dynamic array segmentation fault - c

Here are two codes one uses 2d array and other uses pointers.
The catch is the code with pointers runs fine and the other with 2d array gives segmentation fault in some test cases(not all). It passes all testcases n<=1000 .
Limit of n is 1< n <10000.
The logic of both code are correct.
Problem:
https://www.hackerrank.com/challenges/dynamic-array/problem
I need to know why the use of 2d array is giving segmentation fault. Is it bad to use 2d array and use pointers instead. But both are same things.
This is the code which gives segmentation fault
int main()
{
/* Enter your code here. Read input from STDIn. Print output to STDOUT */
int n, q, o, x, y, ti, pi, la, j;
scanf("%d%d",&n,&q);
int v[n][n],a[n];
for (j = 0; j<n; j++)
a[j] = 0;
la=0;
while(q-->0)
{
scanf("%d%d%d",&o,&x,&y);
ti=((x^la)%n);
switch(o)
{
case 1:
v[ti][a[ti]]=y;
a[ti]++;
break;
case 2:
pi = y % a[ti];
printf("%d\n", v[ti][pi]);
la = v[ti][pi];
break;
}
}
return 0;
}
This code run fine
int main()
{
int n, q, o, x, y, ti, pi, la;
int *a, *s;
int **v;
scanf("%d %d", &n, &q);
v = (int **)malloc(n * sizeof(int *));
a = (int *)malloc(n * sizeof(int));
s = (int *)malloc(n * sizeof(int));
for (int i = 0; i < n; i++)
{
s[i] = 1;
v[i] = (int *)malloc(s[i] * sizeof(int));
a[i] = 0;
}
la = 0;
for (int i = 0; i < q; i++)
{
scanf("%d %d %d", &o, &x, &y);
ti = (x ^ la) % n;
if (o == 1)
{
v[ti][a[ti]] = y;
a[ti] ++;
if (a[ti] == s[ti])
{
s[ti] *= 2;
v[ti] = realloc(v[ti], s[ti] * sizeof(int));
}
}
else
{
pi = y % a[ti];
printf("%d\n", v[ti][pi]);
la = v[ti][pi];
}
}
free(s);
free(a);
free(v);
return 0;
}
THE TEST CASES ARE AS FOLLOWS
n=99999 q=99977
1 624371833 212297989
1 93347849 234235833
1 683498501 669278504
1 364574347 423183716 continues..
view full testcase here
http://www.writeurl.com/text/z2qlksiktbd7aeuvy3zj/dvgaqxcmeugthbkutrtn

Related

Finding Prime and Composite Elements in an array. Print primes in ascending order and composite in descending order

