I'm working on a C implementation for Conway's game of life, I have been asked to use the following header:
#ifndef game_of_life_h
#define game_of_life_h
#include <stdio.h>
#include <stdlib.h>
// a structure containing a square board for the game and its size
typedef struct gol{
int **board;
size_t size;
} gol;
// dynamically creates a struct gol of size 20 and returns a pointer to it
gol* create_default_gol();
// creates dynamically a struct gol of a specified size and returns a pointer to it.
gol* create_gol(size_t size);
// destroy gol structures
void destroy_gol(gol* g);
// the board of 'g' is set to 'b'. You do not need to check if 'b' has a proper size and values
void set_pattern(gol* g, int** b);
// using rules of the game of life, the function sets next pattern to the g->board
void next_pattern(gol* g);
/* returns sum of all the neighbours of the cell g->board[i][j]. The function is an auxiliary
function and should be used in the following function. */
int neighbour_sum(gol* g, int i, int j);
// prints the current pattern of the g-board on the screen
void print(gol* g);
#endif
I have added the comments to help out with an explanation of what each bit is.
gol.board is a 2-level integer array, containing x and y coordinates, ie board[x][y], each coordinate can either be a 1 (alive) or 0 (dead).
This was all a bit of background information, I'm trying to write my first function create_default_gol() that will return a pointer to a gol instance, with a 20x20 board.
I then attempt to go through each coordinate through the 20x20 board and set it to 0, I am getting a Segmentation fault (core dumped) when running this program.
The below code is my c file containing the core code, and the main() function:
#include "game_of_life.h"
int main()
{
// Create a 20x20 game
gol* g_temp = create_default_gol();
int x,y;
for (x = 0; x < 20; x++)
{
for (y = 0; y < 20; y++)
{
g_temp->board[x][y] = 0;
}
}
free(g_temp);
}
// return a pointer to a 20x20 game of life
gol* create_default_gol()
{
gol* g_rtn = malloc(sizeof(*g_rtn) + (sizeof(int) * 20 * 20));
return g_rtn;
}
This is the first feature I'd like to implement, being able to generate a 20x20 board with 0's (dead) state for every coordinate.
Please feel free to criticise my code, I'm looking to determine why I'm getting the segmentation fault, and if I'm allocating memory properly in the create_default_gol() function.
Thanks!
The type int **board; means that board must contain an array of pointers, each of which points to the start of each row. Your existing allocation omits this, and just allocates *g_rtn plus the ints in the board.
The canonical way to allocate your board, supposing that you must stick to the type int **board;, is:
gol* g_rtn = malloc(sizeof *g_rtn);
g_rtn->size = size;
g_rtn->board = malloc(size * sizeof *g_rtn->board);
for (int i = 0; i < size; ++i)
g_rtn->board[i] = malloc(size * sizeof **g_rtn->board);
This code involves a lot of small malloc chunks. You could condense the board rows and columns into a single allocation, but then you also need to set up pointers to the start of each row, because board must be an array of pointers to int.
Another issue with this approach is alignment. It's guaranteed that a malloc result is aligned for any type; however it is possible that int has stricter alignment requirements than int *. My following code assumes that it doesn't; if you want to be portable then you could add in some compile-time checks (or run it and see if it aborts!).
The amount of memory required is the sum of the last two mallocs:
g_rtn->board = malloc( size * size * sizeof **g_rtn->board
+ size * sizeof *g_rtn->board );
Then the first row will start after the end of the row-pointers (a cast is necessary because we are converting int ** to int *, and using void * means we don't have to repeat the word int):
g_rtn->board[0] = (void *) (g_rtn->board + size);
And the other rows each have size ints in them:
for (int i = 1; i < size; ++i)
g_rtn->board[i] = g_rtn->board[i-1] + size;
Note that this is a whole lot more complicated than just using a 1-D array and doing arithmetic for the offsets, but it was stipulated that you must have two levels of indirection to access the board.
Also this is more complicated than the "canonical" version. In this version we are trading code complexity for the benefit of having a reduced number of mallocs. If your program typically only allocates one board, or a small number of boards, then perhaps this trade-off is not worth it and the canonical version would give you fewer headaches.
