Template function “Subscript requires array type”, but works in smaller projects. Why?

Question

The following code is part of an interpolation function I wrote as part of a larger project. The first version of this function returned the myScalar yval, but I modified it to return a flag on whether or not the function worked.

My question is this. The following code compiles when run by itself both on codepad.org and in a smaller Visual Studio project. In my larger project, though, I am getting error C2109 "subscript requires array or pointer type." What could be going wrong?

Thanks in advance! -- Joe

using namespace std;

template <class myScalar, class myXVec, class myYVec>
int finterp(int mode, myXVec xarray, myYVec yarray, int num_pts, myScalar xval, myScalar &yval)
{
   myScalar dx, dydx, xmin, xmax;
   int success_flag = 0;

   if (num_pts < 1) return success_flag;
   yval = yarray[0]; //Visual Studio error C2109

   //some more calculations are here

   success_flag = 1;
   return success_flag;
}

int main()
{
   double *xvec, *yvec;
   xvec = new double [5]; yvec = new double [5];
   for (int i = 0; i < 5; i++)
   {
      xvec[i] = (double)i;
      yvec[i] = (double)i+1;
   }
   double x, y;
   x = 3.0;
   int success = finterp(1, xvec, yvec, 5, x, y);
   cout << y << "  " << success << endl;
   return 0;
}

Output:

1> j:\london_study\irasshell_2011-05-13\iras_src\templateutilityfunctions.h(74): 
   error C2109: subscript requires array or pointer type 
1> j:\london_study\irasshell_2011-05-13\iras_src\hydpowclass.cpp(41) : 
   see reference to function template instantiation 'int finterp<double,std::vector<_Ty>,double>(int,myXVec,myYVec,int,myScalar,myScalar &)' being compiled 
1> with 
1> [ 
1> _Ty=double, 
1> myXVec=std::vector<double>,
1> myYVec=double, 
1> myScalar=double 
1> ] 

Answer 1

As per your comment, in your real code a mere double is being passed in for yarray rather than a double* or std::vector<double>. This is a simple case of having a sufficiently small, but incorrect, repro -- the real error lies in your real code.

Answer 2

In the code you've posted, you're calling finterp with myYVec = double*. This can be indexed just fine with [0].

When you use this in a larger project, how are you calling finterp? Visual Studio should tell you in the error messages after the c2109.

Whatever type you're passing in as the third parameter is apparently non-indexable.

EDIT Ah, you've updated your question with the error message. The error occurs when you call finterp with myYVec = double -- which is NOT indexable. I think you meant to use a double*.