如何在编译时检查类型

Question

I'm giving option to compile the program with either float or double type, but there is a problem: I need to manually set either GL_FLOAT or GL_DOUBLE, because I dont know how can I do the following:

typedef float MYTYPE;

#if MYTYPE == float
    #define GL_MYTYPE GL_FLOAT // used for vertex array parameters.
#else
    #define GL_MYTYPE GL_DOUBLE
#endif

Note: I dont have C++11 or whatsoever, just the good old C++.

Answer 1

In C++11, you can use std::conditional and std::is_same as:

#define GL_MYTYPE  std::conditional                       \
                   <  std::is_same<MYTYPE,float>::value,  \
                      GL_FLOAT,                           \
                      GL_DOUBLE                           \
                   >::type                            

In C++03, you can implement these functionalities yourself as:

template<bool B, class T, class F>
struct conditional { typedef T type; };

template<class T, class F>
struct conditional<false, T, F> { typedef F type; };

and

template<class T, class U>
struct is_same { static const bool value = false; };

template<class T>
struct is_same<T, T> { static const bool value = true; };

Note that the implementation of conditional is taken from the site itself.

Answer 2

#define s are processed by the preprocessor, before the compiler runs. So it doesn't see the typedef .

Answer 3

Use preprocessor macros to change what your typedef line does

#if defined(TYPE_IS_FLOAT)
typedef float MYTYPE;
#else
typedef double MYTYPE;
#end

And then use your build system to set TYPE_IS_FLOAT as appropriate (and give it a better name of course).

Or wrap all the code which behaves like this up in a template class and use a template parameter for this type, then use whichever version is appropriate at the time.

Answer 4

You can do it this way:

#define MYFLOAT

#ifdef MYFLOAT
    typedef float MYTYPE;
     #define GL_MYTYPE GL_FLOAT // used for vertex array parameters.
#else
    typedef double MYTYPE;
    #define GL_MYTYPE GL_DOUBLE
#endif

You would need to define MYFLOAT if you want to use float , or omit it to use double .

PS: keep in ind this is not evaluated at compile time, but at pre-processing time.

Answer 5

You may use #ifdef as follow :

#ifdef DOUBLE
   #define GL_MYTYPE GL_FLOAT
   //or
   typdef ...
#endif

#ifdef FLOAT
   #define GL_MYTYPE GL_DOUBLE
   //or
   typdef ...
#endif

You have to compile your code with DOUBLE or FLOAT as the symblos defined.

Answer 6

I think Nawaz's answer is nearly correct. In the special case of choosing between GL_FLOAT and GL_DOUBLE , std::conditional is not needed. Just use a ternary if :

#define GL_MYTYPE std::is_same<MYTYPE,float>::value?GL_FLOAT:GL_DOUBLE

This is because GL_FLOAT and GL_DOUBLE are not types, but rather literal GLenum values. So GL_MYTYPE should also not be a typename but a GLenum . So I'd rather use:

const GLenum GL_MYTYPE = std::is_same<MYTYPE,float>::value?GL_FLOAT:GL_DOUBLE;