Finally - it would be possible to allocate both *g_rtn and the board in the single malloc, as you attempted to do in your question. However my advice (based on experience) is that it is simpler to keep the board separate. It makes your code clearer, and your object easier to use and make changes to, if the board is a separate allocation to the game object.
create_default_gol() misses to initialise board, so applying the [] operator to it (in main() ) the program accesses "invaid" memory and with ethis provokes undefined behaviour.
Although enough memory is allocated, the code still needs to make board point to the memory by doing
gol->board = ((char*) gol) + sizeof(*gol);
Update
As pointed out by Matt McNabb's comment board points to an array of pointers to int, so initialisation is more complicate:
gol * g_rtn = malloc(sizeof(*g_rtn) + 20 * sizeof(*gol->board));
g_rtn->board = ((char*) gol) + sizeof(*gol);
for (size_t i = 0; i<20; ++i)
{
g_rtn->board[i] = malloc(20 * sizeof(*g_rtn->board[i])
}
Also the code misses to set gol's member size. From what you tell us it is not clear whether it shall hold the nuber of bytes, rows/columns or fields.
Also^2 coding "magic numbers" like 20 is bad habit.
Also^3 create_default_gol does not specify any parameters, which explictily allows any numberm and not none as you might perhaps have expected.
All in all I'd code create_default_gol() like this:
gol * create_default_gol(const size_t rows, const size_t columns)
{
size_t size_rows = rows * sizeof(*g_rtn->board));
size_t size_column = columns * sizeof(**g_rtn->board));
gol * g_rtn = malloc(sizeof(*g_rtn) + size_rows);
g_rtn->board = ((char*) gol) + sizeof(*gol);
if (NULL ! = g_rtn)
{
for (size_t i = 0; i<columns; ++i)
{
g_rtn->board[i] = malloc(size_columns); /* TODO: Add error checking here. */
}
g_rtn->size = size_rows * size_columns; /* Or what ever this attribute is meant for. */
}
return g_rtn;
}
gol* create_default_gol()
{
int **a,i;
a = (int**)malloc(20 * sizeof(int *));
for (i = 0; i < 20; i++)
a[i] = (int*)malloc(20 * sizeof(int));
gol* g_rtn = (gol*)malloc(sizeof(*g_rtn));
g_rtn->board = a;
return g_rtn;
}
int main()
{
// Create a 20x20 game
gol* g_temp = create_default_gol();
int x,y;
for (x = 0; x < 20; x++)
{
for (y = 0; y < 20; y++)
{
g_temp->board[x][y] = 10;
}
}
for(x=0;x<20;x++)
free(g_temp->board[x]);
free(g_temp->board);
free(g_temp);
}
main (void)
{
gol* gameOfLife;
gameOfLife = create_default_gol();
free(gameOfLife);
}
gol* create_default_gol()
{
int size = 20;
gol* g_rtn = malloc(sizeof *g_rtn);
g_rtn = malloc(sizeof g_rtn);
g_rtn->size = size;
g_rtn->board = malloc(size * sizeof *g_rtn->board);
int i, b;
for (i = 0; i < size; ++i){
g_rtn->board[i] = malloc(sizeof (int) * size);
for(b=0;b<size;b++){
g_rtn->board[i][b] = 0;
}
}
return g_rtn;
}
Alternatively, since you also need to add a create_gol(size_t new_size) of custom size, you could also write it as the following.
main (void)
{
gol* gameOfLife;
gameOfLife = create_default_gol();
free(gameOfLife);
}
gol* create_default_gol()
{
size_t size = 20;
return create_gol(size);
}
gol* create_gol(size_t new_size)
{
gol* g_rtn = malloc(sizeof *g_rtn);
g_rtn = malloc(sizeof g_rtn);
g_rtn->size = new_size;
g_rtn->board = malloc(size * sizeof *g_rtn->board);
int i, b;
for (i = 0; i < size; ++i){
g_rtn->board[i] = malloc(sizeof (int) * size);
for(b=0;b<size;b++){
g_rtn->board[i][b] = 0;
}
}
return g_rtn;
}
Doing this just minimizes the amount of code needed.
