Front end function for qsorting linked lists.. written on my way out the
door, so I wouldn't use it on something that might screw up too bad:
/* merc.h */
#define NEXTPTR(list) ((char *)&(list->next) - (char *)list)
void *list_sort(void *list, int nextpos, int(*compar)(const void *,
const void *));
/* some .c file */
void *list_sort(void *list, int nextpos, int(*compar)(const void *,
const void *))
{
int count = 0, i = 0;
unsigned char *ptr = list;
unsigned char **array;
for (; ptr != NULL; count++)
ptr = (void *)*(long *)(ptr+nextpos);
array = malloc(sizeof(void *)*count);
ptr = list;
for (; ptr != NULL; ptr = (void *)*(long *)(ptr+nextpos))
array[i++] = ptr;
qsort(array, count, sizeof(unsigned char *), compar);
list = NULL;
for (count--;count >= 0; count--)
{
(void *)*(long *)(array[count]+nextpos) = list;
list = array[count];
}
free(array);
return list;
}
/* to use it */
list = list_sort(list, NEXTPTR(list), comparison_function);
--Palrich.
On Wed, 2004-01-28 at 15:43, Chad Simmons wrote:
> --- Valnir <[EMAIL PROTECTED]> wrote:
> > if you look at his original call to Qsort he IS passing it a count.
> >
> > he is passing it "i" which is counted during a for loop prior to the Qsort
> > call.
> >
>
>
> Yes, he's passing a count. What he is NOT passing is an array. Qsort doesn't
> operate on linked lists. If it did you wouldn't need to pass a count, cause it
> would just stop when it got to the NULL record on the end. The point of
> contention has been that (as per the man page) qsort takes the address to the
> first element in an array, the count of elements in the array, and a pointer
> to
> a comparison function.
>
> ~Kender
