🚀 go-pugleaf

In article <1992Aug10.204928.14917@frog.CRDS.COM> miket@frog.CRDS.COM (vulture) writes:
>According to my current book which may be outdated (if so, please offer
>better books to use)

For questions involving a standard, a better book would be the standard
itself.  If you need a tutorial (textbook) first, then the standard isn't
the best.

>it says that the prototype 'void *' was introduced
>to ANSI C "to represent a 'universal' data pointer."

Any type of pointer-to-data can be converted to void* and back without
problems.  However, if you try to hide the fact that conversion might
be occuring, then you can still get in trouble for it.

>		typedef struct prot1 {
>			int (*mycall)(char *parm1, void *parm_ptr );
>			} PRO1;

The function pointed to by a "mycall" component must take a second parameter
of type void*.  At the time of calling, you can put any kind of pointer in
the second parameter, and it will get converted to void*.

>		static int myfunc1 ( char *myparm1, PRO2 *ptr1 ) {}

myfunc1() doesn't take a second parameter of type void*.
If you assign myfunc1 to the "mycall" component of some struct, then you
should expect a type mismatch because the parameter types are different.

WARNING:  If you shut up the compiler by putting a cast in front of myfunc1,
so that you assign a compatible pointer-to-function type, then you should
expect random failures during execution (on some machines), because PRO2*
still isn't the same as void*, and you're preventing the conversion.

It's probably easiest to change myfunc1(), e.g.
		static int myfunc1 ( char *myparm1, void *ptr0 ) {
			PRO2 *ptr1 = ptr0;
			/* ... */
		}

>			/* PRO2 not the same as PRO1 */

Irrelevant.
--
Norman Diamond       diamond@jit081.enet.dec.com
If this were the company's opinion, I wouldn't be allowed to post it.
"Yeah -- bad wiring.  That was probably it.  Very bad."