Pointer math vs. Array index

Question

I know this has been hashed over a number of time, but I have come across a case today that shook my understanding of the pointer math/ array index.

As I have allways understood it, &mybuff[10] and (&mybuff+10) are equivilent ways of referancing the same memory.

However I spent the morning fighting a case where:

memcpy(&mybuff+10,&in,8); 

overflowed the buffer, when compiled with optimization on and worked just fine when compiled for debugging.

While athe same time,

memcpy(&mybuff[10],&in,8); 

worked just fine in both cases.

Many thanks for any ideas or pointers.

Answer 1

&mybuff[10] is equivalent to &mybuff[0] + 10 which is equivalent to mybuff + 10

Array indexing is defined in terms of pointer arithmetic. p[i] means *(p+i) (I'm ignoring the need for extra parentheses in case p or i is a more complex expression), where p is a pointer value and i is an integer value.

Trivia: Since addition, even pointer+integer addition, is commutative, p[i] can also be written as i[p]. Yes, 4["Hello"] == 'o'. For the sake of anyone who reads your code in the future, please do not make use of this knowledge.

An excellent reference on the relationship between arrays and pointer in C (and C++, where the rules are nearly identical) is section 6 of the comp.lang.c FAQ. (I could have linked directly to section 6, but I like to encourage people to browse it; the whole thing is worth reading.)