Struct manipulation needs some explanation - c

Question

I have the following piece of code:
//length, width, height
typedef struct {
float L, W, H;
} TDim;
//contains details for each geometrical form
typedef struct {
char *id; // each geometrical form has a unique id
float volume;
TDim *dim;
} TForm;
TForm* FormAlloc()
{
TForm *F = (TForm*) malloc(MAX * sizeof(TForm));
F->id = (char*) malloc(MAX * sizeof(char));
F->dim = (TDim*) malloc(MAX * sizeof(TDim));
return F;
}
which is a good exercise of manipulating variables and pointers within more structures.
What I want to do, is to store the id (basically the letter 'F' followed by 1,2,3...), length, width and height of some rectangular cuboid in order to compute its volume.
I need some explanation around the usage of -> vs .
I understand that -> works with addresses and . with members.
In int main(), I have:
F[0].id = "F1"; //the first cuboid has the id "F1"
F[0].dim[0].L = 1; //the first cuboid has length = 11
//could have used F[0].dim->L = 1; as well?
F[0].dim[0].W = 2;
F[0].dim[0].H = 3;
F[1].id = "F2"; //the second cuboid has the id "F2"
F[1].dim[1].L = 4; //this is where it breaks down - SEG_FAULT (I've used gdb to catch it)
What causes the SEG_FAULT in that line?

Answer 1

In this function:
TForm* FormAlloc()
{
TForm *F = (TForm*) malloc(MAX * sizeof(TForm));
F->id = (char*) malloc(MAX * sizeof(char));
F->dim = (TDim*) malloc(MAX * sizeof(TDim));
return F;
}
you create a dynamic array of MAX elements of TForm, but then you only allocate memory for the id and dim members of the first, hence the segfault when you try to store memory in the second element.
You'll need something like:
TForm* FormAlloc(void)
{
TForm *F = malloc(MAX * sizeof(TForm));
if ( !F ) {
perror("couldn't allocate memory");
exit(EXIT_FAILURE);
}
for ( size_t i = 0; i < MAX; ++i ) {
if ( (F[i].id = malloc(MAX)) == NULL ) {
perror("couldn't allocate memory");
exit(EXIT_FAILURE);
}
if ( (F[i].dim = malloc(MAX * sizeof(TDim))) == NULL ) {
perror("couldn't allocate memory");
exit(EXIT_FAILURE);
}
}
return F;
}
Most likely, you don't want to use MAX for all three purposes, here, i.e. the chances you need exactly the same number of elements of TForm as you need characters in each id string is probably low.
Other more minor points:
malloc() can return NULL, and you should check for that.
In the code in your question, F[0].id = "F1"; should be strcpy(F[0].id, "F1"); or similar, since otherwise you're going to lose your reference to the memory you malloc()ed.
You don't need to, and really shouldn't, cast the return from malloc() in C.
sizeof(char) is always 1 by definition, so you can leave it out.
For the question in your comment:
What about F[0].dim[0].L = 1; vs F[0].dim->L = 1;?
the two are equivalent, here. Using the subscript operator, dim[0], automatically dereferences dim for you, so you'd use the . operator to get at the elements (dim[n] is equivalent to *(dim + n) in C). Without the [0], dim is a straight pointer, so you'd use the -> operator. Obviously the first form is much more convenient when you want to access the other elements in the array. For the others, you'd have to replace F[0].dim[1].L = 1; with (F[0].dim + 1)->L = 1; or similar, to use the -> operator, which is unnecessarily convoluted.