Added a few more large integer functions.
diff --git a/dlls/kernel/time.c b/dlls/kernel/time.c
index dcd0cb2..d7e39c6 100644
--- a/dlls/kernel/time.c
+++ b/dlls/kernel/time.c
@@ -214,26 +214,11 @@
* 1) Divided by CLK_TCK
* 2) Time is relative. There is no 'starting date', so there is
* no need in offset correction, like in UnixTimeToFileTime
- * FIXME: This function should be moved to a more appropriate .c file
- * FIXME: On floating point operations, it is assumed that
- * floating values are truncated on conversion to integer.
*/
static void TIME_ClockTimeToFileTime(clock_t unix_time, LPFILETIME filetime)
{
- double td = (unix_time*10000000.0)/CLK_TCK;
- /* Yes, double, because long int might overflow here. */
-#if SIZEOF_LONG_LONG >= 8
- unsigned long long t = td;
- filetime->dwLowDateTime = (UINT) t;
- filetime->dwHighDateTime = (UINT) (t >> 32);
-#else
- double divider = 1. * (1 << 16) * (1 << 16);
- filetime->dwHighDateTime = (UINT) (td / divider);
- filetime->dwLowDateTime = (UINT) (td - filetime->dwHighDateTime*divider);
- /* using floor() produces wierd results, better leave this as it is
- * ( with (UINT32) convertion )
- */
-#endif
+ LONGLONG secs = RtlEnlargedUnsignedMultiply( unix_time, 10000000 );
+ ((LARGE_INTEGER *)filetime)->QuadPart = RtlExtendedLargeIntegerDivide( secs, CLK_TCK, NULL );
}
/*********************************************************************
