Mix-Language Program Problem: Array Access

[deadpickle]

I'm writing a C++ program that calls fortran subroutines. The C++ program also uses structs to access fortran common blocks. Here is the C++ code. Unfortunately, the Fortran code is to much to far to post but I think the problem originates in the C++ code.

#include <iostream>
#include <cstdlib>
#include <fstream>
#include <string>
#include <boost/tokenizer.hpp>
using namespace std;
using namespace boost;

//external Fortran function prototype
extern"C" {
	void getrec_(int *irw,const char *stname,int *sampin,int *mofm,int *nparm,\
					int *pc,double *val,const char *flag,const char *dflg1,\
					const char *dflg2,int *osmin,int *oemin,int *amin,int *iflag);
	void rsif_(int *iflag);
	void getfn_(int *iflag);

	// fortran common block declaration
	extern struct
	{
		bool prompt;
	} pr_;
	extern struct
	{
		char fnsif[80],fndat[80],fnerr[80],fnonly[80];
		char fnhis[80],fncrp[80],fnnp[80],action[72];
	} c6_;
	extern struct
	{
		float lat, lon, elev;
		int cr210t, pcatot, valwin, sifem, sifsm, trlqc, trlsa, trlrrd;
		short int statn, staten, outtab[6][5], parttab[4][50];
	} c2_;
}

int main(int argc, char *argv[])
{
	int arg;
	char *inFilename;
	fstream rawFile;
	string rawData;
	char_separator<char> sep(", -");
	int iflag;
	int irw = 1;
	
	// handle command-line arguments
	if (argc == 3)
	{
		for (arg = 1; arg < argc; ++arg)
		{
			char *argCp = argv[arg];
			
			if (argCp[0] != '-')
				cerr << "Illegal switch ignored: " << argCp << '\n';
			else
				switch (argCp[1])
				{
					case 'f':
						inFilename = argv[++arg];
						break;
					default:
						cerr << "Illegal switch ignored: " << argCp << '\n';
						break;
				}
		}
	}
	else
	{
		cerr << "Specify Command-line Arguments (-h for help)" << '\n';
		return(EXIT_FAILURE);
	}

	// Prompt user for station ID
	pr_.prompt = true;
	getfn_(&iflag);
	if (iflag != 0)
	{
		cerr << "Error During getfn" << '\n';
		return(EXIT_FAILURE);
	}
	
	rsif_(&iflag);
	if (iflag != 0)
	{
		cerr << "Error During rsif" << '\n';
		return(EXIT_FAILURE);
	}
	
	// read data from inFile
	cout << "Reading data from File: " << inFilename << '\n';
	rawFile.open(inFilename);
	
	// tokenize data 
	while (rawFile.good())
	{
		getline(rawFile, rawData);
		tokenizer< char_separator<char> > data(rawData, sep);	
	}
	
	return(EXIT_SUCCESS);
}

The problem I'm encountering is that one of the common blocks contains an array named parttab. This array is populated by the fortran subroutine rsif_. when I try to access the array elements in parttab such as parttab[0][0] it is not the correct value but if I access element parttab[0][-12] i get the expected first element. The parttab variable in c2 is declared in fortran as parttab[4][50] so I believe the array is declared correctly in C++. Whats the deal here? I appreciate any help. Thanks.

[Cubbi]

What's your compiler? That struct might be missing __attribute_common__ or __attribute__((__common__)). or what-have-you
You could post the definition of your c2 common area, or even add a dummy rsif to make it a compilable test that others could run.
You remember that fortran arrays are column-major while C arrays are row-major, right?

[deadpickle]

I'm not sure what you mean by _attribute_common_?
Yes, the arrays have been switched such that:
Fortran: parttab[50][4]
C++:parttab[4][50]
Compiler is gfortran and g++

[deadpickle]

The problem was that my variable order did not match the order within the common block. Now that they match the byte alignment is correct.