Integral constant overflow

nitin_ucer · 09-10-2010, 02:56 PM

Following piece of code is giving "Integral constant overflow" Error :

#define MAXSIZE (512*1024*1024*1024)
UINT64 size;
If(size > MAXSIZE)
{
// do something
}

UINT64 can have values ranging from 0 to 18,446,744,073,709,551,615.

Then why i am getting this compiler error ? How to handle this ?

dugan · 09-10-2010, 02:59 PM

The macro expands to:

Code:

if (size > (512*1024*1024*1024))

Then when the right-hand side is evaluated, a temporary variable is automatically created to store the result, before it is compared with size. That variable is a 32-bit int.

nitin_ucer · 09-10-2010, 03:28 PM

I defined MAXSIZE as below and it worked

#define MAXSIZE 549755813888 // actual value of (512*1024*1024*1024)

thanks !!!!

Is there any other way to handle this ? putting a big number in Macro doesn't look good.

Sergei Steshenko · 09-10-2010, 05:37 PM

Quote:

Originally Posted by nitin_ucer

I defined MAXSIZE as below and it worked

#define MAXSIZE 549755813888 // actual value of (512*1024*1024*1024)

thanks !!!!

Is there any other way to handle this ? putting a big number in Macro doesn't look good.

Yes, there is - add (IIRC) 'ul' suffix to the numbers to be multiplied. Better yet check http://www.open-std.org/JTC1/SC22/wg...docs/n1124.pdf - C99 standard.

dugan · 09-10-2010, 05:46 PM

Quote:

Originally Posted by nitin_ucer

Is there any other way to handle this ? putting a big number in Macro doesn't look good.

I'm curious as to why you think so, considering that the big-number-constant version executes faster.

nitin_ucer · 09-12-2010, 11:15 AM

Quote:

Originally Posted by dugan

I'm curious as to why you think so, considering that the big-number-constant version executes faster.

I do agree with you.MAXSIZE is representing 512GB , so for readability I wanted to put it as 512*1024*1024*1024.