C Array being overwritten? [closed] - c

Question

Closed. This question needs debugging details. It is not currently accepting answers.
Edit the question to include desired behavior, a specific problem or error, and the shortest code necessary to reproduce the problem. This will help others answer the question.
Closed 6 years ago.
Improve this question
When I run this program : http://hastebin.com/asorawoluw.m
I get this error in GDB :
Program received signal SIGSEGV, Segmentation fault.
0x0000000000401f91 in resoudre (baie=...) at lineaire.c:291
291 printf("type[%d] : %d\n", i, helper_glpk.col_bounds[i]->type);
When i ask gdb to print iI get :
$1 = 1
So the first iteration is failing, but I'm sure I did write into the first case of helper_glpk.col_bounds at the line 200-204 and I did malloc so there's no way (I think ?) that my data is being overwritten or deleted.. so I don't understand why I get this error.
Edit : here's the minimal code :
My structs :
typedef struct Bounds Bounds;
struct Bounds
{
int type;
double lb;
double ub;
};
typedef struct HelperGlpk HelperGlpk;
struct HelperGlpk
{
double *matrix_coefs;
double *obj_coefs;
Bounds **row_bounds;
Bounds **col_bounds;
int *column_of_coef;
int *row_of_coef;
int cpt_coef;
int cpt_contrainte;
};
My functions to generates constraint :
void genere_contrainte_1(int i, int j, HelperGlpk *helper_glpk, Baie baie){
helper_glpk->col_bounds[index_ouverture_serveur(i)]->type = GLP_DB;
helper_glpk->col_bounds[index_ouverture_serveur(i)]->lb = 0;
helper_glpk->col_bounds[index_ouverture_serveur(i)]->ub = 1;
helper_glpk->cpt_coef++;
helper_glpk->col_bounds[index_connexion(i, j, baie.nbr_serveur)]->type = GLP_LO;
helper_glpk->col_bounds[index_connexion(i, j, baie.nbr_serveur)]->lb = 0;
helper_glpk->col_bounds[index_connexion(i, j, baie.nbr_serveur)]->ub = 0;
helper_glpk->cpt_coef++;
}
and the main program is :
void resoudre(Baie baie){
glp_prob *lp;
const int nbr_rows = baie.nbr_client + baie.nbr_serveur * baie.nbr_client; // nombre de contrainte
const int nbr_colums = baie.nbr_serveur + baie.nbr_serveur * baie.nbr_client; // nombre de variable
const int nbr_coefs = 3 * baie.nbr_serveur * baie.nbr_client;
int i, j;
HelperGlpk helper_glpk;
helper_glpk.matrix_coefs = malloc((nbr_coefs + 1) * sizeof(double));
helper_glpk.matrix_coefs[0] = 0;
helper_glpk.obj_coefs = malloc((nbr_colums + 1) * sizeof(double));
helper_glpk.obj_coefs[0] = 0;
helper_glpk.column_of_coef = malloc((nbr_colums + 1) * sizeof(int));
helper_glpk.column_of_coef[0] = 0;
helper_glpk.row_of_coef = malloc((nbr_rows + 1) * sizeof(int));
helper_glpk.row_of_coef[0] = 0;
helper_glpk.col_bounds = malloc((nbr_colums + 1) * sizeof(Bounds *));
for (int index = 0; index <= nbr_colums; index++)
{
helper_glpk.col_bounds[index] = malloc(sizeof(Bounds));
}
helper_glpk.row_bounds = malloc((nbr_rows + 1) * sizeof(Bounds *));
for (int index = 0; index <= nbr_rows; index++)
{
helper_glpk.row_bounds[index] = malloc(sizeof(Bounds));
}
helper_glpk.cpt_coef = 1;
for(i = 1; i <= baie.nbr_serveur; i++)
for(j = 1; j <= baie.nbr_client; j++)
genere_contrainte_1(i, j, &helper_glpk, baie);
for(i = 1; i <= nbr_colums; i++)
printf("type[%d] : %d\n", i, helper_glpk.col_bounds[i]->type);
for(j = 1; j <= baie.nbr_client; j++)
genere_contrainte_2(j, &helper_glpk, baie.nbr_serveur);
The error I get is while I try to printf after the call to generate_contrainte_1

Answer 1

This code is wrong:
helper_glpk.col_bounds = malloc((nbr_colums + 1) * sizeof(Bounds));
You need to fix it with (provided you need nbr_colums + 1 elements):
helper_glpk.col_bounds = malloc((nbr_colums + 1) * sizeof(Bounds *));
for (int index = 0; index < nbr_colums + 1; index++)
{
helper_glpk.col_bounds[index] = malloc(sizeof(Bounds));
}
I have not checked the rest of the code, there could be other errors.
Edit: maybe you don't need the for loop depending on what your genere_contrainte_1 does, but you need to correct your malloc with the right sizeof.
Edit2: I read your genere_contrainte_1, you definitely need all these mallocs. But I really doubt you need row_bounds and col_bounds to be Bounds **, it seems to me Bounds * would have been fine, and that way a single malloc for each field would be enough.