Revision history for Perl module Math::NV
2.04
- Add cmp_2 (new function).
- Remove all occurrences of "scalar reverse" from NV.pm, NV.pod,
t/08subnormal_mpfr.t and t/09subnormal_mpfr_alt.t
- Having removed all of those instances of "scalar reverse", I subsequently
discovered a need to (conditionally) reinstate a couple of them in NV.pm,
because of bugginess in pre-5.20.0 32-bit builds of Windows perl. Because
of that same bugginess, these perls are now also excluded from running
t/08subnormal_mpfr.t and t/09subnormal_mpfr_alt.t. (Some bugs call for
more fucking about than even I am prepared to tolerate.)
- Enable building on Cygwin -Dusequadmath perls. (Add '-lquadmath' link to LIBS.)
2.03
- "use constant" to replace $Math::NV::mpfr_strtofr_bug with MPFR_STRTOFR_BUG
and to replace $Math::NV::_ld_subnormal_bug with LD_SUBNORMAL_BUG
- Accommodate changes made to Math-MPFR private functions _d_bytes, _ld_bytes,
_dd_bytes and _f128_bytes that were made in Math-MPFR-4.13
2.02
- Remove "Perl_strtod" from NV.xs.
- Remove references to Perl_strtod from NV.pod.
- Add is_inexact (new function).
2.01
- Require version 4.07 of Math::MPFR as Math-MPFR-4.07 altered the way
that atonv() is called. The atonv function is used only in
t/08subnormal_mpfr.t and t/09subnormal_mpfr_alt.t
- Fix bug in test 7 of t/08subnormal_mpfr.t and t/09subnormal_mpfr_alt.t
Bug exposed by:
http://www.cpantesters.org/cpan/report/2d66ef6c-19fc-11e9-9376-4c7cbe00b72a
2.0
- Move POD to separate file (NV.pod)
- cater correctly for subnormal values in is_eq_mpfr() and nv_mpfr()
- add set_C() and set_mpfr
- make Math::MPFR-4.03 a prerequisite
1.02
- Update nv_mpfr() to accommodate IEEE 754 quad long double type.
- Small alteration to output of is_eq() and is_eq_mpfr(). (In the
event of discrepancy, they now identify which value belongs to
which method.)
1.01
- Correct tests in t/07nv_mpfr.t.
1.0
- Add nv_mpfr() and is_eq_mpfr() functions. (Also added $Math::NV::no_mpfr.)
- Add $Math::NV::no_warn; and check is_arg(), is_arg_mpfr(), nv_mpfr() and nv()
for non-string args.
0.08
- Change tack to concentrate on simply checking that the Math::NV functions
work correctly, and don't worry about trying to uncover discrepancies between
respective values assigned by perl and C. (This calls for a major change of
the test suite, and a small change to the documentation.)
- Remove Data::Float dependency.
- Remove mant2binary() and mant_str2binary(). (They were incorrect, and unnecessary.)
0.07
- Make corrections to POD documentation.
- Remove Data::Float::DoubleDouble from list of prereqs on doubledouble builds
(as that module already lists this module as a prereq).
0.06
- Accommodate nvtype of __float128
- Remove the 2nd arg (flag) from ld2binary() and ld_str2binary(). (Now takes
only one argument.)
- In t/07ld2binary.t, attempt to make allowances for locale settings of
decimal point.
- Add t/07ld2binary_doubledouble.t
0.05
- Add extra diagnostics to t/07ld2binary.t in an attempt to determine
the cause of the failures that occur on some machines.
0.04
- Minor rewrite of _ld2binary and _ld_str2binary to work with doubles
as well as long doubles
- Add bin2val function
- Add Cprintf and Csprintf functions
0.03
- Add t/10exponent_0_to_12.t.
- Define PERL_NO_GET_CONTEXT.
- Add metadata to Makefile.PL.
0.02
- Add 08compiler_bugs.t and 09list_context.t.
- Add is_eq(), mant2binary() and mant_str2binary() functions,
along with test scripts 06is_eq.t and 07mant2binary.t.
- Rewrite 04mpfr_str_checks.t and 05mpfr_nv_checks.t.
- nv() can now be called in either scalar or array context,
like POSIX::strtod().
0.01 2013-11-01
- First release to CPAN