I am successful in identifying prime and composite from an array. But my qsort function seem to not have any effect when I print the output. I need the primes to be ascending and composite to be descending. When I run the code, it does not sort the output, though it identifies primes and composites.
#include <stdio.h>
#include <stdlib.h>
int compare_Asc(const void *a_void, const void *b_void) {
int a = *(int *)a_void;
int b = *(int *)b_void;
return a - b;
}
int compare_Desc(const void *a_void, const void *b_void) {
int a = *(int *)a_void;
int b = *(int *)b_void;
return b - a;
}
int main() {
int i = 0, n, x, p, c, z, w, j = 0, k = 0, cmpst, null;
int prm;
int prime[50], composite[50], input[50];
printf("How many inputs are you be working with?\nNote: 50 Maximum Inputs\n");
scanf("%d", &n);
printf("Enter the numbers.\n", n);
for (i = 0; i < n; i++) {
scanf("%d", &input[i]);;
}
for (i = 0; i < n; i++) {
if (input[i] % 2 != 0) {
prime[p++] = input[i];
prm = p;
} else
if (input[i] >= 2 && input[i] % 2 == 0) {
composite[c++] = input[i];
cmpst = c;
}
}
printf("Prime Numbers:");
qsort(prime, prm, sizeof(int), compare_Asc);
for (i = 0; i < p; i++) {
printf("%d", prime[p]);
}
printf("Composite Numbers:");
qsort(composite, cmpst, sizeof(int), compare_Desc);
for (i = 0; i < c; i++) {
printf("%d", composite[c]);
}
return 0;
}
There are some major issues, in the posted code, worth mentioning.
Variables
Declaring all the variables at the beginning of the scope, instead of just before where they are used, can hide bugs.
Uninitialized variables, are an even worse source of errors, because their values are indeterminated.
int i=0, n, x, p, c, z, w, j=0, k=0, cmpst, null;
// ^ ^ ^^^^ ?
// ... Later, in the code:
prime[p++] = input[i];
// ^^^ What value is incremented?
// Where is [p++]? Is it inside the prime array?
A correct initialization would prevent undefined behavior.
int p = 0, c = 0;
int composite[50], input[50];
for(int i = 0; i < n ; ++i) {
if ( is_prime(input[i]) ) { // <-- More on this, later.
prime[p++] = input[i];
}
else {
composite[c++] = input[i];
}
}
Loops
This happens a couple of times, just because the code itself is duplicated (another code smell):
for(i=0;i<p;i++){
// ^^^^^^^^^^^ We want to iterate over [0, p).
printf("%d",prime[p]);
// ^ But this always prints the element one past the end
}
Even if it's just a simple loop, it could be a good idea to write a (testable and reusable) function
void print_arr(size_t n, int arr[n])
{
for (size_t i = 0; i < n; ++i) {
printf("%d ", arr[i]);
} // ^
putchar('\n');
}
// ... Later, in main:
print_arr(p, prime);
print_arr(c, composite);
Primes or composite
I am successful in identifying prime and composite from an array
Well, no. Not with this code, I'm sorry.
if (input[i]%2 != 0) { // Those are ALL the ODD numbers!
prime[p++]=input[i];
}
else if(input[i]>=2 && input[i]%2==0){ // Those are the EVEN numbers greater than 0
composite[c++]=input[i];
}
// What about 0 and the even numbers less than 0?
Not all the odd numbers are prime number (it's a little more complicated than that) and 2 itself is a prime, not a composite.
It's unclear to me if this is a terminology issue or if the snippet is only a placeholder for a proper algorithm. In any case, there are multiple examples of primality test functions in SE sites (I'm quite confident some are posted almost every day).
Overflow risk
See chux - Reinstate Monica's comment:
return a-b; risks overflow when a, b are large int values.
Consider return (a > b) - (a < b); for a full range solution.
Single letter variables names are to be avoided... except for i, j and k used in for() loops only.
You're not updating the index of the arrays c and p as the numbers are being printed out. The arrays are being sorted fine.
In the code below I also remove redundant variables, and rename n to input_count, c to compo_count and p to prime_count.
#include <stdio.h>
#include <stdlib.h>
int compare_Asc(const void *a_void, const void *b_void)
{
int a = *(int *) a_void;
int b = *(int *) b_void;
return a - b;
}
int compare_Desc(const void *a_void, const void *b_void)
{
int a = *(int *) a_void;
int b = *(int *) b_void;
return b - a;
}
int main ()
{
int i = 0;
int input_count = 0;
int prime_count = 0;
int compo_count = 0;
int prime[50];
int composite[50];
int input[50];
printf("How many inputs are you be working with?\nNote: 50 Maximum Inputs\n");
scanf("%d", &input_count);
printf("Enter the %d numbers.\n", input_count);
for (i = 0; i < input_count; i++)
{
scanf("%d", &input[i]);
}
for (i = 0; i < input_count; i++)
{
if (input[i] % 2 != 0)
{
prime[prime_count] = input[i];
prime_count += 1;
}
else if (input[i] >= 2 && input[i] % 2 == 0)
{
composite[compo_count] = input[i];
compo_count += 1;
}
}
printf("Prime Numbers:");
qsort(prime, prime_count, sizeof(int), compare_Asc);
for (i = 0; i < prime_count; i++)
{
printf("%d ", prime[i]); // <<-- HERE, not [p]
}
printf( "\n" );
printf ("Composite Numbers:");
qsort(composite, compo_count, sizeof(int), compare_Desc);
for (i = 0; i < compo_count; i++)
{
printf("%d", composite[i]); // <<-- HERE, not [c]
}
printf( "\n" );
return 0;
}

