Thread: Wierd problem

The problem I'm having with my code is located at

Code:

void fGR(double x[NSTEPS], double y[NSTEPS], double sig[NSTEPS],
         double **GRalpha, double **GRbeta, int ma, double chi,
         int nu, double Q, double GRa[GRma])

 * * *

  fout1 << "GRma:" << ma << " GR[]:" << GRa[0] << " " << GRa[1] << endl;

  fout1 << "FF L50:  GRalpha[][0]\tGRalpha[][1]\tGRbeta[][0]" << endl;
  for(i=1; i<=ma; i++)
    {
        fout1 << endl << "54GR:" << GRa[0] << " " << GRa[1] << endl;
      fout1 << "[" << i << "][] \t ";
      for(j=1; j<=ma; j++)                                        
        {
          GRalpha[i][j] = 0.0;
          fout1 << GRalpha[i][j] << "\t\t";
        }
        fout1 << endl << "61GR:" << GRa[0] << " " << GRa[1] << endl;
      GRbeta[i][0] = 0.0;
      fout1 << GRbeta[i][0] << endl;
    }

As you can see, there doesn't seem to be any trouble; I'm mostly just outputting some data to a file. The problem I'm having is (output; not code):

Code:

GRma:2 GR[]:430.819 6.0904
FF L50:  GRalpha[][0]   GRalpha[][1]    GRbeta[][0]

54GR:430.819 6.0904
[1][]    0              0               
61GR:430.819 6.0904
0

54GR:430.819 6.0904
[2][]    0              0               
61GR:0 6.0904
0

I input into the function GRa[0] = 430.819, GRa[1] = 6.0904. After the second iteration of the outer for loop, it suddenly changes GRa[0] to 0 for no apparent reason. Why would a program suddenly do this?

If you just look at the output "fout <<", yeah, it would be C++, but that wasn't the point since I'm mostly writing C with C++ convenience.

I can see that the could be confusing, but this isn't the entire code. Basically it comes down to my final question: why would a for-loop suddenly change the value through one iteration when I'm not even adjusting that term in the function?

Here is a cleaner version of that portion (lines labeled where I send output):

Code:

void fGR(double x[NSTEPS], double y[NSTEPS], double sig[NSTEPS],
         double **GRalpha, double **GRbeta, int ma, double chi,
         int nu, double Q, double GRa[GRma])
{
 * * *

(50)

  for(i=1; i<=ma; i++)
    {
(54)
      fout1 << "[" << i << "][] \t ";
      for(j=1; j<=ma; j++)                                        
        {
          GRalpha[i][j] = 0.0;
          fout1 << GRalpha[i][j] << "\t\t";
        }
(61)
      GRbeta[i][0] = 0.0;
      fout1 << GRbeta[i][0] << endl;
    }

 * * *
}

Remove the file output stream and use printf instead, then check to see if does the same thing...

Yeah, it does (output, not code):

Code:

GRma:2 GR[]:430.819400 6.090400

54GR:430.819400 6.090400

61GR:430.819400 6.090400

54GR:430.819400 6.090400

61GR:0.000000 6.090400

I only replaced the lines for that particular problem (separately replacing all of those in the loop). Additionally, I turned it into a dynamically allocated pointer (changed to double *GRa). It works up to this point as well.

Here is if I increase the number:

Code:

GRma:2 GR[]:4308.194000 6.090400

54GR:4308.194000 6.090400

61GR:4308.194000 6.090400

54GR:4308.194000 6.090400

61GR:0.000000 6.090400

Here is if I decrease the value:

Code:

GRma:2 GR[]:4.308194 6.090400

54GR:4.308194 6.090400

61GR:4.308194 6.090400

54GR:4.308194 6.090400

61GR:0.000000 6.090400

Here is the pertinent declaration in my main() code:

Code:

int main()
{
 * * *
  double GRbeta[GRma][0], GRalpha[GRma][GRma];
  double **beta, **alpha, *a;
  double GRX, GRN;

//  GRX = 430.8194;
  GRX = 4.308194;
  GRN = 6.0904;

  GR[0] = GRX;
  GR[1] = GRN;

  alpha = dmatrix(1,GRma,1,GRma);
  beta  = dmatrix(1,GRma,0,0);
  a     = dvector(1,GRma);

  for(i=1; i<=GRma; i++)
    {
      alpha[i] = GRalpha[i];
      beta[i]  = GRbeta[i];
    }
  a = GR;

  oFile << "GR:" << a[0] << " " << a[1] << endl;

  fGR(x, y, sig, alpha, beta, GRma, GRchi, GRnu, GRQ, a);

 * * *
}

Here is the function as of now:

Code:

void fGR(double x[NSTEPS], double y[NSTEPS], double sig[NSTEPS],
         double **GRalpha, double **GRbeta, int ma, double chi,
         int nu, double Q, double *GRa)
{}

That is everything relevant to this problem. If there's anything I've missed, I'm basing this on the Numerical Recipes' gaussj and mrqmin functions.

As Mats pointed out, GRalpha[GRma] and GRbeta[GRma] just don't exist (arrays in C go from 0 to n-1, not 1 to n).