__int64 and problems

[Radius]

I need to calculate the sizes in bytes of some very large files so I started playing around with unsigned __int64 datatypes.

As a test, I took 0x40000000 and said basically: num = _GB * 2;

(_GB being the hex number above)

and the result in num was: 0xffffffff80000000

Is there a compiler option I need to make this work properly, or am I not understanding something about these 64 bit types? I thought it should just be 0x80000000 without the leading 0xffffffff? Or am I totally out to lunch here?

Thanks!

[0xC0000005]

Since 0x40000000 and 2 are both valid integers the multiplication will be done as integers before assigning to int64 so you will have overflow.

The solution is to explicitly cast at least one of the integers to an int64 before the multiplcation:

__int64 num = (__int64)0x40000000 * 2;

[Radius]

Gotcha, thank you!

[Mark Xa]

The solution is to explicitly cast at least one of the integers to an int64 before the multiplcation:

__int64 num = (__int64)0x40000000 * 2;

No need to cast. Simply use the LL suffix on one of the integers:

    __int64 a;
    a = 12345678LL * 12345678;

(LL means "long long").

In any case it is recommended (by Bjarne himself ) that old C-style casts are not used since their intended meaning is less well understood by the compiler: http://www.research.att.com/~bs/bs_f...ml#static-cast