On 2 Dec 2007, at 1:54 AM, Adam R. Maxwell wrote: > > On Dec 1, 2007, at 4:41 PM, Christiaan Hofman wrote: > >> On 2 Dec 2007, at 1:28 AM, Adam R. Maxwell wrote: >> >>> If I compile and run the following code, the ivars in the parent and >>> child classes are released. >>> >>> For some reason I was thinking that the subclass' dealloc >>> wouldn't be >>> called. Since it is called, we don't have to do anything if super >>> returns nil...right? I'm fully confused now. >>> >> >> That's correct. In -[Parent init] (called from within -[Child >> init]) , the object is released. Therefore it's -dealloc is called. >> As the object is a Child, this is -[Child dealloc]. So _string is >> released. And it also calls [super dealloc], so also _parentName is >> released. So there never has to be an explicit release of ivars in >> init, cleanup is all taken care of by dealloc (as it should). > > Right. So we can get rid of that release-in-init madness and not > worry about freeing anything twice. > > I think I started out with this idea that the if(self=[super init]) > check was there to avoid leaking ivars, and never bothered to think it > through. I guess if you're going to fire off threads or do something > expensive it's worthwhile to see if super's init failed, but otherwise > it doesn't do much. > >> Anyway, in practice for us this would never occur, as -[NSObject >> init] is empty and never returns nil. > > Unless NSAllocateObject() (and presumably malloc) fails, but by then > you're SOL anyway. > > -- > adam
No, that should have happened already earlier, before any init is started. NSAllocateObject is not called from within init. BTW, I was wondering about a compiler issue: what object do the ivars actually belong to, is it the receiver of the method, or the current value of self? So if I do: receiver = self; self = replacement; myIvar; Is myIvar now equivalent to receiver->myIvar or replacement->myIvar? This would be relevant in this case if super's init could replace the object, because it may make assigning ivare before [super init] unreliable. Christiaan ------------------------------------------------------------------------- SF.Net email is sponsored by: The Future of Linux Business White Paper from Novell. From the desktop to the data center, Linux is going mainstream. Let it simplify your IT future. http://altfarm.mediaplex.com/ad/ck/8857-50307-18918-4 _______________________________________________ Bibdesk-develop mailing list Bibdesk-develop@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bibdesk-develop