Sum elements of array structure with a pointer to that structure

I have a problem where I have this structure
typedef struct _COMPLEX {
double real, imag;
}COMPLEX;
and I need to declare an array of this structure and sum the elements of the array. Here is my code.
#include <stdio.h>
#include <stdlib.h>
void* xmalloc(size_t nrOcteti);
int main()
{
COMPLEX* v = 0, *s = 0;
int n, i;
printf("\n n = ");
scanf("%d", &n);
v = (COMPLEX*)xmalloc((n)*sizeof(COMPLEX));
s = (COMPLEX*)xmalloc(sizeof(COMPLEX));
for(i = 0; i < n; i++)
{
printf("\n v[%d].real and imag:", i);
scanf("%lf %lf", &v[i].real, &v[i].imag);
}
for(i = 0; i < n; i++)
printf("V%d after scan=%.2lf + %.2lf * i\n", i, v[i].real, v[i].imag);
s->real = 0; s->imag = 0;
for(i = 0; i < n; i++)
{
s->real = s->real + v[i].real;
s->imag = s->imag + v[i].imag;
}
for(i = 0; i < n; i++)
printf("V%d after sum=%.2lf + %.2lf * i\n", i, v[i].real, v[i].imag);
printf("\nS=%lf + %lf\n", s->real, s->imag);
if(s) free(s);
s = 0;
if(v) free(v);
v = 0;
return 0;
}
void* xmalloc(size_t nrOcteti)
{
void *p = 0;
p = malloc(sizeof(nrOcteti));
if(!p)
{
fprintf(stderr, "Allocation failed!");
exit(EXIT_FAILURE);
}
return p;
}
After I give the elements of the array I print them everything is alright, but just before the sum the elements of the array are changed(and for multiple test, apparently the element with the index 2 is modified) and the sum at the end is incorrect. And sometimes(I think because of different inputs) it gives me at the end Segmentation fault because of the free().
The array and sum needs to be pointers to that struct and need to be dynamically allocated.
I tried many times and I can't manage to make it work properly.
If somebody can help me solve this it will be a blessing =))
The line
p = malloc(sizeof(nrOcteti));
in the function xmalloc() is wrong. This line is ignoring what is passed as the argument and just allocating for one size_t.
It should be
p = malloc(nrOcteti);
to allocate specified size.

