[cfe-commits] r160839 - in /cfe/trunk: lib/Sema/SemaExprObjC.cpp test/ARCMT/checking.m test/ARCMT/no-canceling-bridge-to-bridge-cast.m test/ARCMT/nonobjc-to-objc-cast-2.m test/SemaObjC/arc-bridged-cast.m test/SemaObjC/arc-cf.m test/SemaObjC/arc-type-conversion.m test/SemaObjC/arc-unbridged-cast.m test/SemaObjC/arc.m test/SemaObjCXX/arc-unbridged-cast.mm

[Jordan Rose]

On Jul 27, 2012, at 9:50 , jahanian <fjahanian at apple.com> wrote:

> 
> On Jul 27, 2012, at 9:35 AM, Jordan Rose wrote:
> 
>> This commit was bothering me for a while, and I think I know why:
>> 
>> - (id)getSpecialName:(CFDictionaryRef)dict {
>> return [(NSDictionary *)dict objectForKey:MySpecialKey];
>> }
>> 
>> - (voidd)extractSpecialName:(CFDictionaryRef)dict {
>> _ivar = (id)CFDictionaryGetValue(dict, MySpecialCFKey);
>> }
>> 
>> Using CFBridgingRelease here would be very very wrong, in either case.
> 
> This commit did not change the message on CFBridgingRelease.  If it is wrong, it was wrong before this
> patch. Please file a new radar and provide the reasoning.

The commit removed the suggestion of __bridge, which would be the only correct thing to do here.