Rein Klazes wrote:
Reading the comment again, I think you are mistaken: the "actual" bias is not tzi.Bias but either tzi.Bias + tzi.StandardBias or tzi.Bias + tzi.DaylightBias. If the flag is set, tzi.DaylightBias is zero, that is how the "actual" bias takes the disabled DST correction into account.
*nods* ... got confused by the different units and didn't see the connection.
Well, I could just duplicate the code but wouldn't it make more sense to move this code from kernel32 to ntdll and let GetTimeZoneInformation() call NtQuerySystemInformation()?
That is a possibility. But some code must be duplicated still for the support of TzSpecificLocalTime kernel32 API's, unless you know NT equivalents.
Hm... NT most likely stores it in UserSharedData (does anyone have the DDK handy?). How to deal with that?
-flx