Related
I've a problem about deallocating memory using free() in C.
My program generates a random genealogic tree using a matrix. This matrix can be very huge depending on the number of family members. The program seemed to work fine until I decided to generate more than one tree. I noticed that generating about 100 trees causes my 8GB RAM to fill! I'm sure I can make a better code to reduce the demand of memory, but my problem remains.
I use free() to deallocate memory and there's no error. I installed Valgrind to se what's happening and it says that about 100 million byte per tree are definitely lost. This means that free() doesn't work fine. I don't now where is the problem. I link some functions that I think are correlated to the problem.
typedef struct{
int f_id;
char f_name[L_NAMES];
int generations;
int n_members;
type_people *members;
int_mtx *mtx;
}type_family;
The struct above is for the family.
typedef struct temp{
int p_id;
char name[L_NAMES];
char f_name[L_NAMES];
int generation;
int n_sons;
struct temp **sons;
int f_id;
int sex;
int age;
}type_people;
This is for the members.
typedef struct{
int i;
int j;
int **val;
}int_mtx;
And the matrix.
In the main i call the function to initialize the tree:
type_family *family_a;
family_a = malloc(sizeof(type_family));
family_a = init_family_n_gen(family_a, 6);
This is the frist part of init_family_n_gen():
type_family *init_family_n_gen(type_family *family, int n){
...
family->members = malloc(max_people * sizeof(type_people));
family->mtx = mtxcalloc(family->mtx, max_people, max_people - 1);
...
This code is for mtxcalloc that initializes the matrix:
int_mtx *mtxcalloc(int_mtx *mtx, int i, int j){
mtx = malloc(sizeof(int_mtx));
mtx->i = i;
mtx->j = j;
mtx->val = malloc(i * sizeof(int *));
for(int a = 0; a < i; a++){
mtx->val[a] = malloc(j * sizeof(int));
for(int b = 0; b < j; b++){
mtx->val[a][b] = 0;
}
}
return mtx;
}
And to conclude the code to deallocate the family:
void free_family(type_family *family){
for(int m = 0; m < family->n_members; m++){
if(family->members[m].n_sons != 0){
free(family->members[m].sons);
}
}
mtxfree(family->mtx);
free(family->members);
}
And the one to deallocate the matrix:
void mtxfree(int_mtx *mtx){
for(int i = 0; i < mtx->i; i++){
free(mtx->val[i]);
}
free(mtx->val);
free(mtx);
}
Screen capture of Valgrind output
So I call the free_family(family_a) every time i need to regenerate the family but the memory still increases. (In the photo above the number of byte become 1 billion if i regenerate the family for 50 times).
Thanks for the support!
EDITED
I made a minimal reproducible example that emulates my original code. The structs and variables are the same but I changed the functions according to Weather Vane: they are all void and I pass them the double **.
The init_family_n_gen becomes:
void init_family(type_family **f){
type_family *family = malloc(sizeof(type_family));
family->members = malloc(100 * sizeof(type_people));
for(int m = 0; m < 100; m++){
family->members[m].n_sons = 0;
}
mtxcalloc(&family->mtx, 100, 99);
family->mtx->val[0][1] = 7;
family->mtx->val[9][8] = 1;
mtxrealloc(&family->mtx, 5, 4);
*f = family;
}
The main is:
type_family *family_a;
init_family(&family_a);
free_family(&family_a);
The only thing I added is this function(Is the code right?):
void mtxrealloc(int_mtx **mtx, int i, int j){
(*mtx)->i = i;
(*mtx)->j = j;
(*mtx)->val = realloc((*mtx)->val, (*mtx)->i * sizeof(int *));
for(int a = 0; a < (*mtx)->i; a++){
(*mtx)->val[a] = realloc((*mtx)->val[a], (*mtx)->j * sizeof(int));
}
}
I noticed that the problem occours when i use the realloc function and i can't figure why. I link the images of Valgrind with and without the function mtxrealloc. (I see that there is aslo a 48 byte leak...).
Valgrind with realloc
Valgrind without realloc
Thanks again for your support!
