[llvm] r190304 - Silencing a warning about control flow reaching the end of a non-void function.

[Duncan Sands]

On 09/09/13 16:34, Aaron Ballman wrote:
> On Mon, Sep 9, 2013 at 10:20 AM, Duncan Sands <duncan.sands at gmail.com> wrote:
>> Hi Aaron,
>>
>>
>> On 09/09/13 15:22, Aaron Ballman wrote:
>>>
>>> Author: aaronballman
>>> Date: Mon Sep  9 08:22:45 2013
>>> New Revision: 190304
>>>
>>> URL: http://llvm.org/viewvc/llvm-project?rev=190304&view=rev
>>> Log:
>>> Silencing a warning about control flow reaching the end of a non-void
>>> function.
>>>
>>> Modified:
>>>       llvm/trunk/lib/Target/XCore/XCoreLowerThreadLocal.cpp
>>>
>>> Modified: llvm/trunk/lib/Target/XCore/XCoreLowerThreadLocal.cpp
>>> URL:
>>> http://llvm.org/viewvc/llvm-project/llvm/trunk/lib/Target/XCore/XCoreLowerThreadLocal.cpp?rev=190304&r1=190303&r2=190304&view=diff
>>>
>>> ==============================================================================
>>> --- llvm/trunk/lib/Target/XCore/XCoreLowerThreadLocal.cpp (original)
>>> +++ llvm/trunk/lib/Target/XCore/XCoreLowerThreadLocal.cpp Mon Sep  9
>>> 08:22:45 2013
>>> @@ -122,6 +122,7 @@ createReplacementInstr(ConstantExpr *CE,
>>>        default:
>>>          assert(0 && "Unhandled constant expression!\n");
>>>      }
>>> +  llvm_unreachable("Unhandled constant expression!\n");
>>
>>
>> how about removing the default case, since it isn't doing anything
>> useful any more.
>
> Wouldn't that trigger -Wcovered-enum warnings?

If so, maybe move the llvm_unreachable up into the default case, or
let the default fall through?

Ciao, Duncan.