Changes for version 0.06
- Doc fix for FakeDD.pod -see https://github.com/sisyphus/math-fakedd/issues/1 Thanks, Peter Acklam.
- Add overload_copy() which overloads the assignment operator ('=') Also add dd_copy() and dd_clone() - both of which return a copy of their Math::FakeDD object argument. Thanks again, Peter Acklam.
- Have dd_repro() check that Math::MPFR is at version 4.24 or later. (This is done in case Math::FakeDD is installed into a location where it becomes accessible to a perl whose version of Math::MPFR is pre-4.24.)
- Have mpfr2dd() return an (appropriately signed) Inf whenever it has calculated the DoubleDouble to be either [DBL_MAX, 2**970] or [-DBL_MAX, -(2**970)]. (DBL_MAX is 0x1.fffffffffffffp+1023, which is the largest possible finite double.) Add checks for this to t/assign.t and t/add.t. Thanks, Peter Acklam.
- Add overloading of '++' and '--'.
- Add dd_nextup() and dd_nextdown().
- Add is_subnormal() and ulp_exponent().
- Add any2dd()
Modules
DoubleDouble precision arithmetic for all architectures