"Protoman" <Pr**********@gmail.com> wrote in message
news:11**********************@g49g2000cwa.googlegr oups.com...
| Why can't I write this?: int* array=new int[2]{1,010}; It keeps
saying:
|
| 8 C:\Dev-Cpp\test6.cpp expected `,' or `;' before '{' token
| 9 C:\Dev-Cpp\test6.cpp invalid type argument of `unary *'
|
| Any suggestions? Thanks!!!!
|
Ask yourself the following:
a) What does int* array do exactly?
answer: it *may* hold a pointer to a single integer.
So in the fragment:
int *p = new int(5);
delete p;
p is a pointer to a single integer. Not a container of integers.
b) What does
int array[2] = { 1, 10 };
do exactly?
answer: it creates a container of 2 integers initialized to 1 and 10
respectfully.
c) Since its now clear that a pointer to an integer can't do anything
else but potentially point to an integer, one can use the pointer to
"iterate" through a container of integer elements.
int array[2] = { 1, 10 };
int * p = array[0]; // p points to the first element
++p; // p now points to the second element: array[1]
So you can argue that setting the pointer to the container's first
element will naturally return the base address of the container. Except
that this is a rather dangerous statement since the grand majority of
containers are not organized in sequence in memory. In fact, the only
containers that are organized sequentially are:
primitive arrays
std::vector
a stack
(and thats a very short list of the available containers out there)
Hence, the C++ programmer makes a clear, unambiguous and critical
distinction between a pointer to a given element and a pointer (or
reference) to a given container.
Just cause you hold a pointer to a single car in a parking lot does not
give you the address of the parking lot.