This:
init_family(&family_a);
Causes this code from mtxcalloc to execute:
mtx->val = malloc(i * sizeof(int *));
for(int a = 0; a < i; a++){
mtx->val[a] = malloc(j * sizeof(int));
for(int b = 0; b < j; b++){
mtx->val[a][b] = 0;
}
}
, with i, j = 100, 99. That is, you allocate space for 100 pointers, and for each one, you allocate space for 99 ints. These are then accessible via family_a->mtx.
Very shortly thereafter, you make this call:
mtxrealloc(&family->mtx, 5, 4);
, which does this, among other things:
(*mtx)->val = realloc((*mtx)->val, (*mtx)->i * sizeof(int *));
That loses all the pointers (*mtx)->val[5] through (*mtx)->val[99], each of which is the sole pointer to allocated space sufficient for 99 ints. Overall, sufficient space for 9405 ints is leaked before you even perform any computations with the object you are preparing.
It is unclear why you overallocate, just to immediately (attempt to) free the excess, but perhaps that's an artifact of your code simplification. It would be much better to come up with a way to determine how much space you need in advance, and then allocate only that much in the first place. But if you do need to reallocate this particular data, then you need to first free each of the (*mtx)->val[x] that will be lost. Of course, if you were going to reallocate larger, then you would need to allocate / reallocate all of the (*mtx)->val[x].
Disclaimer: This is homework. I am attempting it and do not expect or want anyone to do it for me. Just a few pointers (hehe) where I'm going wrong would be appreciated.
The homework requires me to create an int* array that holds 10 elements, and then attempt to insert a million ints into it. Each insertion checks if the array needs to be resized, and if it does, I increase it's size so it can hold one more element.
When I insert 10,000 elements, it works fine, but if I try 100,000 elements, I get the following error:
*** glibc detected *** ./set2: realloc(): invalid old size: 0x00000000024dc010 ***
This is the code I'm running. I've commented it so it's easily readable.
void main()
{
//begin with a size of 10
int currentsize = 10;
int* arr = malloc(currentsize * sizeof(int));
int i;
//initalize with all elements set to INT_MAX
for(i = 0; i < currentsize; i++) {
arr[i] = INT_MAX;
}
// insert random elements
for(i = 0; i < 100000; i++) {
currentsize = add(rand() % 100,arr,currentsize);
}
free(arr);
}
/*
Method resizes array if needed, and returns the new size of the array
Also inserts the element into the array
*/
int add(int x, int* arr, int size)
{
//find the first available location
int newSize = size;
int i;
for(i = 0; i < size; i++) {
if (arr[i] == INT_MAX)
break;
}
if (i >= size) {
//need to realloc
newSize++;
arr = realloc(arr, newSize * sizeof(int) );
}
arr[i] = x;
return newSize;
}
The error is probably because you properly use realloc to change arr in the function add, but this modified value is lost when add returns. So the next call to add will receive the old, now bad value.
Also I can't understand why you're using a the for loop to search. You know you want to add at the last element, so why search? Just reallocate the array and plug the new value in the new slot.
Incidentally I'm pretty sure your teacher is trying to get you to see that reallocating for each member causes an asymptotic run time problem. Most implementations of realloc will do a lot of copying with this algorithm. This is why real programs grow the array size by a factor greater than one (often 1.5 or 2) rather than by fixed amounts.
The usual idiom is to abstract the variable size array in a struct:
typedef struct array_s {
int *elts;
int size;
} VARIABLE_ARRAY;
void init(VARIABLE_ARRAY *a)
{
a->size = 10;
a->elts = malloc(a->size * sizeof a->elts[0]);
// CHECK FOR NULL RETURN FROM malloc() HERE
}
void ensure_size(VARIABLE_ARRAY *a, size_t size)
{
if (a->size < size) {
// RESET size HERE TO INCREASE BY FACTOR OF OLD SIZE
// size = 2 * a->size;
a->elts = realloc(size * sizeof a->elts[0]);
a->size = size;
// CHECK FOR NULL RETURN FROM realloc() HERE
}
}
// Set the i'th position of array a. If there wasn't
// enough space, expand the array so there is.
void set(VARIABLE_ARRAY *a, int i, int val)
{
ensure_size(a, i + 1);
a->elts[i] = val;
}
void test(void)
{
VARIABLE_ARRAY a;
init(&a);
for (int i = 0; i < 100000; i++) {
set(&a, i, rand());
}
...
