23 nov 2009 kl. 21.35 skrev Eli Friedman:
>> 
> 
> I'm not sure if this matters other than the cleanliness of the IR, but
> this isn't quite the same as the normal handling for incomplete
> function types in CodeGenTypes::ConvertNewType.  Specifically, this
> bypasses the handling to resolve the opaque type to a proper function
> type.

Nope, it doesn't matter since we don't need/want to resolve the opaque type.

> 
> Also, HasIncompleteReturnTypeOrArgumentTypes is roughly equivalent to
> VerifyFuncTypeComplete in CodeGenTypes.cpp.
> 

Yeah, I noticed that. Will refactor.

Anders


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

Reply via email to