Troubles with implementation of cooley FFT

I am doing a convolution of two integer signals with the help of FFT, but somehow I can't get it right. I am not sure if my implementation of FFT is correct. Especially the math part.
big edit:
I posted all the code now. My apologies for not starting with it. I was sure the error was only in FFT part, but there might be more problems I overlooked. I know the code is messy and not clean. Everything is a bit fragmented and can be programmed in a simpler and cleaner way, but I was testing bit by bit. As for input it reads two signals from the command line. build up as a number indicating how big the signal is and the signal presented as an integer array e.q 2: [1,-1] and 10: [0,0,0,1,1,1,1,0,0,0]. It should then do a convolution on the signals by performing a FFT on both of them then do bit wise multiplication. With a inverse FFT on the resulting signal. Printing it again with the length and then the array consiting of integers. The printing itself is correct, but the values in the resulting array is not correct. I hope it is all a bit clearer now again my apologies and thank you for your help so far.
#include <stdio.h>
#include <stdlib.h>
#include <math.h>
#include <complex.h>
double PI;
int *readSignal(int *len) { //reads the signal
int *x;
char c;
scanf("%d:", len);
x = calloc(*len, sizeof(int));
do c = getchar(); while (c != '[');
if (len > 0) {
scanf("%d", &x[0]);
for (int i=1; i < *len; i++) scanf(",%d", &x[i]);
}
do c = getchar(); while (c != ']');
return x;
}
void printSignal(int len, int *x) { //prints the signal
printf("%d: [", len);
if (len > 0) {
printf("%d", x[0]);
for (int i=1; i < len; i++)
printf(",%d", x[i]);
}
printf("]\n");
}
void *padSignal(int len, int lenSig, int *x) { //ensures that the signal is of size 2^n by padding it with 0's
int *padded;
padded = calloc(len, sizeof(int));
for (int i=0; i < lenSig; i++) {
padded[i] = x[i];
}
return padded;
}
void fft(double complex signal[], int length, int power) {
if (length == 1) {
return;
}
double complex *signalODD = calloc((length/2+1), sizeof(double complex));
double complex *signalEVEN = calloc((length/2+1), sizeof(double complex));
int index1 = 0;
int index2 = 0;
for(int i = 0; i < length; i++) {
if(i % 2 ==0) {
signalEVEN[index1] = signal[i];
index1++;
}
else {
signalODD[index2] = signal[i];
index2++;
}
}
fft(signalEVEN,length/2, power+1);
fft(signalODD,length/2, power+1);
for(int i = 0; i<length/2-1; i++) {
signal[i] = signalEVEN[i] + cexp((I*2*PI*i)/length)*signalODD[i];
signal[i+length/2] = signalEVEN[i]-cexp((I*2*PI*i)/length)*signalODD[i];
}
free(signalODD);
free(signalEVEN);
}
void ifft(double complex signal[], int length, int power) {
if (length == 1) {
return;
}
double complex *signalODD = calloc((length/2+1), sizeof(double complex));
double complex *signalEVEN = calloc((length/2+1), sizeof(double complex));
int index1 = 0;
int index2 = 0;
for(int i = 0; i < length; i++) {
if(i % 2 ==0) {
signalEVEN[index1] = signal[i];
index1++;
}
else {
signalODD[index2] = signal[i];
index2++;
}
}
fft(signalEVEN,length/2, power+1);
ifft(signalODD,length/2, power+1);
for(int i = 0; i<length/2-1; i++) {
signal[i] = signalEVEN[i] + cexp((I*-2*PI*i)/length)*signalODD[i];
signal[i+length/2] = signalEVEN[i]-cexp((I*-2*PI*i)/length)*signalODD[i];
}
free(signalODD);
free(signalEVEN);
}
int checkPowerofTwo(double len) { //checks for the closed power of 2
double x = 1;
while(len > pow(2,x)) {
x++;
}
return pow(2,x);
}
int main(int argc, char *argv[]) {
int lenH, *H;
int lenX, *X;
int *paddedX;
int *paddedH;
double length;
H=readSignal(&lenH); //reads in the signal H
X=readSignal(&lenX); //reads in signal X
length = lenH+lenX-1;
paddedH=padSignal((length),lenH,H); //pads the signal to the length
paddedX=padSignal((length),lenX,X); // pads the signal to the length
double complex *signalX = calloc(length, sizeof(double complex)); //creats a complex signal X and fills it with paddedX
for (int i = 0; i<length; i++) {
signalX[i] = paddedX[i];
}
double complex *signalH = calloc(length, sizeof(double complex)); // same for H
for (int i = 0; i<length; i++) {
signalH[i] = paddedH[i];
}
fft(signalX, length, 1); //performs the fast fourier transform on X
fft(signalH,length, 1); // performs the fast fourier transfom on H
double complex *signalY = calloc(length, sizeof(double complex)); //makes complex signal Y
for (int i = 0; i<length; i++) { //performs the convolution
signalY[i] = signalX[i]*signalH[i];
}
ifft(signalY, length,1);
int *output = calloc(length, sizeof(int)); //creates the final output signal
for (int i = 0; i<length; i++) {
output[i] = creal(signalY[i]);
}
printSignal(length,output);
free(signalX);
free(signalH);
free(signalY);
free(H);
free(X);
free(paddedH);
free(paddedX);
free(output);
return 0;
}
In:
if(i % 2 ==0 && i != 0)
Why do you exclude i == 0? Change that to if(i % 2 ==0) in both fft and ifft.
In both fft and ifft, the line:
for(int i = 0; i<lenght/2-1; i++) {
should be:
for(int i = 0; i<lenght/2; i++) {
In ifft, the recursion accidentally uses fft:
fft(signalEVEN,lenght/2, power+1);
fft(signalODD,lenght/2, power+1);
Change those to ifft.
The calloc calls do not need this much space:
calloc((lenght/2 + 1), sizeof(double complex));
That can be:
calloc((lenght/2), sizeof(double complex));
Also, the proper spelling of is “length”.
With those fixed, the fft and ifft routines appear to work for some superficial cases.

