Indeed.  This seems symptomatic of an invariant being broken elsewhere.

On Apr 29, 2011, at 8:46 AM, jahanian wrote:

> Shouldn't this be caught at Sema type? A test case would clarify the 
> situation.
> 
> - Fariborz
> 
> On Apr 29, 2011, at 7:10 AM, David Chisnall wrote:
> 
>> Author: theraven
>> Date: Fri Apr 29 09:10:35 2011
>> New Revision: 130500
>> 
>> URL: http://llvm.org/viewvc/llvm-project?rev=130500&view=rev
>> Log:
>> Don't crash if the AST doesn't have a sensible ObjC id type.
>> 
>> 
>> Modified:
>>   cfe/trunk/lib/CodeGen/CGObjCGNU.cpp
>> 
>> Modified: cfe/trunk/lib/CodeGen/CGObjCGNU.cpp
>> URL: 
>> http://llvm.org/viewvc/llvm-project/cfe/trunk/lib/CodeGen/CGObjCGNU.cpp?rev=130500&r1=130499&r2=130500&view=diff
>> ==============================================================================
>> --- cfe/trunk/lib/CodeGen/CGObjCGNU.cpp (original)
>> +++ cfe/trunk/lib/CodeGen/CGObjCGNU.cpp Fri Apr 29 09:10:35 2011
>> @@ -688,11 +688,13 @@
>>  PtrTy = PtrToInt8Ty;
>> 
>>  // Object type
>> -  ASTIdTy = 
>> CGM.getContext().getCanonicalType(CGM.getContext().getObjCIdType());
>> -  if (QualType() == ASTIdTy) {
>> -    IdTy = PtrToInt8Ty;
>> -  } else {
>> +  QualType UnqualIdTy = CGM.getContext().getObjCIdType();
>> +  ASTIdTy = CanQualType();
>> +  if (UnqualIdTy != QualType()) {
>> +    ASTIdTy = CGM.getContext().getCanonicalType(UnqualIdTy);
>>    IdTy = cast<llvm::PointerType>(CGM.getTypes().ConvertType(ASTIdTy));
>> +  } else {
>> +    IdTy = PtrToInt8Ty;
>>  }
>>  PtrToIdTy = llvm::PointerType::getUnqual(IdTy);
>> 
>> 
>> 
>> _______________________________________________
>> cfe-commits mailing list
>> [email protected]
>> http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
> 
> _______________________________________________
> cfe-commits mailing list
> [email protected]
> http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits

_______________________________________________
cfe-commits mailing list
[email protected]
http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits

Reply via email to