Modulo with long long integers in Objective C

Question

I'm trying use the modulo operator (%) on long longs, and it seems to return 0 if the number is above the range of an unsigned int. Is there an operator or function that I should be using instead, or should I roll my own?

-- Update: sorry for the lack of example code before, I was in a hurry, heading out the door --

Here's the code:

long long val = 12345678991245; // always less than LLONG_MAX
int power = 0;
NSMutableArray *powers = [[NSMutableArray alloc] init];
while (pow(10, power) <= val) {
    tVal = (val % (lround(pow(10, power + 1)))) / lround(pow(10, power));
    [powers addObject:[NSNumber numberWithInt:tVal]];
    NSLog(@"Power: %d tVal: %d", power, tVal);
    val -= val % lround(pow(10, power + 1));
    power++;
}

When you look at the output made by that NSLog statement, though, you see this when the number exceeds UINT_MAX (for the number 1,111,111,111,111):

2009-03-18 20:42:16.471 Numbers[11197:20b] Power: 0 tVal: 1
2009-03-18 20:42:16.472 Numbers[11197:20b] Power: 1 tVal: 1
2009-03-18 20:42:16.473 Numbers[11197:20b] Power: 2 tVal: 1
2009-03-18 20:42:16.476 Numbers[11197:20b] Power: 3 tVal: 1
2009-03-18 20:42:16.476 Numbers[11197:20b] Power: 4 tVal: 1
2009-03-18 20:42:16.477 Numbers[11197:20b] Power: 5 tVal: 1
2009-03-18 20:42:16.477 Numbers[11197:20b] Power: 6 tVal: 1
2009-03-18 20:42:16.477 Numbers[11197:20b] Power: 7 tVal: 1
2009-03-18 20:42:16.477 Numbers[11197:20b] Power: 8 tVal: 1
2009-03-18 20:42:16.478 Numbers[11197:20b] Power: 9 tVal: 0
2009-03-18 20:42:16.478 Numbers[11197:20b] Power: 10 tVal: 0
2009-03-18 20:42:16.479 Numbers[11197:20b] Power: 11 tVal: 0
2009-03-18 20:42:16.479 Numbers[11197:20b] Power: 12 tVal: 0

Everything is kosher until you get up to past UINT_MAX, then it just returns 0.

I'm compiling against the iPhone SDK using Xcode 3.1.2.

Any ideas?

Answer 1

lround returns a long, which has 32 bits. Try llround instead.

Better yet, don't use floating point to do integer math. Since you're increasing the exponent by one each time through the loop, you don't have to compute the power each time. Just store it in a variable, and multiply it by 10 on each iteration.

Also, instead of dividing the number by larger and larger powers of ten, try dividing the number by ten each time.

When you're looking for a bug, break the code into smaller steps, and check the results from the intermediate operations.