Bitset operation in c++

I am trying to pass bitset parameter to the below function but I am not able to:

error: no matching function for call to search(std::bitset<100000000ul>&, int, unsigned int, unsigned int)

Here is the code:

#include <bitset>


#define ELEMENTS 100000000
#define TRANSITION_POINT 500
using namespace std;

template<unsigned int N>
unsigned int search(bitset<N> &array, int value, unsigned int low, unsigned int high)  {

      unsigned int middle;
      .........
}

int main开发者_JAVA技巧()  {

    const unsigned int NUMBER_OF_ELEMENTS = ELEMENTS;
    bitset<NUMBER_OF_ELEMENTS> b;

    unsigned int i = 0;
    i = TRANSITION_POINT;
    while(i < NUMBER_OF_ELEMENTS)  {

        b[i] = 1;
        i++;
    }

    pos = search(b, (int)1, (unsigned int)0, (unsigned int)NUMBER_OF_ELEMENTS);
    return 1;
}

I tried to debug the code but couldnt find what is the problem. Could some one please help.

Your function template expects its first argument to be of type

std::bitset<(unsigned int)NUMBER_OF_ELEMENTS>

but b is of type

std::bitset<(std::size_t)NUMBER_OF_ELEMENTS> 

This subtle difference means that argument-based type inference will fail to match your template (unless std::size_t happens to be an alias of unsigned int).

There are two ways to fix it:

• Change the parameter type of your template to std::size_t to match that of std::bitset
• Explicitly use the function template:
  pos = search<NUMBER_OF_ELEMENTS>(b, 1, 0, NUMBER_OF_ELEMENTS);

The code works fine in my g++ compiler. The only error I get is 'pos' is not declared in this scope. (Also, I think its redundant to typecast to int and unsigned int while calling search(); that's not needed)

With 32 bit gcc 4.4.3 here are the changes that were needed to make it not segfault:

#define ELEMENTS 10000000

Apparently there's a max size, because 100000000 caused it to segfault on bitset<NUMBER_OF_ELEMENTS> b;

b.set(i,1);

The original b[i] = 1 also caused a segfault.