How can a MEX file give parameters to a C++ program?

Hello!

I am coding a program in MATLAB, in which I have to call a C++ library. To do so, I created a C++ file that calls the library and then a MEX-file that help me to call the desired function in MATLAB. The function I want to access in the library returns me a value, but I have to give it parameters. I am currently able to retrieve a value, because I write my parameteres directly in the C++ code as you can see here (my file's name is Test704()):

// Test704.cpp : Defines the entry point for the console application.
#define _AFXDLL
#include "StdAfx.h"
#include "704IO.h"
#include "Test704.h"
#include "mex.h"
#ifdef _DEBUG
  #define new DEBUG_NEW
#endif
/////////////////////////////////////////////////////////////////////////////

CWinApp theApp;  // The one and only application object

/////////////////////////////////////////////////////////////////////////////

using namespace std;

/////////////////////////////////////////////////////////////////////////////
//int _tmain(int argc, TCHAR *argv[], TCHAR *envp[])
int _tmain(double port[], double rack[], double offset[])
{
double  valueRead;
//short port;
//short rack;
//short offset;

    //      valueRead = PortRead(1, 780, -1);
          valueRead = PortRead(port[0], rack[0], offset[0]);
          mexPrintf("Value Read = %i\n",valueRead);
  return valueRead;
  }

/////////////////////////////////////////////////////////////////////////////
void mexFunction(int nlhs, mxArray *plhs[], int nrhs, const mxArray *prhs[])
{
    double *port, *rack, *offset;
    // Creates a 1-by-1 real integer. 
    //plhs[0] = mxCreateNumericMatrix(1, 1, mxINT32_CLASS, mxREAL);
    plhs[0] = mxCreateNumericMatrix(1, 1, mxINT32_CLASS, mxREAL);

      int* data = (int*) mxGetData(plhs[0]); 
      port = mxGetPr(prhs[0]);
      rack = mxGetPr(prhs[0]);
      offset = mxGetPr(prhs[0]);
      //valueRead = mxGetPr(plhs[0]);

      //data[0]=_tmain(0,0,0); 
      //return ;
      data[0] = _tmain(port,rack,offset);
    return ;
  }

You will notice that I have commented parts of the code in order to make researches. Indeed, I want now to be able to give parameters by this way (MATLAB code):

x = 1;
y = 780;
z = 1;
myVal = double(Test704(x,y,z));

and to still be able to retrieve a value in myVal. I helped myself with the example timestwo.c provided by MathWorks, but unfortunately, it just returns me a value of 0 instead of 2 or 4, and I cannot figure out why.

Thanks for help,

Mana

Answer 1

[James Tursa]

In this line you are using an integer output format %i to print a floating point double value:

mexPrintf("Value Read = %i\n",valueRead);

That will not work. You need to use one of the floating point formats for this. E.g.,

mexPrintf("Value Read = %g\n",valueRead); // or %f or %e

Once that is fixed you will get the proper value printed to the screen and see if it is correct at that point in the code. Also, I find it a bit strange that you are using a double for valueRead even though the function _tmain returns an int.

Then there are typos in your inputs. The following lines of code get the pointers from the 1st argument, not the 1st three arguments:

        port = mxGetPr(prhs[0]);
        rack = mxGetPr(prhs[0]);
        offset = mxGetPr(prhs[0]);

So change that to:

        port = mxGetPr(prhs[0]);
        rack = mxGetPr(prhs[1]);
        offset = mxGetPr(prhs[2]);

And if you want the output to be a double at your m-file level, why not just return a double variable from the mex routine? E.g.,

plhs[0] = mxCreateDoubleScalar( _tmain(port,rack,offset) );