With the array decaying into a pointer you’re losing one important piece of information: its size.
With an array reference your range based loop works:
void foo(int (&bar)[3]);
int main() {
int bar[3] = {1,2,3};
for (int i : bar) {
cout << i << endl;
}
foo(bar);
}
void foo(int (&bar)[3]) {
for (int i : bar) {
cout << i << endl;
}
}
or, in a generic fashion (i.e. without specifying the array size in the function signature),
template <std::size_t array_size>
void foo(int (&bar)[array_size]) {
for (int i : bar) {
cout << i << endl;
}
}