[cfe-commits] r147900 - in /cfe/trunk: lib/Sema/SemaExprCXX.cpp lib/Sema/SemaType.cpp test/SemaObjC/property.m test/SemaObjCXX/properties.mm

[John McCall]

On Jan 10, 2012, at 4:32 PM, jahanian wrote:
> On Jan 10, 2012, at 4:14 PM, John McCall wrote:
>> Author: rjmccall
>> Date: Tue Jan 10 18:14:46 2012
>> New Revision: 147900
>> 
>> URL: http://llvm.org/viewvc/llvm-project?rev=147900&view=rev
>> Log:
>> Do placeholder conversions on array bounds in both declarators and
>> new-expressions.
>> 
>> Modified:
>>   cfe/trunk/lib/Sema/SemaExprCXX.cpp
>>   cfe/trunk/lib/Sema/SemaType.cpp
>>   cfe/trunk/test/SemaObjC/property.m
>>   cfe/trunk/test/SemaObjCXX/properties.mm
> 
> Curious as to why testing objc/objc++. Is this specific to properties?
> Otherwise, this may not be tested for those who don't care 
> about these two dialects.

At the moment, we have no placeholder expressions other than ObjC
property references that can get resolved to something with integer type
and therefore be valid in these locations.  You're right that we could
nonetheless write tests in pure C++, but those tests would just be
testing that we emit one of two very similar possible diagnostics.
And given the nature of the code to check for placeholders, I'm
not really worried about completeness across languages.

John.