I'm not quite sure I follow the question, but...

If I wanted to do the same thing with a structure, I would pass a=20
pointer to the stucture (even if it were simply a byte array) into a=20
function that takes a void * pointer as a parameter.  I would then cast=20
the pointer within the function.

typedef union tagFoonion {
     unsigned char bytes[2];
     unsigned int word_u16;
} Foonion;

     char theBytes[2];
     Foonion gUnion;
     gUnion.bytes[0] =3D 0xbe;
     gUnion bytes[1] =3D 0xba;
     memcpy(theBytes, &gUnion, sizeof(Foonion));

     useUnion(theBytes);    // pass bytes as union

     ...

     void useUnion(void * anArg) {
         Foonion *pUnion =3D (Foonion *) anArg;
         Foonion myUnion;
         memcpy(myUnion, pUnion, sizeof(Foonion));    // a local copy of
         // do stuff with myUnion, or even pUnion (which is gUnion)
     }

     There're probably some errors with casting or specifying pointers,=20
but it's my expression mof a general technique.

Joe W





On 01/16/2013 07:01 PM, Bob Blick wrote:
> Is this a correct way to declare and call a function that uses a union
> as the argument, and pass it a value that is not declared as union?
>
> //declaring the function
> void myDumbFunction(union {unsigned int uint; signed int ssint; unsigned
> char[2] unchar});
>
> // calling it from somewhere in my program
>    unsigned int dumbvar;
>    dumbvar =3D somedumbvalue;
>    myDumbFunction(dumbvar);
>
> // the function
> void myDumbFunction(somevar){
>    if (somevar.unchar[1])
>      do_something();
> }
>
> Is that going to work, the compiler should know that somevar is an
> unsigned int(not that it should matter since it's just memory space that
> is the size of the union), so within my function I can access
> somevar.uint somevar.ssint as well as somevar.unchar[0] and
> somevar.unchar[1]
>
> That all seem OK? I know that in most cases using a union serves no real
> purpose other than to be dangerous, there are other ways to do things.
> But my question is whether this is accurate or how it would be made
> correct if it isn't.
>
> Thanks,
>
> Bob
>


--=20
Joe Wronski
Stillwater Embedded Engineering
www.stillwatereng.net

--=20
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
.