I see in the documentation - and from a compiler error - that some classes are
not compatible with weak references.
What makes these classes incompatible?
Perhaps I'm using the wrong search terms, but I can't seem to find information
on why this should be the case. Can anyone shed some light on
On Aug 13, 2012, at 8:42 AM, Ben wrote:
> I see in the documentation - and from a compiler error - that some classes
> are not compatible with weak references.
>
> What makes these classes incompatible?
They have custom implementations of -retain and -release.
When NextSTEP was first released
Le 13 août 2012 à 17:56, Kyle Sluder a écrit :
> On Aug 13, 2012, at 8:42 AM, Ben wrote:
>
>> I see in the documentation - and from a compiler error - that some classes
>> are not compatible with weak references.
>>
>> What makes these classes incompatible?
>
> They have custom implementati
On Aug 13, 2012, at 9:29 AM, Jean-Daniel Dupas wrote:
> Le 13 août 2012 à 17:56, Kyle Sluder a écrit :
>> On Aug 13, 2012, at 8:42 AM, Ben wrote:
>>
>>> I see in the documentation - and from a compiler error - that some classes
>>> are not compatible with weak references.
>>>
>>> What makes th
Le 13 août 2012 à 19:54, John McCall a écrit :
> On Aug 13, 2012, at 9:29 AM, Jean-Daniel Dupas wrote:
>> Le 13 août 2012 à 17:56, Kyle Sluder a écrit :
>>> On Aug 13, 2012, at 8:42 AM, Ben wrote:
>>>
I see in the documentation - and from a compiler error - that some classes
are no
On Mon, Aug 13, 2012, at 11:30 AM, Jean-Daniel Dupas wrote:
>
> Thanks for the details. I know what the fragile ABI, and the modern ABI
> are but where I don't agree is when Kyle says that it was not possible to
> add a refcount ivar when retain/release was introduced. As retain/release
> was alre
On 13 Aug 2012, at 19:30, Jean-Daniel Dupas wrote:
>
> Le 13 août 2012 à 19:54, John McCall a écrit :
>
>> On Aug 13, 2012, at 9:29 AM, Jean-Daniel Dupas wrote:
>>> Le 13 août 2012 à 17:56, Kyle Sluder a écrit :
On Aug 13, 2012, at 8:42 AM, Ben wrote:
> I see in the documenta
On Mon, Aug 13, 2012, at 02:47 PM, Mike Abdullah wrote:
> An idea I've vaguely wondered about would be turning the isa variable
> into a tagged pointer. If you know nothing is accessing it directly (to
> do so was deprecated with the modern runtime), then, say, the last 4 bits
> of the pointer coul
On Aug 13, 2012, at 9:29 AM, Jean-Daniel Dupas wrote:
>
> Le 13 août 2012 à 17:56, Kyle Sluder a écrit :
>
>> On Aug 13, 2012, at 8:42 AM, Ben wrote:
>>
>>> I see in the documentation - and from a compiler error - that some classes
>>> are not compatible with weak references.
>>>
>>> What
Le 13 août 2012, à 23:47, Mike Abdullah scripsit:
> An idea I've vaguely wondered about would be turning the isa variable into a
> tagged pointer. If you know nothing is accessing it directly (to do so was
> deprecated with the modern runtime), then, say, the last 4 bits of the
> pointer could
On 13 Aug 2012, at 16:56, Kyle Sluder wrote:
> On Aug 13, 2012, at 8:42 AM, Ben wrote:
>
>> I see in the documentation - and from a compiler error - that some classes
>> are not compatible with weak references.
>>
>> What makes these classes incompatible?
>
> They have custom implementation
Ahh the Atari ST… Now those are fond memories indeed…
-Stevo Brock
Sunset Magicwerks, LLC
www.sunsetmagicwerks.com
818-609-0258
On Aug 14, 2012, at 12:09 AM, Vincent Habchi wrote:
> Le 13 août 2012, à 23:47, Mike Abdullah scripsit:
>
>> An idea I've vaguely wondered about would be turni
On Tue, Aug 14, 2012, at 12:09 AM, Vincent Habchi wrote:
> Using part of pointer to do something else than record an address is
> perilous.
NSNumber does exactly that on 10.7 and higher. Tagged pointers are a
well-understood technique.
You just have to be careful not to dereference them directly.
On Aug 14, 2012, at 2:09 AM, Vincent Habchi wrote:
> Le 13 août 2012, à 23:47, Mike Abdullah scripsit:
>
>> An idea I've vaguely wondered about would be turning the isa variable into a
>> tagged pointer. If you know nothing is accessing it directly (to do so was
>> deprecated with the modern
That exists already in Lion with NSNumber objects: some types of pointers to
NSNumbers actually use the low bits of the pointer to indicate the type and the
rest of the pointer to pack the value; resulting in not needing to actually
allocate an object off the heap (which is why direct isa access
Le 14 août 2012 à 09:09, Vincent Habchi a écrit :
> Le 13 août 2012, à 23:47, Mike Abdullah scripsit:
>
>> An idea I've vaguely wondered about would be turning the isa variable into a
>> tagged pointer. If you know nothing is accessing it directly (to do so was
>> deprecated with the modern
On 16 août 2012, at 03:11, Britt Durbrow
wrote:
> objects must be 16-byte aligned - and that alignment requirement isn't likely
> to go down, but rather up in the future; using the low bits doesn't really
> cause the same issue that using the high bits did.
Well, I partly disagree. There is n
On Aug 16, 2012, at 4:37 AM, Vincent Habchi wrote:
> On 16 août 2012, at 03:11, Britt Durbrow
> wrote:
>
>> objects must be 16-byte aligned - and that alignment requirement isn't
>> likely to go down, but rather up in the future; using the low bits doesn't
>> really cause the same issue that
18 matches
Mail list logo
