Stroustrup 5.4.1, pointers and constants

[arnuld]

int a = 1;
const int c = 2;

now Stroustrup declares pointers for "a" and "c":

const int* p1 = &c; // ok
const int* p2 = &a; // he says it is OK

i don't get it, "a" is not a constant and we are declaring a "pointer
to a constant" . what is this ?

after that, is this legal:

int* p3 = &a;

?

[Alf P. Steinbach]

* arnuld:

int a = 1;
const int c = 2;

now Stroustrup declares pointers for "a" and "c":

const int* p1 = &c; // ok
const int* p2 = &a; // he says it is OK

i don't get it, "a" is not a constant and we are declaring a "pointer
to a constant" . what is this ?

"pointer to a constant" is a restriction on how that pointer can be
used. p2 cannot be used to change a. It does not guarantee that a
isn't changed by other means, it only guarantees that given p2, and
using only p2, you can't change a.

after that, is this legal:

int* p3 = &a;

?

Yes.

--
A: Because it messes up the order in which people normally read text.
Q: Why is it such a bad thing?
A: Top-posting.
Q: What is the most annoying thing on usenet and in e-mail?

[arnuld]

On Mar 28, 12:18 pm, "Alf P. Steinbach" <[e-mail]:

* arnuld:

[SNIP]

"pointer to a constant" is a restriction on how that pointer can be
used. p2 cannot be used to change a. It does not guarantee that a
isn't changed by other means, it only guarantees that given p2, and
using only p2, you can't change a.

then what about this:

const int* p2 = &a;

you said i can not change "a" using p2 but i can using some other
pointer. then

int *const p22 = &a;

this means, "a" will *never* be changed. ... right ?

after that, is this legal:

int* p3 = &a;

?

Yes.

*p3 = 78;

and "a" will get changed to value 78 but i can not use "p3" on:

int *const p22 = &a;

then what does this mean:

const int *const p4 = &a;

?

[Alf P. Steinbach]

* arnuld:

>On Mar 28, 12:18 pm, "Alf P. Steinbach" <[e-mail]:

>* arnuld:

>[SNIP]

>"pointer to a constant" is a restriction on how that pointer can be
used. p2 cannot be used to change a. It does not guarantee that a
isn't changed by other means, it only guarantees that given p2, and
using only p2, you can't change a.

then what about this:

const int* p2 = &a;

you said i can not change "a" using p2 but i can using some other
pointer. then

int *const p22 = &a;

this means, "a" will *never* be changed. ... right ?

No, it means p22 will never be changed.

Reading the declaration backwards, p22 is a constant pointer to int.

Since p22 is constant it cannot be changed; since it's a pointer to
non-constant it can be used to change whatever it points to.

--
A: Because it messes up the order in which people normally read text.
Q: Why is it such a bad thing?
A: Top-posting.
Q: What is the most annoying thing on usenet and in e-mail?

[arnuld]

On Mar 28, 12:48 pm, "Alf P. Steinbach" <[e-mail]:

* arnuld:

int *const p22 = &a;

this means, "a" will *never* be changed. ... right ?

No, it means p22 will never be changed.

Reading the declaration backwards, p22 is a constant pointer to int.

Since p22 is constant it cannot be changed; since it's a pointer to
non-constant it can be used to change whatever it points to.

ok i will try one more time:

int a = 1;

const int* p = &a;

it is a pointer to a constant, whereas neither "p" nor "a" is not a
constant. it only means can not change "a" using "p" but "some other
pointer" can and we can also make "p" to point some where else.
int *const p1 = &a;

"constant pointer" to a variable. we can not change the "p1" to point
to anything else. we can change the value of "a" using "p1" or some
other pointer.

const int *const p2 = &a;

we can not change anything here.

[Juha Nieminen]

Alf P. Steinbach wrote:

it only guarantees that given p2, and
using only p2, you can't change a.

Technically speaking, that's not actually true: const_cast<>

[Alf P. Steinbach]

* arnuld:

>On Mar 28, 12:48 pm, "Alf P. Steinbach" <[e-mail]:

>* arnuld:

>>int *const p22 = &a;
this means, "a" will *never* be changed. ... right ?

No, it means p22 will never be changed.

Reading the declaration backwards, p22 is a constant pointer to int.

Since p22 is constant it cannot be changed; since it's a pointer to
non-constant it can be used to change whatever it points to.

ok i will try one more time:

int a = 1;

const int* p = &a;

it is a pointer to a constant, whereas neither "p" nor "a" is not a
constant. it only means can not change "a" using "p" but "some other
pointer" can and we can also make "p" to point some where else.
int *const p1 = &a;

"constant pointer" to a variable. we can not change the "p1" to point
to anything else. we can change the value of "a" using "p1" or some
other pointer.

const int *const p2 = &a;

we can not change anything here.

Yep.

--
A: Because it messes up the order in which people normally read text.
Q: Why is it such a bad thing?
A: Top-posting.
Q: What is the most annoying thing on usenet and in e-mail?

[Alf P. Steinbach]

* Juha Nieminen:

Alf P. Steinbach wrote:

>it only guarantees that given p2, and
using only p2, you can't change a.

Technically speaking, that's not actually true: const_cast<>

Technically speaking there are at least 101 more ways to circumvent the
type safety.

But don't list them, please.

--
A: Because it messes up the order in which people normally read text.
Q: Why is it such a bad thing?
A: Top-posting.
Q: What is the most annoying thing on usenet and in e-mail?