segmantation fault malloc pointers functions

hello guys this is my code :
#include <stdio.h>
#include <stdlib.h>
int power(int a, int b) {
int exponent = b, result = 1;
while (exponent != 0) {
result = result * a;
exponent--;
}
//printf("%d",result);
return result;
}
int fill_it(char ** p, int N, int fliptimes, int column2) {
if (N < 0) return 0;
int counter = 0, l;
char a = 'H';
for (l = 0; l < power(2, fliptimes); l++) {
p[l][column2] = a;
counter++;
if (counter == (power(2, N) / 2)) {
counter = 0;
if (a == 'H') a = 'T';
if (a == 'T') a = 'H';
}
}
fill_it(p, N--, fliptimes, column2++);
}
int main() {
int i, fores, j, l, m;
char ** p;
printf("how many times did you toss the coin?:");
scanf("%d", & fores);
p = (char ** ) malloc((power(2, fores)) * sizeof(char * ));
for (i = 0; i < fores; i++)
p[i] = (char * ) malloc(fores * sizeof(char));
fill_it(p, fores, fores, 0);
for (l = 0; l < power(2, fores); l++) {
for (m = 0; m < fores; m++) {
printf("%c", p[l][m]);
}
}
printf(",");
}
it does compile.But when i run the program it returns a "segmantation fault (core dumped)" error
i know it means that i tried to access memory,i dont have acces to but i dont understand which part of the program is defective
The problem is, you're not allocating enough memory. This line is fine
p = (char ** ) malloc((power(2, fores)) * sizeof(char * ));
but this loop is only allocating memory for part of the 2-dimensional array.
for (i = 0; i < fores; i++)
p[i] = (char * ) malloc(fores * sizeof(char));
The memory allocation should look more like this...
foresSquared = power(2, fores);
p = malloc(foresSquared*sizeof(char *));
for (i = 0; i < foresSquared; i++)
p[i] = malloc(fores);
Since the result of power is going to be consistent, it makes sense to store the value in a variable and use that rather than recalculating it. It'll make the code clearer too.
You also don't need to cast the return value of malloc as C handles that for you. And sizeof(char) isn't needed as it's guaranteed to always be 1.

Passing argument makes a pointer without cast?

