fix warning std::abs

[kerautret]

Small PR just to remove warning appeared from last OS X update

[rolanddenis]

Actually, it is more than a simple warning: if both abs and std::abs are in scope, and without explicit use of std::abs (or of the directive using std::abs;), the compiler will choose abs that works on integers.

For example, the following code:

#include <iostream>
#include <cstdlib>
#include <cmath>

int main()
{
    double v = 2.5;
    std::cout << "abs(2.5) = " << abs(v) << std::endl;
    std::cout << "std::abs(2.5) = " << std::abs(v) << std::endl;
    return 0;
}

compile silently with g++ (v4.9.1 on Linux), even with -Wall -Wextra, and with a warning with clang++ (v3.5.0):

test.cpp:8:35: warning: using integer absolute value function 'abs' when argument is of floating point type [-Wabsolute-value]
    std::cout << "abs(2.5) = " << abs(v) << std::endl;
                                  ^
test.cpp:8:35: note: use function 'std::abs' instead
    std::cout << "abs(2.5) = " << abs(v) << std::endl;
                                  ^~~
                                  std::abs
1 warning generated.

but both binaries output is:

abs(2.5) = 2
std::abs(2.5) = 2.5

I've made a fast search through the DGtal source and it seems to be the only file where abs (alone) is used on floats.

[kerautret]

yes, I also detects strange behavior with abs in my dev prog ;)

[dcoeurjo]

Thanks @kerautret for the PR and @rolanddenis for the explanation :)

[dcoeurjo]

@kerautret could you please merge the master to this branch?

[kerautret]

ok

[kerautret]

@dcoeurjo done

[dcoeurjo]

merging.