0.16
- Work around perl bug in https://github.com/Perl/perl5/issues/19550
0.15
- Fix bug re inconsistent settings of POK and NOK flag, which manifests itself as (eg):
http://www.cpantesters.org/cpan/report/dc17e330-900b-11ec-bfc9-d1f1448276d4
It's a bug that rarely bites, and applies mainly (but not only) to perl-5.18.x and
earlier.
- In the Makefile.PL, do not re-insert stripped '-lm' or '-lquadmath' on
Windows as this won't ever achieve the desired result on that OS. (Right ??)
0.14
- Fix https://github.com/sisyphus/math-decimal64/pull/1 (also aplies to Math::Float128)
Thanks to @hiratara
- tweak the alignment conditions in math_float128_include.h to exclude
alignment directives for 32-bit mingw-w64 compilers
0.13
- Prototype subroutines listed in 'use subs()'.
- Move pod from Float128.pm to (newly created) Float128.pod.
- Add fromSTR(), fromNV(), fromIV() and fromUV().
- corrections to t/03arith.t and t/13constants.t.
( See https://rt.cpan.org/Ticket/Display.html?id=132376 )
0.12
- Fix typo ("Reinsterting") in Makefile.PL
- In Float128.xs fix alignment issue that arises on some -DDEBUGGING builds
- Attend to overloading of POK/NOK dualvars
0.11
- Replace SvUV/SvIV/SvNV with SvUVX/SvIVX/SvNVX where appropriate. (In typemap, too.)
- Add code to Makefile.PL to ensure that '-lquadmath' doesn't get stripped
(Thanks to kent frederic.)
0.10
- Fix overloading of comparison operators (for when operands are reversed).
0.09
- Change default ouptut precision to 36 - which brings it into line with the mpfr
library.
0.08
- In t/11stringification.t add a test to check that F128toSTR is, in fact, not
outputting insane values.
- Fix the problem that necessitated the addition of that F128toSTR test. For
defined(__MINGW32__) && !defined(__MINGW64__)
we now:
typedef __float128 float128 __attribute__ ((aligned(32)));
instead of simply:
typedef __float128 float128;
0.07
- Allow for MinGW64-v4 (runtime) bugfixes in Float128.xs.
0.06
- Have the Makefile.PL perform a test that works around the brokenness of
Bingos' NetBSD smoker(s).
- Add nnumflag(), set_nnum() and clear_nnum() functions. Also check for
non-numeric strings handed to strtoflt128()..
0.05
- Allow UV, IV, NV and PV overloading, and alter accordingly the way that new() assigns
these types.
0.04
- Add metadata to Makefile.PL.
- Add f128_bytes()
- Add extra quadmath.h functions
- Add typemap
0.03
- Fix t/03arith.t, t/05cmp.t, t/08log_exp.t, t/11stringification.t, and t/13constants.t
to accommodate the possibility that the locale in use represents the decimal point
with a comma.
0.02
- mingw64 compilers segfault on exp(). For these compilers we now have _overload_exp()
do e ** arg instead of doing exp(arg).
- Add constants defined in quadmath.h - and add t/13constants.t to check them.
- Define PERL_NO_GET_CONTEXT.
0.01
- First release