[cfe-commits] r126897 - /cfe/trunk/lib/Analysis/CFG.cpp

[Matt Beaumont-Gay]

On Wed, Mar 2, 2011 at 16:35, Ted Kremenek <kremenek at apple.com> wrote:
> +1.
>
> All the enum cases were enumerated, and I want a compiler warning (from Clang) to tell us when we miss updating the switch statement instead of waiting for an assertion failure.  A 'default' case removes this checking.

Done in r126907.

>
> On Mar 2, 2011, at 3:33 PM, John McCall wrote:
>
>> On Mar 2, 2011, at 3:25 PM, Matt Beaumont-Gay wrote:
>>> Author: matthewbg
>>> Date: Wed Mar  2 17:25:06 2011
>>> New Revision: 126897
>>>
>>> URL: http://llvm.org/viewvc/llvm-project?rev=126897&view=rev
>>> Log:
>>> Keep GCC from complaining about falling off the end of the function.
>>>
>>> Modified:
>>>   cfe/trunk/lib/Analysis/CFG.cpp
>>>
>>> Modified: cfe/trunk/lib/Analysis/CFG.cpp
>>> URL: http://llvm.org/viewvc/llvm-project/cfe/trunk/lib/Analysis/CFG.cpp?rev=126897&r1=126896&r2=126897&view=diff
>>> ==============================================================================
>>> --- cfe/trunk/lib/Analysis/CFG.cpp (original)
>>> +++ cfe/trunk/lib/Analysis/CFG.cpp Wed Mar  2 17:25:06 2011
>>> @@ -2774,6 +2774,7 @@
>>>
>>> const CXXDestructorDecl *CFGImplicitDtor::getDestructorDecl() const {
>>>  switch (getKind()) {
>>> +    default: assert(0 && "Unknown CFGElement");
>>>    case CFGElement::Invalid:
>>>    case CFGElement::Statement:
>>>    case CFGElement::Initializer:
>>
>> The general style in Clang is (1) to use llvm_unreachable(msg) instead of assert(0 && msg) and (2) to not add default cases to switches which are supposed to be exhaustive.  Please move this to after the switch and follow it with a return 0.
>>
>> John.
>> _______________________________________________
>> cfe-commits mailing list
>> cfe-commits at cs.uiuc.edu
>> http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
>
>