[cfe-commits] [PATCH] ObjC: interface decls pointing to implementation decls

[Fariborz Jahanian]

On Jul 20, 2009, at 1:00 PM, Argyrios Kyrtzidis wrote:

>
> On Jul 20, 2009, at 12:36 PM, Fariborz Jahanian wrote:
>
>>
>> On Jul 20, 2009, at 12:01 PM, steve naroff wrote:
>>
>>>
>>> On Jul 20, 2009, at 2:45 PM, Argyrios Kyrtzidis wrote:
>>>
>>>>
>>>> On Jul 20, 2009, at 11:34 AM, steve naroff wrote:
>>>>
>>>>>
>>>>> Looks great.
>>>>
>>>> Ok, unless someone objects (soon), I'm going to commit the set of
>>>> patches.
>>>>
>>>
>>> From my perspective, you are good to go.
>>
>> Since I raised a question related to addition of the data member, I  
>> should say something here :).
>> From my perspective benefits outweighs the drawbacks. In the past,  
>> we used coca.m to measure
>> the additional memory consumptions caused by patches such as this.
>
> Ah sorry, I didn't explain that in my second patch I followed  
> Steve's advice and I don't add a new data member in ObjCContainerDecl,
> only a DenseMap in ASTContext that associates interfaces to  
> implementations, if an implementation exists.

Yes, I like that. Then all is OK with me :).

- Thanks, Fariborz

>
>
> So no memory increase for ObCContainterDecl :-)
>
> -Argiris
>