float128 implementation of fma missing

[charles@…]

Compiling the following code (g++ 4.8.3 with -fext-numeric-literals) gives

fmatest.cpp:9:16: error: ‘fma’ is not a member of ‘boost::math’

#include <iostream>
#include <boost/cstdfloat.hpp>
#include <boost/multiprecision/float128.hpp>
#include <boost/math/special_functions.hpp>

int main () {
  typedef boost::multiprecision::float128 real;
  real a = 10, b = 1/a, c = -1;
  std::cout << boost::math::fma(a, b, c) << " " << a * b + c << "\n";
}

[John Maddock]

Fixed in develop.