Revision history for Perl module Math::Decimal128

    - Alter typedef of _Decimal128 for __MINGW64__ . (This became necessary with perl-5.28.0,
      but I've applied it universally whenever __MINGW64__ is defined.)
    - "use 5.008;" in Makefile.PL.
    - In XS code, replace (where practical) dynamic allocation of string buffer with
      static allocation.
    - Prototype subroutines listed in 'use subs()'.
    - Move POD out of and into (newly created) Decimal128.pod.

    - Adjust _is_neg_zero() to circumvent gcc optimization bug.

    - Remove occurrences of 'aTHX' that had been auto-implanted in some error messages
    - Replace SvIV/SvUV/SvNV with SvIVX/SvUVX/SvNVX where appropriate.
    - Amend _atodecimal() to correctly handle the case where the string begins with '-+'
    - In Decimal128, following the _Decimal128 typedef, replace all occurrences of
      "_Decimal128" with "D128".
    - Fix alignment issue on nvtype=double -DDEBUGGING perls

    - Fix overloading of comparison operators when operands are reversed.

    - Add $Math::Decimal128::NNW (set to 0). If this variable is set to 1,
      then non-numeric warnings will be given whenever a string containing
      any non-numeric characters is given to Math::Decimal128::_atodecimal().
    - Remove looks_like_number() from _atodecimal.
    - Rename nnumflagl, set_nnuml, clear_nnuml to nnumflag, set_nnum and
      clear_nnum and no longer export them.

    - Workarounds in _atodecimal() and assignPVl.t to cater for broken
      compiler/libc handling of signed zero _Decimal128 values.

    - Add PV overloading.
    - Add assignD128().
    - Rewrite assignPVl in XS. (Now uses _atodecimal XSub.)
    - Rewrite PVtoD128. (Now uses _atodecimal XSub.)

    - Fix bug in overload_sub by adding overload_neg.
    - No longer allow new() to assign an NV. (NVtoD128 must now be used instead.)
    - Add assignIVl(), assignUVl() and assignNVl().
    - Add D128toFSTR() and D128toRSTR()

    - Add DPDtoD128() and d128_bytes()

    - First release