Parameter pack in non-last position in function formal parameter list?

[Avi Kivity]

On 06/19/2015 07:31 PM, Marc Glisse wrote:
> On Fri, 19 Jun 2015, Avi Kivity wrote:
>
>> Consider
>>
>>  template <typename... X, typename Y>
>>  void f(X... x, Y y) {
>>  }
>>
>>
>>  int main(int ac, char** av) {
>>    f(1, 2, 3, 4);
>>    return 0;
>>  }
>>
>> is this legal?
>
> No.
>
>> So it's not complaining that f() is illegal, yet it id not deducing 
>> X... as int, int, int.
>
> That's what the standard says. You can still call it as:
> f<int,int,int>(1, 2, 3, 4);
>
> (the question is not specific to gcc, so it would be more appropriate 
> on a general C++ forum like stackoverflow.com)
>

Thanks.

I wonder if gcc can give a better diagnostic here.  Instead of

   ./variadic-function-nonlast.cc:9:15: note:   candidate expects 1 
argument, 4 provided

it might say,

    ./variadic-function-nonlast.cc:9:15: note:   candidate expects 1 
argument, 4 provided
./variadic-function-nonlast.cc:9:15: note:   parameter pack `X' in 
non-last position deduced as empty parameter pack