}
I would pass arr to add() as a pointer (to a pointer), so that it can be modified inside of add()
int add(int x, int** arr, int size)
{
// ...
*arr = realloc(*arr, newSize * sizeof(int) );
}
And calling it....
currentsize = add(rand() % 100, &arr, currentsize);
Note that that your code (and my suggested change) is not doing any error checking. You should be checking the return value of malloc and realloc for NULL.
I am an unexperienced C-programmer: I want all the numbers below 5000 that are multiples of 5. Here is how I do this currently:
int main()
{
int i;
const int max =5000-1;
for(i=2; i<(max+1); i++)
{
if(!(i%5))
{
printf("%d\n", i);
}
}
return 0;
}
Say that I want them all listed in an array. What I could do is just to pre-allocate an integer array and fill out the various position. Naturally I can't know the exact required length beforehand, so I would over estimate it length.
However, I come from a C++ background, so normally what I would do there is to pushback a vector, all clean and tidy. But what is the professional way to do this in C? Would you guys pre-allocate or dynamically resize the array?
I am currently using Herbert Schildt's "Turbo C/C++", I'm sure there are much better (and up-to-date) references out there when I get more into things.
realloc does everything you're talking about. Allocating an array, growing an array, shrinking an array: it does it all.
int max = 5000; /* why subtract one if you have to add one to use it? */
int *arr = NULL;
int i;
arr = realloc(arr, max * sizeof *arr); /* allocate generous array */
for (i = 0; i < max; i++) {
/* ... */
}
max = 10000;
arr = realloc(arr, max * sizeof *arr); /* grow array */
max = 100;
arr = realloc(arr, max * sizeof *arr); /* shrink array */
Now there is some popular advice that you should always save the return value from realloc as a separate variable and check it for NULL before overwriting your real pointer variable. This is because there are bizarre situations where the realloc may fail, even on something as innocuous as shrinking an array. This can happen if the malloc subsystem is implemented using fixed-sized buckets, among other possibilities. A shrinking request may fail with a fixed-sized bucket system if there simply aren't any more "small" regions available.
If realloc fails, it returns NULL, but the original allocation is left intact. If you just write the return value into your pointer variable, that data will lost. So, in general, you should try to do this instead:
int *tmp;
tmp = realloc(arr, max * sizeof *arr);
if (tmp) {
arr = tmp;
} else {
/* maybe issue an error message? */
}
If you want to allocate the perfect size, you could try this :
#include <stdio.h>
#include <stdlib.h>
int main(){
int i, j;
int max = 5000;
int * ourNumbers = 0;
int count = 0;
for(i = 2; i < max; i++){
if (i % 5 == 0){
count += 1;
}
}
printf("\ncount = %d\n", count);
ourNumbers = (int *) malloc(sizeof (int) * count);
// and after you can populate your array with those values;
// like this you will allocate the exact memory
}
I know that is not so efficient, but I hope it will help you :)
Disclaimer: This is homework. I am attempting it and do not expect or want anyone to do it for me. Just a few pointers (hehe) where I'm going wrong would be appreciated.
The homework requires me to create an int* array that holds 10 elements, and then attempt to insert a million ints into it. Each insertion checks if the array needs to be resized, and if it does, I increase it's size so it can hold one more element.