I'm having an issue with compling the code.
I keep getting these errors.
"C: 194 warning passing argument 3 of 'matrix_column_subtract' makes pointer from integer without a cast"
"C: 12 note: expected 'double**' but argument is type 'int'
"C: 194 error too few arguments to function 'matrix_column_subtract'
I think I know what is going on I'm calling matrix_column_multiply is a void and I need to be calling it a pointer pointer I think and I don't know how to chage that. If anyone has some idea on how I can get this to compile that would be highly appreciated!!
#include <stdio.h>
#include <stdlib.h>
#include <math.h>
#define DEBUG 0
#define MAX_ITER 10000
double *eigen (int n, double **A);
void qr_decomp (int n, double **A, double **Q, double **R);
void matrix_copy_column(double **msrc, int col1, double **mdst,int col2, int rows);
void matrix_column_subtract(double **m1, int c1, double **m2, int c2, int rows);
void matrix_column_multiply(double **m, int c, double k, int rows);
int main() {
int i, n = 126;
double *eig, **Am;
FILE *BinInp, *TxtOut;
/* Input Code: Reads bv, Am in binary form from CGrad.bin */
if ( (BinInp = fopen("CGrad.bin","r")) == NULL ) {
fprintf(stderr, "Cannot open matrix binary file INPUT... exiting\n");
exit(-1);
}
Am = (double**)malloc (n*sizeof(double*));
Am[0] = (double*)malloc (n*n*sizeof(double));
for (i = 1; i < n; ++i) {
Am[i] = Am[0] + i*n;
}
for (i = 0; i < n; i++) {
if (i==0) { /* Read one extra line that is discarded (bv is still in bin file) */
if (!fread(Am[i], sizeof(double), n, BinInp)) {
fprintf(stderr, "Cannot read row Am[%03d] of matrix... exiting\n", i+1);
exit(1);
}
}
if (!fread(Am[i], sizeof(double), n, BinInp)) {
fprintf(stderr, "Cannot read row Am[%03d] of matrix... exiting\n", i+1);
exit(1);
}
}
if (fclose(BinInp) == EOF) {
fprintf(stderr, "Cannot close matrix binary file INPUT... exiting\n");
exit(-1);
}
/* COMPUTE EIGENVALUES HERE USING FUNCTIONS. RETURN EIGENVALUES (AS 1D VECTOR) TO eig. */
if (DEBUG) printf ("Calling eigen\n");
eig = eigen (n, Am);
/* Output Code: Writes eig in text form to Eigs.txt */
if ( (TxtOut = fopen("Eigs.txt", "w")) == NULL ) {
fprintf(stderr, "Cannot open matrix text file OUTPUT... exiting\n");
exit(-1);
}
for (i = 0; i < n; i++) {
fprintf (TxtOut, "%18.14e ", eig[i]);
}
if (fclose(TxtOut) == EOF) {
fprintf(stderr, "Cannot close matrix text file INPUT... exiting\n");
exit(-1);
}
return 0;
}
double* eigen (int n, double **Acur)
{
double err = 1, eps = 1e-2;
double ndenom, nnumer, temp;
double *eig, **Anex, **Qsub, **Rsub;
int i, j, k, iters = 1;
/* Malloc memory for the three matricies */
Anex = malloc (n*sizeof(double*));
Anex[0] = malloc (n*n*sizeof(double));
for (i = 1; i < n; ++i) {
Anex[i] = Anex[0] + i*n;
}
Qsub = malloc (n*sizeof(double*));
Qsub[0] = malloc (n*n*sizeof(double));
for (i = 1; i < n; ++i) {
Qsub[i] = Qsub[0] + i*n;
}
Rsub = malloc (n*sizeof(double*));
Rsub[0] = malloc (n*n*sizeof(double));
for (i = 1; i < n; ++i) {
Rsub[i] = Rsub[0] + i*n;
}
/* Malloc memory for the return eig vector */
eig = malloc (n*sizeof(double));
for (i = 0; i < n; i++) {
eig[i] = 0;
}
/* Enter main iteration loop for eigenvalues */
while (err > eps && iters < MAX_ITER) {
/* QR Decompose Acur then find next iterate value in Anex */
qr_decomp (n, Acur, Qsub, Rsub);
// FIND NEXT ITERATE VALUE, PUT IN Anex.
/* Determine relative error change, reset "old" iterate value. */
ndenom = 0;
nnumer = 0;
for (i = 0; i < n; i++) {
for (j = 0; j < n; j++) {
temp = Anex[i][j]-Acur[i][j];
ndenom += temp*temp;
