maxmspjit

I'm a little unclear on how to write a working compare function for 

linklist_sort. This is what I've got at the moment:

return ((t_foo *)objectA)->property - ((t_foo *)objectB)->property;

Do I need to return A-B, or B-A, or A==B, or ??. Nothing I've tried 

so far seems to sort the list properly, although some choices do seem 

to change the list ordering somewhat (usually between unintended 

I'm a little unclear on how to write a working compare function for  
linklist_sort. This is what I've got at the moment:

long
cmpfn(void *objectA, void *objectB) {
     return ((t_foo *)objectA)->property - ((t_foo *)objectB)->property;
}

Do I need to return A-B, or B-A, or A==B, or ??. Nothing I've tried  
so far seems to sort the list properly, although some choices do seem  
to change the list ordering somewhat (usually between unintended  
pairs of objects).

linklist_sort

On Sep 9, 2007, at 12:22 PM, Ritchie Argue wrote:

> Do I need to return A-B, or B-A, or A==B, or ??. Nothing I've tried 

> so far seems to sort the list properly, although some choices do 

> seem to change the list ordering somewhat (usually between 

> Do I need to return A-B, or B-A, or A==B, or ??. Nothing I've tried  
> so far seems to sort the list properly, although some choices do  
> seem to change the list ordering somewhat (usually between  
> unintended pairs of objects).

On 07-09-09, at 1243, Joshua Kit Clayton wrote:

> On Sep 9, 2007, at 12:22 PM, Ritchie Argue wrote:

>> Do I need to return A-B, or B-A, or A==B, or ??. Nothing I've 

>> tried so far seems to sort the list properly, although some 

>> choices do seem to change the list ordering somewhat (usually 

> cmpfn(x,y) returns 0 if x>y, otherwise 1

On 07-09-09, at 1243, Joshua Kit Clayton wrote:
>
> On Sep 9, 2007, at 12:22 PM, Ritchie Argue wrote:
>
>> Do I need to return A-B, or B-A, or A==B, or ??. Nothing I've  
>> tried so far seems to sort the list properly, although some  
>> choices do seem to change the list ordering somewhat (usually  
>> between unintended pairs of objects).
>
> The way it works currently is:
>
> cmpfn(x,y) returns 0 if x>y, otherwise 1
>
Aha, thanks!