C / C++ / MFC

__int64 *buf = new __int64;
// this is the result I want, but it only shifts 32 bits??
*buf = (((((((data[6] | data[7] << 8) | data[8] << 16 ) | data[9] << 24 ) | data[10] << 32 ) | data[11] << 40 ) | data[12] << 48 ) | data[13] << 56 );
// these two produce the exact same result
__int64 foo = ((data[10] | data[11] << 8) | data[12] << 16 ) | data[13] << 24;
__int64 bar = (((data[10] << 32) | data[11] << 40) | data[12] << 48 ) | data[13] << 56;
// this corrects the error
*foo <<= 32;
// cast the value
double* dub = (double *)buf;
double val = *dub;