Am Di,05.08.2008 um 12:00 schrieb mmalc crawford:
On Aug 5, 2008, at 2:48 AM, Negm-Awad Amin wrote:
This is the "mirror" of the problem, when you initialize an object.
Of course, theoretically in both cases the usage of setters are
dangerous. In most cases the deallocation of the object in reverse
order to its initialization will not lead to any problems. It is no
problem, when you have synthisized accessors, because they have
noch side-effects.
No, this is not the case -- Ken's reply was correct. If you use
automatic KVO notifications, your accessors will still have side
effects.
Again per Ken's reply, you should typically set instance variables
directly in init methods, and release them directly in dealloc.
No, I think, that you should never, never have an observation, when -
dealloc is running. *This* is the problem, not the KVO-Messages. If
you are in -dealloc, the instance *will* be deleted. Since this
happens running up the class hierarchy, the observing object has no
chance to remove its observation before the object is deallocated. So
in the moment of -dealloc (NSObject) you have an observation to an
"dead object". *This* is the fault.
You always have to remove the observation before there is the chance
of losing the object.
mmalc
Amin Negm-Awad
[EMAIL PROTECTED]
_______________________________________________
Cocoa-dev mailing list (Cocoa-dev@lists.apple.com)
Please do not post admin requests or moderator comments to the list.
Contact the moderators at cocoa-dev-admins(at)lists.apple.com
Help/Unsubscribe/Update your Subscription:
http://lists.apple.com/mailman/options/cocoa-dev/archive%40mail-archive.com
This email sent to [EMAIL PROTECTED]
