I've spent quite some time on this issue, but I just can't see what I'm doing wrong. Take a look at this:

struct test {
int array[5];
int value;
};

struct test abc;

Well, guess what: abc.value and abc.array[5] are the same thing - they always have the same address. So if you write to abc.array[5] you automatically overwrite abc.value, and vice versa. But if the array and value aren't inside a structure, this doesn't happen. It's like the array inside the structure has 5 elements, while outside it has 6 elements (which is what I want ). So what is wrong here?

George

Erm, an array of length 5 has entries 0 to 4, no? The memory address after [4] is the next variable's.

You're writing past the end of the array, like Proud said.
So, when you're array are "not inside a structure", and you write past the array, some other variable (or arguably worse, part of the stack) may, or may not get overwritten. If tou're "lucky" you get a "segmentation fault" runtime error...

yes, in C array numbering starts from zero
so
array[4] is the fifth element

(or is it bruce willis?)

Yup the above answers are true, if the memory allocated to your
struct test abc will be something like

memory of abc.array[0]
memory of abc.array[1]
memory of abc.array[2]
memory of abc.array[3]
memory of abc.array[4]
memory of abc.value

And as there are no bound check for array in C/C++ and the address of array element is calculated like.

Address of array + (Size of data type * index number)