Question

I have a some pieces of code in my project in which a structure z is being used. This structure declaration changes as per a condition flag being set in my system. I want the structure declaration to be dynamic.

if condition flag is 0 then,

struct x z;

if condition flag is 1 then,

struct y z;

but the operation to be performed on z would be the same. Currently this is handled by declaring both the structure :

struct x z_x;
struct y z_y;

then as per the condition flag use the appropriate structure in the code. This method doesn’t seem to be optimal. Any suggestions on handling this type of code ?

Answer 1

As the flag is run-time, you cannot avoid checking this flag all the time. At least, I can’t think of such way.

But you still can reduce the memory and the code. I’d do it like this:

---

To reduce the memory usage:

struct z_x* = NULL;
struct z_y* = NULL;

//...

if( flag )
{
    z_x = malloc( sizeof( struct x ) );
}
else
{
    z_y = malloc( sizeof( struct y ) );
}

---

To reduce the code, instead of writing all the time things like:

if( flag )
{
    // do operations_N with z_x
}
else
{
    // do operations_N with z_y
}

I’d write:

operations_N( z_x, z_y, flag );

where

void operations_N( struct x* z_x, struct y* z_y, flag )
{
    if( flag )
    {
        // do operations_N with z_x
    }
    else
    {
        // do operations_N with z_y
    }
}

of course, operations_1 and flag should have some good and logical names.

BUT this could lead do some problems, so just be careful, if you decide to use this approach.