FROM : j o a r
DATE : Tue Mar 25 23:26:15 2008
On Mar 25, 2008, at 3:19 PM, James Hober wrote:
> On the other hand, the documentation for dealloc says:
>
> "Note that when an application terminates, objects may not be sent a
> dealloc message since the process’s memory is automatically cleared
> on exit—it is more efficient simply to allow the operating system to
> clean up resources than to invoke all the memory management methods."
>
> Since it says "may," theoretically it could be considered an
> implementation detail that dealloc is not called during termination
> and Apple could change things in a future release. Admittedly, this
> is not very likely and admittedly worrying about code that almost
> certainly will never execute is probably not the best use of time.
> And yet it is very common to have a number of singleton objects of
> this kind in Cocoa apps. So nearly all of us are "solving" this one
> way or another.
That's a valid argument, but I feel pretty confident that Apple
wouldn't be able to make such a chance for existing binaries without
creating havoc in the installed application base.
If they do change this, I would expect them to communicate it first,
and then only actually update the behaviour for apps compiled on a
newer version of the OS.
j o a r
DATE : Tue Mar 25 23:26:15 2008
On Mar 25, 2008, at 3:19 PM, James Hober wrote:
> On the other hand, the documentation for dealloc says:
>
> "Note that when an application terminates, objects may not be sent a
> dealloc message since the process’s memory is automatically cleared
> on exit—it is more efficient simply to allow the operating system to
> clean up resources than to invoke all the memory management methods."
>
> Since it says "may," theoretically it could be considered an
> implementation detail that dealloc is not called during termination
> and Apple could change things in a future release. Admittedly, this
> is not very likely and admittedly worrying about code that almost
> certainly will never execute is probably not the best use of time.
> And yet it is very common to have a number of singleton objects of
> this kind in Cocoa apps. So nearly all of us are "solving" this one
> way or another.
That's a valid argument, but I feel pretty confident that Apple
wouldn't be able to make such a chance for existing binaries without
creating havoc in the installed application base.
If they do change this, I would expect them to communicate it first,
and then only actually update the behaviour for apps compiled on a
newer version of the OS.
j o a r
| Related mails | Author | Date |
|---|---|---|
| Samvel | Mar 24, 02:55 | |
| Sherm Pendley | Mar 24, 04:24 | |
| Samvel | Mar 24, 04:36 | |
| Sherm Pendley | Mar 24, 04:42 | |
| Ken Thomases | Mar 24, 21:12 | |
| James Hober | Mar 25, 17:26 | |
| Steve Bird | Mar 25, 17:34 | |
| j o a r | Mar 25, 17:44 | |
| James Hober | Mar 25, 23:19 | |
| j o a r | Mar 25, 23:26 |






Cocoa mail archive

