NEWS

---

cppdoubles 0.5.0

• cppdoubles has migrated its C++ dependency from cpp11 to cppally.

• Major speed improvements to most functions through the use of parallelisation and compiler optimisations.

• New functions get_threads and set_threads for getting and setting threads for parallel computation.

• double_equal now always returns TRUE when equality is exact, even when tolerance is negative.

cppdoubles 0.4.0 (2025-06-09)

New features

• get_tolerance and set_tolerance as helpers to get/set the package-wide tolerance.

• New argument scale added to rel_diff to allow users to optionally specify a scale for each relative difference.

Changes

• The default argument for tol now uses get_tolerance().

Internal updates

• Inlined most of the simple C++ functions which provides some moderate speed improvements.

• Some C++ functions were returning double when they should return bool. This has now been fixed.

cppdoubles 0.3.0 (2024-11-12)

Breaking changes

• The 'C++' functions have been renamed for readability and consistency though the equivalent R function names remain unchanged.

Internal improvements

• Vectorised loops have been internally sped-up.
• Code has been simplified for readability.

Bug fixes

• rel_diff(0, 0) now correctly returns 0

cppdoubles 0.2.0 (2024-02-21)

• New function all_equal as an efficient alternative to all.equal for numeric vectors.

• NA values are now returned when tol is NA.

cppdoubles 0.1.1 (2023-11-10)

• CRAN resubmission.

cppdoubles 0.1.0 (2023-11-08)

• Initial CRAN submission.