> wouldn't you want to provide the source range to the microsoft-mode
> diagnostic too?
r148683, good catch.
>> +namespace PR11791 {
>> + template<class _Ty>
>> + void del(_Ty *_Ptr) {
>> + _Ptr->~_Ty(); // expected-warning {{pseudo-destructors on type void
>> are a Microsoft extension}}
>> + }
>> +
>> + void f() {
>> + int* a = 0;
>> + del((void*)a); // expected-note {{in instantiation of function
>> template specialization}}
>> + }
>> +}
>
> This test case seems a bit more convoluted than necessary - would:
>
> typedef void *foo;
> foo f;
> f.~foo();
>
> suffice?
typedef void *foo;
foo f;
f->foo::~foo();
does suffice. The current test looks easier to understand to me
(pseudo-destructors were new to me), which is why I went with that.
Nico
_______________________________________________
cfe-commits mailing list
[email protected]
http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
