Frederick Gotham wrote:

> This is forbidden by the C++ Standard. Every object must have a unique
> address.

What do you mean by "address" in this statement?
I wasn't aware that any language standard used the term "address."

Do you mean by that "pointer value"?  (Which can only be compared
for identical types)

Or pointer bit pattern (what a binary dump would show)?

Or value after casting the pointer value to void *?

Or value after some sort of reinterpret_cast<> of the pointer value?

Or are you talking about something unrelated to pointer values?

Note that casts may change the bit pattern of pointer values,
and that the contents of a pointer may include more than a hardware
address, for some reasonable interpretation of the term "hardware
address", if there is any reasonable interpretation of the
term on the architecture you're interested in.

-- Alan McKenney

---
[ comp.std.c++ is moderated.  To submit articles, try just posting with ]
[ your news-reader.  If that fails, use mailto:std-c++@ncar.ucar.edu    ]
[              --- Please see the FAQ before posting. ---               ]
[ FAQ: http://www.comeaucomputing.com/csc/faq.html                      ]

"Alan McKenney" <alan_mckenney1@yahoo.com> writes:
> Frederick Gotham wrote:
>
>> This is forbidden by the C++ Standard. Every object must have a unique
>> address.
>
> What do you mean by "address" in this statement?
> I wasn't aware that any language standard used the term "address."

The C and C++ standards both use the term "address" extensively.  For
example, unary "&" is the address-of operator.  I don't see a formal
definition for the term in either standard, but it seems to be
synomymous with "pointer value".

--
Keith Thompson (The_Other_Keith) kst-u@mib.org  <http://www.ghoti.net/~kst>
San Diego Supercomputer Center             <*>  <http://users.sdsc.edu/~kst>
We must do something.  This is something.  Therefore, we must do this.

---
[ comp.std.c++ is moderated.  To submit articles, try just posting with ]
[ your news-reader.  If that fails, use mailto:std-c++@ncar.ucar.edu    ]
[              --- Please see the FAQ before posting. ---               ]
[ FAQ: http://www.comeaucomputing.com/csc/faq.html                      ]

Alan McKenney:

>> This is forbidden by the C++ Standard. Every object must have a unique
>> address.
>
> What do you mean by "address" in this statement?


    Pointer value.


> I wasn't aware that any language standard used the term "address."


    "pointer value" will have to do then.


> Do you mean by that "pointer value"?  (Which can only be compared
> for identical types)


    A "pointer value" is an address. It's an identifier which indicates
where something is located. An address can take several forms:

    0x2987D4F422
    17 Greenlawns Avenue
    The 2nd left past the garden centre.
    554, byte position 3.


> Or pointer bit pattern (what a binary dump would show)?


    Presumably, a pointer bit-pattern would be the result of memcpy'ing a
pointer.

    int *p = 0;
    char unsigned buf[sizeof p];
    memcpy(buf,&p,sizeof p);


> Or value after casting the pointer value to void *?


    If you cast (or implicitly convert) an object pointer to a void*, then
you have the address in the form of a byte address. This may be identical
to the object pointer, e.g. a value something like:

    3536

, or it might be something more complex like:

    3536, byte position 2


> Or value after some sort of reinterpret_cast<> of the pointer value?


    If the casted-to pointer type describes an object type whose alignment
requirements are no stricter than the casted-from type, then you simply
have the address stored in a different pointer type. The addresses may be
represented identically, or they may be different insofar as the following
two addresses are different:

    17 Greenlawns Avenue

and:

    17 Ascaill na nGlasfhaichi

   These addresses above are represented differently (in this case,
different spoken languages rather than different bit-patterns), but they're
still the same address. If we perform an equality test on these two
addresses, we'd expect it to be true:

    (void*)(17 Greenlawns Avenue) == (void*)(17 Ascaill na nGlasfhaichi)


> Or are you talking about something unrelated to pointer values?


No.


> Note that casts may change the bit pattern of pointer values,
> and that the contents of a pointer may include more than a hardware
> address, for some reasonable interpretation of the term "hardware
> address", if there is any reasonable interpretation of the
> term on the architecture you're interested in.


If an implementation so desires, a pointer can hold a menagarie of
information, such as bounds-checking information.

--

Frederick Gotham

---
[ comp.std.c++ is moderated.  To submit articles, try just posting with ]
[ your news-reader.  If that fails, use mailto:std-c++@ncar.ucar.edu    ]
[              --- Please see the FAQ before posting. ---               ]
[ FAQ: http://www.comeaucomputing.com/csc/faq.html                      ]