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.


Attachment: signature.asc
Description: OpenPGP digital signature

Reply via email to