Ad
  • Custom User Avatar
    08-03-1646 21:20
    Etc/GMT-2 (+2)
    Jamaica (-5)
    expected:<08-03-1646 14:[12]> but was:<08-03-1646 14:[20]>
    
    30-04-1875 13:56
    Asia/Novosibirsk (+7)
    Asia/Khandyga (+9)
    expected:<30-04-1875 1[7:2]6> but was:<30-04-1875 1[5:5]6>
    

    The random tests are using ancient time periods which requires arcane knowledge about all the documented time zone and clock changes, which is, obviously, library-dependent, hence implicitly forcing a specific library to be used in the solution.

    04-08-2703 18:21
    Europe/Simferopol (3.0)
    Europe/Warsaw (1.0)
    

    Also, testing times in far future is even more dangerous: we do not know the possible timezone changes in the future, hence even the currently used library will be suspect (unless it can keep updating the time adjustments till eternity).