C Macros: Function factories, why the macro don't works only in one case?

i've written this two macros:

// Magic Assert Equal Atomic constructor generator
#define _GENERIC_ASSERT_EQ_ATOMIC_CONSTRUCTOR_(n, N, W, tt) \
assert_data_t *assert_eq_##n##_constructor (tt a, tt b, int passed) {          \
    return assert_data_constructor (_ASSERT_EQ_##N##_, passed, W(a), W(b));    \
}

// Magic Assert Equal Vector constructor generator
#define _GENERIC_ASSERT_EQ_VECTOR_CONSTRUCTOR_(n, N, W, tt) \
assert_data_t *assert_eq开发者_如何学运维_##n##_vector_constructor                              \
  (tt * a, tt * b, int n, int passed) {                                        \
    return assert_data_constructor                                             \
      (_ASSERT_EQ_##N##_VECTOR_, passed, W##Vector(a, n), W##Vector(b, n));    \
}

the first macro works well (in the cases i've tried):

_GENERIC_ASSERT_EQ_ATOMIC_CONSTRUCTOR_(int, INT, Int, int)
_GENERIC_ASSERT_EQ_ATOMIC_CONSTRUCTOR_(flt, FLT, Flt, float)
_GENERIC_ASSERT_EQ_ATOMIC_CONSTRUCTOR_(dbl, DBL, Dbl, double)
_GENERIC_ASSERT_EQ_ATOMIC_CONSTRUCTOR_(complex_flt, COMPLEX_FLT, ComplexFlt, complex float)
_GENERIC_ASSERT_EQ_ATOMIC_CONSTRUCTOR_(complex_dbl, COMPLEX_DBL, ComplexDbl, complex double)
_GENERIC_ASSERT_EQ_ATOMIC_CONSTRUCTOR_(str, STR, Str, char *)

but the second macro... doesn't work well in the 'int' case (with float and double yes):

_GENERIC_ASSERT_EQ_VECTOR_CONSTRUCTOR_(int, INT, Int, int) // Here i have an error
_GENERIC_ASSERT_EQ_VECTOR_CONSTRUCTOR_(flt, FLT, Flt, float)
_GENERIC_ASSERT_EQ_VECTOR_CONSTRUCTOR_(dbl, DBL, Dbl, double)

the error that gcc shows me is:

unitarium4c.c:115: error: two or more data types in declaration specifiers
unitarium4c.c: In function ‘assert_eq_int_vector_constructor’:
unitarium4c.c:115: error: parameter name omitted
unitarium4c.c:115: error: expected expression before ‘int’
unitarium4c.c:115: error: expected expression before ‘int’

:p i don't understand why appears that error. (If I copy the macro and expand it, it works well in the 'int' case).

Thanks in advance :) .

Try

// _GENERIC_ASSERT_EQ_VECTOR_CONSTRUCTOR_(int, INT, Int, int) // Here i have an error
   _GENERIC_ASSERT_EQ_VECTOR_CONSTRUCTOR_(integer, INT, Int, int)
//                                   _____^^^^^^^_____

The expansion of the original macro would create

assert_data_t *assert_eq_int_vector_constructor                              \
  (int * a, int * b, int int, int passed) {                                        \
/* __________________^^^^^^^__ */
    return assert_data_constructor                                             \
      (_ASSERT_EQ_INT_VECTOR_, passed, IntVector(a, n), IntVector(b, n));    \
}

You can also tell gcc to stop "compiling" after expanding the macros and check the resulting code. Try

gcc -E source

You have a macro parameter n, and a function parameter int n in the macro body, so that's becoming int int.