[userbinator]

The array->pointer "decay" (as the standard calls it) has 3 exceptions, of which two are when an array is the operand of "sizeof", and when it is the operand of "&". (The third involves a string initialiser, which is not relevant here.)

So your reasoning is not quite correct, it should really be that you think of

    sizeof( &array[0] )

as being

    sizeof( &(something) )

where "something" could be of any type T, and so the '&' operator yields "pointer-to-T", to which sizeof will yield the size of a pointer.

[bluetomcat]

> and when it is the operand of "&"

Another slightly confusing aspect is that `&array` gives you a "pointer to array" (which has the same value as `&array[0]`, but is of a different type). Most importantly, it behaves differently in pointer arithmetic (the implied offset is the size of the array, rather than the size of its elements).

[dsfuoi]

Fabien is hinting at the confusion: sizeof(&array[0]) != sizeof(array)

&array[0] and array decay to the same thing, the pointer to the first element if they are used in an expression. But sizeof gives a different result, because array+0 'decays' to a pointer, and array doesn't.