Cyclical Dependency

[Bryan Bullard]

I have a cyclical dependency problem very much like the one listed below.
Does anyone have a solution?

struct A;
struct B;

struct A
{
struct B b;
};

struct B
{
struct A a;
};

int main() { return 0; }


TIA,
Bryan

 

[Joona I Palaste]

I have a cyclical dependency problem very much like the one listed below.
Does anyone have a solution?

struct A;
struct B;

struct A
{
struct B b;
};

struct B
{
struct A a;
};

int main() { return 0; }

This code is impossible to compile, as you very well know. Both of your
structs contain themselves as elements, which means that their size is
infinite.
The solution is, of course, pointers.

struct A;
struct B;

struct A
{
struct B *b;
};

struct B
{
struct A *a;
}

int main() { return 0; }

Is this your real code or an exercise from a C textbook?

--
/-- Joona Palaste ([email protected]) ------------- Finland --------\
\-- http://www.helsinki.fi/~palaste --------------------- rules! --------/
"I said 'play as you've never played before', not 'play as IF you've never
played before'!"
- Andy Capp

 

[Ben Pfaff]

I have a cyclical dependency problem very much like the one listed below.
Does anyone have a solution?

struct A;
struct B;

struct A
{
struct B b;
};

struct B
{
struct A a;
};

One or both of the structure-type structure members must become a
pointer to a structure.

 

[Martin Ambuhl]

I have a cyclical dependency problem very much like the one listed below.
Does anyone have a solution?

Yes: don't do the impossible. You need an infinite memory to do what
you are attempting.

struct A;
struct B;

struct A
{
struct B b;
};

struct B
{
struct A a;
};

If you are actually trying to do something useful, rather than the
absurd infinite

struct A {
struct B {
struct A {
struct B {
....

use pointers to structs.

 

[Bryan Bullard]

This code is impossible to compile, as you very well know. Both of your
structs contain themselves as elements, which means that their size is
infinite.
The solution is, of course, pointers.

Well, in this instance I need the actual struct resident not pointers.
Here is a closer example:

/* file A.h * * * * * * */
#ifndef A_H
#define A_H

#include "B.h"

struct A
{
struct B b;
};

#endif


/* file B.h * * * * * * */
#ifndef B_H
#define B_H

#include "A.h" /* some other resource in A.h is needed */

struct B
{...};

#endif

/* main.c * * * * * * */
#include "A.h" /* needs resources in A.h and B.h */
#include "B.h"

int main() { return 0; }

This can become a problem when I must always be assured that "B.h" is
preprocessed before "A.h".

Is this your real code or an exercise from a C textbook?

Real code I'm afraid.

 

[Malcolm]

I have a cyclical dependency problem very much like the one listed below.
Does anyone have a solution?

90% of the time the solution is to rework the logic of your progra, so that
cyclic dependencies disappear.
As others have pointed out, structs cannot mutually contain each other (like
a box cannot contain another box containing itself). However if you really
need mutual dependency then you can use pointers.

 

[Bryan Bullard]

90% of the time the solution is to rework the logic of your progra, so that
cyclic dependencies disappear.

This is what I think.

 

[Joona I Palaste]

Well, in this instance I need the actual struct resident not pointers.
Here is a closer example:

/* file A.h * * * * * * */
#ifndef A_H
#define A_H

#include "B.h"

struct A
{
struct B b;
};


/* file B.h * * * * * * */
#ifndef B_H
#define B_H

#include "A.h" /* some other resource in A.h is needed */

struct B
{...};

/* main.c * * * * * * */
#include "A.h" /* needs resources in A.h and B.h */
#include "B.h"

int main() { return 0; }

This can become a problem when I must always be assured that "B.h" is
preprocessed before "A.h".

I don't understand this at all. The code in your closer example here is
completely different from your original code. Your original code was
impossible to compile, even theoretically, as you had structures that
included themselves.
This code, however, has a structure including a structure including
something unknown. As long as that unknown isn't either of these two
structures, your program can be solved with careful attention to
preprocessing.
What are the "resources" you speak of? Structures? Typedefs? Macros?
Actual code? Knowing this is *very* important to determine whether
your problem can be solved.
Do you mean that "A.h" must use types defined in "B.h" and vice versa?
I don't see any way of doing this other than making a separate header
file including all the types that either file needs from the other and
making both #include it.
It's just strange that between your two posts, the entire problem
changed from a struct type problem to a preprocessing problem.

 

[Bryan Bullard]

Do you mean that "A.h" must use types defined in "B.h" and vice versa?
I don't see any way of doing this other than making a separate header
file including all the types that either file needs from the other and
making both #include it.
It's just strange that between your two posts, the entire problem
changed from a struct type problem to a preprocessing problem.

Right. Sorry for the confusion my made by my illustrations but I don't
believe it will be very helpful to get into specifics. Basically, what I
have is a growing code base that is not scaling very well. I believe what I
need to do is to split up my header files into a more granular and hopefully
scalable manner.

"The trouble with the French is they don't have a word for entrepreneur."
- George Bush

Great quote.