[cfe-commits] r96274 - /cfe/trunk/test/Rewriter/rewrite-implementation.mm

Fariborz Jahanian fjahanian at apple.com
Tue Feb 16 09:28:10 PST 2010


On Feb 16, 2010, at 8:45 AM, Daniel Dunbar wrote:

> On Tue, Feb 16, 2010 at 8:12 AM, Fariborz Jahanian <fjahanian at apple.com 
> > wrote:
>>
>> On Feb 15, 2010, at 7:16 PM, Daniel Dunbar wrote:
>>
>>> On Mon, Feb 15, 2010 at 1:37 PM, Fariborz Jahanian <fjahanian at apple.com 
>>> >
>>> wrote:
>>>>
>>>> Author: fjahanian
>>>> Date: Mon Feb 15 15:37:11 2010
>>>> New Revision: 96274
>>>>
>>>> URL: http://llvm.org/viewvc/llvm-project?rev=96274&view=rev
>>>> Log:
>>>> Test to compile the rewritten file for my last patch.
>>>> (radar 7649577 related).
>>>>
>>>> Added:
>>>>   cfe/trunk/test/Rewriter/rewrite-implementation.mm
>>>>
>>>> Added: cfe/trunk/test/Rewriter/rewrite-implementation.mm
>>>> URL:
>>>> http://llvm.org/viewvc/llvm-project/cfe/trunk/test/Rewriter/rewrite-implementation.mm?rev=96274&view=auto
>>>>
>>>>
>>>> = 
>>>> = 
>>>> = 
>>>> = 
>>>> = 
>>>> = 
>>>> = 
>>>> = 
>>>> = 
>>>> = 
>>>> = 
>>>> ===================================================================
>>>> --- cfe/trunk/test/Rewriter/rewrite-implementation.mm (added)
>>>> +++ cfe/trunk/test/Rewriter/rewrite-implementation.mm Mon Feb 15  
>>>> 15:37:11
>>>> 2010
>>>> @@ -0,0 +1,13 @@
>>>> +// RUN: %clang_cc1 -x objective-c++ -fblocks -fms-extensions
>>>> -rewrite-objc %s -o %t-rw.cpp
>>>> +// RUN: %clang_cc1 -DSEL="void *" -S %t-rw.cpp
>>>
>>> Does using -S here make sense, wouldn't -fsyntax-only be just as  
>>> good?
>>
>> I prefer to go all the way to code gen. So, to catch as much as  
>> possible any
>> rewritten error.
>> But, when it makes sense (and I run into clang++ bugs), I will be  
>> using
>> -fsyntax-only. I fact,
>> I may have to in some cases where clang++ Sema is more restrict  
>> than g++.
>
> Ok, but codegen in Clang shouldn't ever generate an error, so it
> should never be the case that -S fails and -fsyntax-only succeeds.
> That would be a bug, but a bug somewhere else.

Yes. That constitutes a clang bug, which would follow with a bugzilla  
filing.

- Fariborz

>



More information about the cfe-commits mailing list