«

I have ran into a problem where I have a struct
that has a member which contains a pointer
to function and is initialized to a function
in the initializer list. With my array of
structs of this type, I have some elements
of this array, thats function pointer member does
not need to be initialized to a function.
I can't simply initialize it to NULL
and I'm not sure if casting NULL(which can
be 0 or (void *)0) to the function pointer type
will work.

If NULL is (void *)0 on some implementations
you would end up with:

(int (*)(unsigned int))(void *)0

is something like that even guaranteed
to work?

I'm not even sure if just plain 0 and casting
it to the function pointer type would work.
Since function pointers might have different
representations than object pointers, it seems
you would end up with implementation-defined
behavior either way.

Has anyone ran into this problem before?
Where portability was a concern to you
and you couldn't rely on implementation-
defined behavior.



--
nethlek

In article < >,
(Mantorok Redgormor) wrote:

> I have ran into a problem where I have a struct
> that has a member which contains a pointer
> to function and is initialized to a function
> in the initializer list. With my array of
> structs of this type, I have some elements
> of this array, thats function pointer member does
> not need to be initialized to a function.
> I can't simply initialize it to NULL
> and I'm not sure if casting NULL(which can
> be 0 or (void *)0) to the function pointer type
> will work.
>
> If NULL is (void *)0 on some implementations
> you would end up with:
>
> (int (*)(unsigned int))(void *)0
>
> is something like that even guaranteed
> to work?

NULL is a null pointer constant. Any null pointer constant can be cast
to any pointer type and the result will be a null pointer of the
corresponding type. That is true both for data pointers and function
pointers. It is guaranteed to work.

> I'm not even sure if just plain 0 and casting
> it to the function pointer type would work.

A plain zero is also a null pointer constant and can be cast to a
function pointer, producing a null pointer of the right type.

> Since function pointers might have different
> representations than object pointers, it seems
> you would end up with implementation-defined
> behavior either way.

No, it is defined by the C Standard.

Mantorok Redgormor wrote:
> I have ran into a problem where I have a struct that has a member
> which contains a pointer to function and is initialized to a
> function in the initializer list. With my array of structs of this
> type, I have some elements of this array, thats function pointer
> member does not need to be initialized to a function. I can't
> simply initialize it to NULL and I'm not sure if casting NULL(which
> can be 0 or (void *)0) to the function pointer type will work.

Just initialize it to an uncast null pointer constant (e.g. NULL or
0).

#include <math.h>
#include <stddef.h>

struct {
double (*one)(double),
(*two)(double);
} functions[] = { {sin, NULL}, {NULL, cos} };

Null pointer constants are special in C: assigning a null pointer
constant to a function pointer type is allowed, and has the desired
behaviour.

> If NULL is (void *)0 on some implementations
> you would end up with:
>
> (int (*)(unsigned int))(void *)0
>
> is something like that even guaranteed
> to work?

It's guaranteed to work

> I'm not even sure if just plain 0 and casting
> it to the function pointer type would work.

There's no need to cast at all.

> Since function pointers might have different
> representations than object pointers, it seems
> you would end up with implementation-defined
> behavior either way.

Casts deal with values, not representations.

Jeremy.

Christian Bau <> wrote in message news:<christian.bau->...
> In article < >,
> (Mantorok Redgormor) wrote:
>
> > I have ran into a problem where I have a struct
> > that has a member which contains a pointer
> > to function and is initialized to a function
> > in the initializer list. With my array of
> > structs of this type, I have some elements
> > of this array, thats function pointer member does
> > not need to be initialized to a function.
> > I can't simply initialize it to NULL
> > and I'm not sure if casting NULL(which can
> > be 0 or (void *)0) to the function pointer type
> > will work.
> >
> > If NULL is (void *)0 on some implementations
> > you would end up with:
> >
> > (int (*)(unsigned int))(void *)0
> >
> > is something like that even guaranteed
> > to work?
>
> NULL is a null pointer constant. Any null pointer constant can be cast
> to any pointer type and the result will be a null pointer of the
> corresponding type. That is true both for data pointers and function
> pointers. It is guaranteed to work.
>
> > I'm not even sure if just plain 0 and casting
> > it to the function pointer type would work.
>
> A plain zero is also a null pointer constant and can be cast to a
> function pointer, producing a null pointer of the right type.
>
> > Since function pointers might have different
> > representations than object pointers, it seems
> > you would end up with implementation-defined
> > behavior either way.
>
> No, it is defined by the C Standard.

what section?



--
nethlek

On 5 Apr 2004 06:18:31 -0700, (Mantorok Redgormor) wrote:


>>
>> A plain zero is also a null pointer constant and can be cast to a
>> function pointer, producing a null pointer of the right type.
>>
>> > Since function pointers might have different
>> > representations than object pointers, it seems
>> > you would end up with implementation-defined
>> > behavior either way.
>>
>> No, it is defined by the C Standard.
>
>what section?

6.3.2.3/3 and 4:

3 An integer constant expression with the value 0, or such an expression
cast to type void *, is called a null pointer constant.55) If a null
pointer constant is converted to a pointer type, the resulting pointer,
called a null pointer, is guaranteed to compare unequal to a pointer to any
object or function.

4 Conversion of a null pointer to another pointer type yields a null
pointer of that type. Any two null pointers shall compare equal.

-leor

--
Leor Zolman --- BD Software --- www.bdsoft.com
On-Site Training in C/C++, Java, Perl and Unix
C++ users: Download BD Software's free STL Error Message Decryptor at:
www.bdsoft.com/tools/stlfilt.html