On 28 Apr 2008, at 17:42, Richard Frith-Macdonald wrote:
>
> On 28 Apr 2008, at 17:09, hns computer.org wrote:
>
>>> See... this is gnustep, so while you could do
that, it would best
>>> be done as:
>>>
>>> RELEASE(willy);
>>
>> Well, I never understood completely why is it
better (or best) to
>> replace some standard Obj-C method call syntax by
standard C function
>> call syntax that breaks compilation on OSX without
adding some
>> defines.
>>
>> I know that it allows to remove/undef all RELEASE()
calls if we
>> have a
>> working GC runtime.
>
> For what it's worth, I no longer think that's a good
reason for
> using RELEASE() ... processor speeds have improved so
much over the
> years that the performance overhead of calls to
-release is really
> not important.
Quick and unscientific testing with the debugger on the
garbage-
collected variant of the NeXT runtime shows that -[NSObject
release]
doesn't even get called in the GC environment, which makes
"the
performance overhead of calls to -release" entirely
unimportant
Graham.
_______________________________________________
Discuss-gnustep mailing list
Discuss-gnustep gnu.org
http://lists.gnu.org/mailman/listinfo/discuss-gnustep
|