nnumer += Anex[i][j]*Anex[i][j];
Acur[i][j] = Anex[i][j];
}
}
err = sqrt(ndenom)/sqrt(nnumer);
/* Increment the iteration count and report error */
if (iters % 25 == 0) {
printf ("Error at end of iteration %05d = %14.10f %%\n", iters, 100*err);
}
++iters;
}
printf ("Error at end of iteration %05d = %14.10f\nCONVERGED.\n", iters-1, err);
if (iters == MAX_ITER) {
printf ("WARNING: MAX_ITER iterations reached!...\n");
}
/* Copy diagonal entries of Acur into eig for return to main */
for (i=0; i<n; i++) {
eig[i] = Acur[i][i];
}
return eig;
}
void qr_decomp (int n, double **Adec, double **myQ, double **myR)
{
int i, j, k; /* Loop Variables: this is all you should need! */
double **T, **S;
double r;
T = malloc (n*sizeof(double*));
T[0] = malloc (n*n*sizeof(double));
for (i = 1; i < n; ++i) {
T[i] = T[0] + i*n;
}
S = malloc (n*sizeof(double*));
S[0] = malloc (n*n*sizeof(double));
for (i = 1; i < n; ++i) {
S[i] = S[0] + i*n;
}
/* Main loop for decomposition */
for (i=0; i<n; i++) {
/* Column i of Q is initially column i of A */
matrix_copy_column(Adec,i,myQ,i,n);
/* For j < i-1, Perform the dot product between the j row of Q and the
i row of A to determine the R(j,i) value, then insure the Q i column
is orthogonal to all other Q columns by subtracting existing Q columns
from it. */
for (j = 0; j < i; j++) {
//r[j,i] = Qj^T * Ui
matrix_copy_column(myQ,j,T,0,n);
matrix_copy_column(Adec,i,S,0,n);
for (k=0; k<n; k++) {
r += T[k][0] * S[k][0];
}
/* Determine the R diagonal as the magnitude of the Q column, then
normalize the Q column (make it a unit vector). */
//Qi = Ui
myR[j][i] = r;
// Something wrong here.
// There is one parameter missing, as matrix_column_subtract needs 5 parameters and
// only 4 are given.
// Also, matrix_column_multiply is defined as returning a void, whereas the 3rd parameter
// of matrix_column_subtract should be a double **
matrix_column_subtract(myQ,i,matrix_column_multiply(T,0,r,n),j);
}
}
}
/* Copies a matrix column from msrc at column col1 to mdst at column col2 */
void matrix_copy_column(double **msrc, int col1, double **mdst,int col2, int rows) {
int i = 0;
for (i=0; i<rows; i++) {
mdst[i][col2] = msrc[i][col1];
}
}
/* Subtracts m2's column c2 from m1's column c1 */
void matrix_column_subtract(double **m1, int c1, double **m2, int c2, int rows) {
int i = 0;
for (i=0; i<rows; i++) {
m1[i][c1] -= m2[i][c2];
}
/*return m1;*/
}
void matrix_column_multiply(double **m, int c, double k, int rows) {
int i = 0;
for (i=0; i<rows; i++) {
m[i][c] *= k;
}
/*return m;*/
}
A problem is here.
matrix_column_subtract(myQ,i,matrix_column_multiply(T,0,r,n),j);
The relevant function signatures are:
void matrix_column_subtract(double **m1, int c1, double **m2, int c2, int rows);
void matrix_column_multiply(double **m, int c, double k, int rows);
Note that matrix_column_multiply returns void, but you're passing it in as if it were double **. I'm guessing the compiler is desperately trying to make it work and returned some sort of garbage integer.
These functions return void because they do their work directly on the matrix in question. matrix_column_multiply alters m. matrix_column_subtract alters m1.
To make it work, call matrix_column_multiply on T then pass T (now modified by the multiplication) into matrix_column_subtract.
matrix_column_multiply(T, 0, r, n);
matrix_column_subtract(myQ, i, T, j, ...rows...);
You're still missing the fifth argument, rows. I'm going to guess that's n, same as has been used for the rows in every other matrix_column_blah call.

Resources