Hi,

I want to write an array of pointer to function "which takes an array
of function pointer of type "int fun(int)" and returns a fptr of same
type

The following works well

typedef int (*fp)(int);
fp (*newFp[5]) (fp [5]);

But when I try to inline the typedefs then it gives me compilation
error in VC++ 2010 saying " syntax error : ')'"

int (*)(int)  (*newFp[5])  ( int (*[5]) (int) )

How do I inline the declarations to remove the typedef (its not
required but am asking just out of curiosity)

Thanks in advance
Rahul

* Rahul, on 05.10.2010 11:38:
>
> I want to write an array of pointer to function "which takes an array
> of function pointer of type "int fun(int)" and returns a fptr of same
> type
>
> The following works well
>
> typedef int (*fp)(int);
> fp (*newFp[5]) (fp [5]);
>
> But when I try to inline the typedefs then it gives me compilation
> error in VC++ 2010 saying " syntax error : ')'"
>
> int (*)(int)  (*newFp[5])  ( int (*[5]) (int) )
>
> How do I inline the declarations to remove the typedef (its not
> required but am asking just out of curiosity)

Well it sounds like homework, but it's often asked and seldom answered.

Starting at the innermost thing, the name of the array:

   newFp

What you can do with the array is to index it:

   newFp[]

The result of that is a function pointer, which can be dereferenced:

   *newFp[]

Yielding an untyped entity (corresponding to a delegate in other languages) that 
can be called:

   (*newFp[])(...)

The result of calling is a function pointer that can be dereferenced:

   *(*newFp[])(...)

And called with int argument:

   (*(*newFp[])(...))(int)

Yielding an int:

   int (*(*newFp[])(...))(int)

The dots formal argument should be an array, and such array can be indexed:

   int (*(*newFp[])(...[]))(int)

The result of indexing is a function pointer that can be dereferenced:

   int (*(*newFp[])(...*...[]))(int)

And called with int argument:

   int (*(*newFp[])(...(*[])(int) ))(int)

Yielding an int:

   int (*(*newFp[])(int (*[])(int)))(int)

Given this it's understandable why both Brian Kernighan (I think it was) and 
Bjarne Stroustrup have described the C declaration syntax as a "failed 
experiment". :-)

In short, use typedef.


Cheers & hth.,

- Alf

-- 
blog at <url: http://alfps.wordpress.com>

On Oct 5, 5:00=A0pm, "Alf P. Steinbach /Usenet" <alf.p.steinbach
+use...@gmail.com> wrote:
> * Rahul, on 05.10.2010 11:38:
>
>
>
>
>
>
>
> > I want to write an array of pointer to function "which takes an array
> > of function pointer of type "int fun(int)" and returns a fptr of same
> > type
>
> > The following works well
>
> > typedef int (*fp)(int);
> > fp (*newFp[5]) (fp [5]);
>
> > But when I try to inline the typedefs then it gives me compilation
> > error in VC++ 2010 saying " syntax error : ')'"
>
> > int (*)(int) =A0(*newFp[5]) =A0( int (*[5]) (int) )
>
> > How do I inline the declarations to remove the typedef (its not
> > required but am asking just out of curiosity)
>
> Well it sounds like homework, but it's often asked and seldom answered.
>
> Starting at the innermost thing, the name of the array:
>
> =A0 =A0newFp
>
> What you can do with the array is to index it:
>
> =A0 =A0newFp[]
>
> The result of that is a function pointer, which can be dereferenced:
>
> =A0 =A0*newFp[]
>
> Yielding an untyped entity (corresponding to a delegate in other language=
s) that
> can be called:
>
> =A0 =A0(*newFp[])(...)
>
> The result of calling is a function pointer that can be dereferenced:
>
> =A0 =A0*(*newFp[])(...)
>
> And called with int argument:
>
> =A0 =A0(*(*newFp[])(...))(int)
>
> Yielding an int:
>
> =A0 =A0int (*(*newFp[])(...))(int)
>
> The dots formal argument should be an array, and such array can be indexe=
d:
>
> =A0 =A0int (*(*newFp[])(...[]))(int)
>
> The result of indexing is a function pointer that can be dereferenced:
>
> =A0 =A0int (*(*newFp[])(...*...[]))(int)
>
> And called with int argument:
>
> =A0 =A0int (*(*newFp[])(...(*[])(int) ))(int)
>
> Yielding an int:
>
> =A0 =A0int (*(*newFp[])(int (*[])(int)))(int)
>
> Given this it's understandable why both Brian Kernighan (I think it was) =
and
> Bjarne Stroustrup have described the C declaration syntax as a "failed
> experiment". :-)
>
> In short, use typedef.
>
> Cheers & hth.,
>
> - Alf
>
> --
> blog at <url:http://alfps.wordpress.com>

Thanks Alf!!

I got the funda.

On Oct 5, 8:00=A0am, "Alf P. Steinbach /Usenet"
[Alf explains a complicated declaration]

Whew! Yes indeed, thanks Alf. That was a ride.
Socks