When I insert 10,000 elements, it works fine, but if I try 100,000 elements, I get the following error:
*** glibc detected *** ./set2: realloc(): invalid old size: 0x00000000024dc010 ***
This is the code I'm running. I've commented it so it's easily readable.
void main()
{
//begin with a size of 10
int currentsize = 10;
int* arr = malloc(currentsize * sizeof(int));
int i;
//initalize with all elements set to INT_MAX
for(i = 0; i < currentsize; i++) {
arr[i] = INT_MAX;
}
// insert random elements
for(i = 0; i < 100000; i++) {
currentsize = add(rand() % 100,arr,currentsize);
}
free(arr);
}
/*
Method resizes array if needed, and returns the new size of the array
Also inserts the element into the array
*/
int add(int x, int* arr, int size)
{
//find the first available location
int newSize = size;
int i;
for(i = 0; i < size; i++) {
if (arr[i] == INT_MAX)
break;
}
if (i >= size) {
//need to realloc
newSize++;
arr = realloc(arr, newSize * sizeof(int) );
}
arr[i] = x;
return newSize;
}
The error is probably because you properly use realloc to change arr in the function add, but this modified value is lost when add returns. So the next call to add will receive the old, now bad value.
Also I can't understand why you're using a the for loop to search. You know you want to add at the last element, so why search? Just reallocate the array and plug the new value in the new slot.
Incidentally I'm pretty sure your teacher is trying to get you to see that reallocating for each member causes an asymptotic run time problem. Most implementations of realloc will do a lot of copying with this algorithm. This is why real programs grow the array size by a factor greater than one (often 1.5 or 2) rather than by fixed amounts.
The usual idiom is to abstract the variable size array in a struct:
typedef struct array_s {
int *elts;
int size;
} VARIABLE_ARRAY;
void init(VARIABLE_ARRAY *a)
{
a->size = 10;
a->elts = malloc(a->size * sizeof a->elts[0]);
// CHECK FOR NULL RETURN FROM malloc() HERE
}
void ensure_size(VARIABLE_ARRAY *a, size_t size)
{
if (a->size < size) {
// RESET size HERE TO INCREASE BY FACTOR OF OLD SIZE
// size = 2 * a->size;
a->elts = realloc(size * sizeof a->elts[0]);
a->size = size;
// CHECK FOR NULL RETURN FROM realloc() HERE
}
}
// Set the i'th position of array a. If there wasn't
// enough space, expand the array so there is.
void set(VARIABLE_ARRAY *a, int i, int val)
{
ensure_size(a, i + 1);
a->elts[i] = val;
}
void test(void)
{
VARIABLE_ARRAY a;
init(&a);
for (int i = 0; i < 100000; i++) {
set(&a, i, rand());
}
...
}
I would pass arr to add() as a pointer (to a pointer), so that it can be modified inside of add()
int add(int x, int** arr, int size)
{
// ...
*arr = realloc(*arr, newSize * sizeof(int) );
}
And calling it....
currentsize = add(rand() % 100, &arr, currentsize);
Note that that your code (and my suggested change) is not doing any error checking. You should be checking the return value of malloc and realloc for NULL.
I am writing a C-program where I need 2D-arrays (dynamically allocated) with negative indices or where the index does not start at zero. So for an array[i][j] the row-index i should take values from e.g. 1 to 3 and the column-index j should take values from e.g. -1 to 9.
For this purpose I created the following program, here the variable columns_start is set to zero, so just the row-index is shifted and this works really fine.
But when I assign other values than zero to the variable columns_start, I get the message (from valgrind) that the command "free(array[i]);" is invalid.
So my questions are:
Why it is invalid to free the memory that I allocated just before?
How do I have to modify my program to shift the column-index?
Thank you for your help.
#include <stdio.h>
#include <stdlib.h>
main()
{
int **array, **array2;
int rows_end, rows_start, columns_end, columns_start, i, j;
rows_start = 1;
rows_end = 3;
columns_start = 0;
columns_end = 9;
array = malloc((rows_end-rows_start+1) * sizeof(int *));
for(i = 0; i <= (rows_end-rows_start); i++) {
array[i] = malloc((columns_end-columns_start+1) * sizeof(int));
}
array2 = array-rows_start; //shifting row-index
for(i = rows_start; i <= rows_end; i++) {
array2[i] = array[i-rows_start]-columns_start; //shifting column-index
}
for(i = rows_start; i <= rows_end; i++) {
for(j = columns_start; j <= columns_end; j++) {
array2[i][j] = i+j; //writing stuff into array
printf("%i %i %d\n",i, j, array2[i][j]);
}
}
for(i = 0; i <= (rows_end-rows_start); i++) {
free(array[i]);
}
free(array);
}
When you shift column indexes, you assign new values to original array of columns: in
array2[i] = array[i-rows_start]-columns_start;
array2[i] and array[i=rows_start] are the same memory cell as array2 is initialized with array-rows_start.
