Let's say, i want to this thing:
std::thread(std::sin,2.9);
But i get unresolved function type error. How can i resolve it?
std::sin is overloaded, not template (why??)
-
see stackoverflow.com/questions/1353757/…sehe– sehe2011-10-07 09:05:45 +00:00Commented Oct 7, 2011 at 9:05
Add a comment
|
4 Answers
Depending on what headers you included, there are the following:
template<class T> complex<T> sin (const complex<T>&); // <complex>
template<class T> valarray<T> sin (const valarray<T>&); // <valarray>
float sin(float); // <cmath>
long double sin(long double); // <cmath>
As you can see, there are templates involved, though that's actually not strictly relevant here.
You simply need to inform the compiler which [potential] overload you want (probably just out of the latter two):
std::thread((float(*)(float))&std::sin, 2.9);
answered Oct 7, 2011 at 9:07
Lightness Races in Orbit
387k77 gold badges670 silver badges1.1k bronze badges
4 Comments
Matthieu M.
If you wish to avoid C-cast, the appropriate C++ cast is
static_cast.Lightness Races in Orbit
@MatthieuM.: I don't particularly (it's completely superfluous here; what am I going to do, accidentally cast away constness?), but thanks.
Matthieu M.
I prefer C++ casts because it's more obvious a cast is going on, subjective, I know :)
Lightness Races in Orbit
@MatthieuM.: I prefer not abusing commas!
You could
typedef double (*sind)(double);
std::thread((sind) std::sin,2.9);
Comments
You have to cast the function pointer to the type of the right overload:
std::thread((float(*)(float))std::sin, 2.9);
Comments
I prefer the lambda solution. It reads much cleaner than anything involving a function pointer. It will also be faster than using a function pointer, but that doesn't usually matter.
std::thread( []( double x ) { return std::sin(x); }, 2.9 );
