Changes for version 2.006006 - 2019-10-01
- change quotify to use longest form of floating point numbers if unable to find a perfectly accurate representation
- updated documentation for quotify to reflect handling of floating point numbers
- don't try to copy hints hash entries that look like references (RT#122698)
Changes for version 2.006_005 - 2019-09-06
- additional fixes for quotify for floating point numbers to be faster and should always be able to maintain accuracy
Changes for version 2.006_004 - 2019-04-24
- change quotify of NaNs to simpler calculation
- avoid ever triggering exceptions when testing for quotify on false values
- skip quotify Inf and NaN tests if perl is compiled without support for them
- fix quotify of very large or very small numbers
- fix accuracy of quotify on high precision numbers when perl is compiled with quadmath
- quotify will use hex floats when needed for full accuracy if they are available
- change tests of quotify for floating point numbers to allow a small amount of inaccuracy, since decimal floats can't always be accurate
Modules
Defer generation of subroutines until they are first called
Efficient generation of subroutines via string eval