So deallocation of memory requires reverse shift. Try the following:
free(array[i] + columns_start);
IMHO, such modification of array indexes gives no benefit, while complicating program logic and leading to errors. Try to modify indexes on the fly in single loop.
#include <stdio.h>
#include <stdlib.h>
int main(void) {
int a[] = { -1, 41, 42, 43 };
int *b;//you will always read the data via this pointer
b = &a[1];// 1 is becoming the "zero pivot"
printf("zero: %d\n", b[0]);
printf("-1: %d\n", b[-1]);
return EXIT_SUCCESS;
}
If you don't need just a contiguous block, then you may be better off with hash tables instead.
As far as I can see, your free and malloc looks good. But your shifting doesn't make sense. Why don't you just add an offset in your array instead of using array2:
int maxNegValue = 10;
int myNegValue = -6;
array[x][myNegValue+maxNegValue] = ...;
this way, you're always in the positive range.
For malloc: you acquire (maxNegValue + maxPosValue) * sizeof(...)
Ok I understand now, that you need free(array.. + offset); even using your shifting stuff.. that's probably not what you want. If you don't need a very fast implementation I'd suggest to use a struct containing the offset and an array. Then create a function having this struct and x/y as arguments to allow access to the array.
I don't know why valgrind would complain about that free statement, but there seems to be a lot of pointer juggling going on so it doesn't surprise me that you get this problem in the first place. For instance, one thing which caught my eye is:
array2 = array-rows_start;
This will make array2[0] dereference memory which you didn't allocate. I fear it's just a matter of time until you get the offset calcuations wrong and run into this problem.
One one comment you wrote
but im my program I need a lot of these arrays with all different beginning indices, so I hope to find a more elegant solution instead of defining two offsets for every array.
I think I'd hide all this in a matrix helper struct (+ functions) so that you don't have to clutter your code with all the offsets. Consider this in some matrix.h header:
struct matrix; /* opaque type */
/* Allocates a matrix with the given dimensions, sample invocation might be:
*
* struct matrix *m;
* matrix_alloc( &m, -2, 14, -9, 33 );
*/
void matrix_alloc( struct matrix **m, int minRow, int maxRow, int minCol, int maxCol );
/* Releases resources allocated by the given matrix, e.g.:
*
* struct matrix *m;
* ...
* matrix_free( m );
*/
void matrix_free( struct matrix *m );
/* Get/Set the value of some elment in the matrix; takes logicaly (potentially negative)
* coordinates and translates them to zero-based coordinates internally, e.g.:
*
* struct matrix *m;
* ...
* int val = matrix_get( m, 9, -7 );
*/
int matrix_get( struct matrix *m, int row, int col );
void matrix_set( struct matrix *m, int row, int col, int val );
And here's how an implementation might look like (this would be matrix.c):
struct matrix {
int minRow, maxRow, minCol, maxCol;
int **elem;
};
void matrix_alloc( struct matrix **m, int minCol, int maxCol, int minRow, int maxRow ) {
int numRows = maxRow - minRow;
int numCols = maxCol - minCol;
*m = malloc( sizeof( struct matrix ) );
*elem = malloc( numRows * sizeof( *elem ) );
for ( int i = 0; i < numRows; ++i )
*elem = malloc( numCols * sizeof( int ) );
/* setting other fields of the matrix omitted for brevity */
}
void matrix_free( struct matrix *m ) {
/* omitted for brevity */
}
int matrix_get( struct matrix *m, int col, int row ) {
return m->elem[row - m->minRow][col - m->minCol];
}
void matrix_set( struct matrix *m, int col, int row, int val ) {
m->elem[row - m->minRow][col - m->minCol] = val;
}
This way you only need to get this stuff right once, in a central place. The rest of your program doesn't have to deal with raw arrays but rather the struct matrix type.