Version 1.4.2

• Fix bug where wrong method flag was returned when using combinations (-c flag) and D could not be calculated due to size.
• Replaced calculation of D when combinations is used (flag -c) with an exact method in all cases except for when both binary and combinations are used (-b and -c flag) and input D is larger than 15 (e.g. resulting D is (2^(15))!).

Version 1.4.1

• Documentation, text and smaller fixes forgotten in v. 1.4.
• Added tests where the project is used as a library (previous tests only used the project's command line API).

Version 1.4

• Added exact method for calculating N given D and P using a numerical approach, this means that from now on
  multiple solution strategies can be used for this calculation as well (earlier this calculation always used Taylor
  approximation).
• Fixed bug in method facultyLog for when input is 0. Since 0! = 1 and the return value is in log space, the
  correct answer is 0 and not 1.
• Added trivial use case for calculating N using D and P when D is 1 and P is neither 0 nor 1 (in this case
  the answer is always 2).

Version 1.3

• Improved exception handling:
  • Error codes
  • Exception class
  • Improved checks on input sizes before starting calculations
  • Returning error codes when terminating program with an error
  • Returning more specific error messages when a method fails
• Simplified conversion to log10 form in _BirthdayProblemNumberFormatter.toLog10ReprOrNone.
• Corrected precision bug in _BirthdayProblemSolverChecked.birthdayProblemInv.
• Corrected minor bug in _BirthdayProblemTextFormatter.methodToText.
• Simplified _BirthdayProblemTextFormatter.methodToText and _BirthdayProblemTextFormatter.methodToDescription.

Version 1.2

• Corrected calculation of adjusted precision during calculations.
• Added constants for repeatedly used values and replaced relevant uses.
• Corrected exact / naive calculation of m take n.
• Adjusted method using Stirling's formula so that it always returns probablity 0 or more (can otherwise return slightly less than 0 due to precision errors).
• Changed preprocessing so that the application fails whenever the max precision is insufficient to represent the resulting log of set size D.
• Small vocabular fixes in descriptions and comments.
• Fixed tests.

Version 1.1

• Added rounding upwards (ceiling) instead of regular rounding (half up) on non-logarithmic solutions for N.
• Removed output approximation character on non-logarithmic solutions for N.
• Added flag --prec for command-line interface allowing the user to choose output precision, where applicable, in [0, 10] with default 10.