On 8/1/2011 5:16 PM, Russ Allbery wrote: > Jeffrey Altman <[email protected]> writes: > >> If you can provide me a list of boundary times to check I can construct >> a test application that uses SYSTEMTIME to FILETIME conversions to test >> the results of various time arithmetic operations. > >> typedef struct _SYSTEMTIME { >> WORD wYear; >> WORD wMonth; >> WORD wDayOfWeek; >> WORD wDay; >> WORD wHour; >> WORD wMinute; >> WORD wSecond; >> WORD wMilliseconds; >> } SYSTEMTIME, *PSYSTEMTIME; > > Usually the interesting ones for Gregorian conversion of software written > in the US are around the date of UK transition from Julian to Gregorian in > September of 1752: > > September 1752 > Su Mo Tu We Th Fr Sa > 1 2 14 15 16 > 17 18 19 20 21 22 23 > 24 25 26 27 28 29 30 > > (gcal is one of the few programs I know that attempts to cope with this.) > > My guess is that Windows is using a backwards-projection of UTC without > leap seconds and not attempting to worry about dates that theoretically > should be in Julian, so you'll get an offset of 10 or 11 days relative to > the actual historical time when converting from a timestamp to calendar > time. >
Provide times for a few known leap seconds and I can test those as well.
signature.asc
Description: OpenPGP digital signature
