.

[SVN r76373]
 2012-01-08 18:51:23 +00:00
John Maddock
fecae229b0 Simplify/improve zeta function still further. 
[SVN r76337]
 2012-01-07 16:57:51 +00:00
John Maddock
532ddf8cb0 Change template argument "L" to "Lanczos" to avoid conflicts with L as a macro. 
[SVN r76278]
 2012-01-02 17:44:19 +00:00
John Maddock
9005ade2a3 Updated big_lanczos.hpp to use new BIG_CONSTANT macros. 
Fix expression template issue in zeta.hpp.

[SVN r76251]
 2012-01-01 11:34:50 +00:00
John Maddock
9922e9134c Remove unused constant. 
Fixes #5982.

[SVN r76099]
 2011-12-21 17:58:28 +00:00
John Maddock
4d57f3e642 Disable SSE2 instruction support under the CUDA compiler. 
Fixes #6234.

[SVN r76097]
 2011-12-21 17:43:55 +00:00
John Maddock
a4a1cca095 Change overflow tests to not generate an infinity. 
Fixes #6307.

[SVN r76095]
 2011-12-21 16:58:36 +00:00
John Maddock
02e5447867 Delete unneeded variable. 
Fixes #6307.

[SVN r76094]
 2011-12-21 16:46:46 +00:00
John Maddock
1a304dccb0 Apply patch from #6192. 
Plus fix a few other issues that arise when BOOST_NO_LIMITS_COMPILE_TIME_CONSTANTS is set.
Fixes #6192.

[SVN r76093]
 2011-12-21 16:37:32 +00:00
John Maddock
af0c106290 Update zeta implementation to use much faster algorithm. 
Change zeta test to use external template instances.

[SVN r76092]
 2011-12-21 16:07:37 +00:00
John Maddock
70f902d332 Misc small workarounds to allow multiprecision tests to function. 
[SVN r76080]
 2011-12-20 16:21:44 +00:00
John Maddock
2902c06d6c Add missing using declarations. 
[SVN r76041]
 2011-12-18 16:05:55 +00:00
John Maddock
381bddafa0 Finish moving test cases into headers. 
Fix array declarations so GCC doesn't warn about them.
Declare constants in headers so they can be used by UDT's larger than type long double.
Suppress a few warnings and fix a couple of bugs that showed up when testing with UDT's.

[SVN r75960]
 2011-12-15 11:23:35 +00:00
John Maddock
434d53bc83 Fix complex number routines to work with signed zeros, changes involve: 
* Use boost::math::signbit rather than comparison to zero.
* Use boost::math::changesign rather than unary negation (unary negation fails for Intel on Linux when the argument is a zero).
* Update to use boost::math::isnan/isinf rather than the old code.
* Update to use boost::math constants.
Fixes #6171.

[SVN r75933]
 2011-12-13 18:11:46 +00:00
John Maddock
8d9372fac9 Begin refactoring math lib test cases as template so they can used by other numerical types. 
Change constant declarations in the Math lib headers to give full digits when used with UDT's that are constructible from strings.

[SVN r75820]
 2011-12-06 09:38:16 +00:00
Paul A. Bristow
cbe12e231d Added "two_thirds" and deprecate "twothirds" Trac 6199 
[SVN r75773]
 2011-12-02 15:01:40 +00:00
Paul A. Bristow
09c2238e7a Corrected complement cdf with Yan Zhou patch. 
[SVN r75591]
 2011-11-21 12:57:33 +00:00
John Maddock
c1e5af9291 Warning suppression. 
[SVN r75585]
 2011-11-21 09:46:33 +00:00
John Maddock
2d5ba01092 Tentative fix for platforms that have fenv.h but no fexcept_t. 
[SVN r75486]
 2011-11-14 12:20:13 +00:00
John Maddock
9373287faa More fixes for expression template types. 
[SVN r75403]
 2011-11-08 12:49:27 +00:00
John Maddock
a8d7299861 Allow distribution parameter checking to be eluded if it's already taken place. 
[SVN r75196]
 2011-10-31 16:38:16 +00:00
John Maddock
c43891eab0 Allow precision.hpp to work with number types in base 10. 
Fix binomial_distribution quantile when success_fraction is 1.
Fixes #6069.

[SVN r75168]
 2011-10-29 17:26:26 +00:00
John Maddock
34a635b621 Fix Clang build issue. 
[SVN r75152]
 2011-10-28 08:29:55 +00:00
John Maddock
681b9c1288 Add missing #include. 
Fixes #6017.

[SVN r74937]
 2011-10-13 17:31:00 +00:00
John Maddock
b7df32ac03 Apply patch from ticket #6014. 
Fixes #6014.

[SVN r74927]
 2011-10-12 17:09:48 +00:00
John Maddock
48e179266b Fix error checking for trunc and round functions. 
Update tests to match.

[SVN r74925]
 2011-10-12 11:06:09 +00:00
John Maddock
f617e8e284 Change nextafter and related functions to handle infinities as arguments the same way as GLIBC and MSVC. 
Fixes #5823.

[SVN r74918]
 2011-10-11 17:28:13 +00:00
John Maddock
a5941a2179 GCC workaround when using TR1 tuple. 
Fixes #5934.

[SVN r74891]
 2011-10-10 12:09:12 +00:00
John Maddock
303503d876 Remove unused constants. 
Fix expression for tolerance in eps_tolerance.
Add include guards to prime.hpp.
Fixes #5982.
Fixes #5927.
Fixes #6001.

[SVN r74890]
 2011-10-10 11:52:05 +00:00
John Maddock
f52587e616 More fixes for UDT's as arguments. 
[SVN r74802]
 2011-10-08 12:30:50 +00:00
John Maddock
8f22d5ebe5 Minor fixes for expression-template types. 
[SVN r74572]
 2011-09-26 11:01:03 +00:00
Paul A. Bristow
a74236ad41 std qualified strlen in int width = (int)std::strlen(prefix) + (int)std::strlen(body); 
to comply with C++ standard.

[SVN r74480]
 2011-09-20 18:03:22 +00:00