Le 10/01/13 22:10, Vicente J. Botet Escriba a écrit :
Le 10/01/13 20:47, Andrey Semashev a écrit :
On Thu, Jan 10, 2013 at 11:31 PM, Vicente J. Botet Escriba <vicente.botet@wanadoo.fr> wrote:
Le 10/01/13 20:12, Andrey Semashev a écrit :
On Thu, Jan 10, 2013 at 10:51 PM, Vicente J. Botet Escriba <vicente.botet@wanadoo.fr> wrote:
Le 10/01/13 19:13, Andrey Semashev a écrit : The constructor doesn't make any difference, it's the at() method body that causes the error. Yes it does. The implicit default constructor is constexpr when possible, so defining it constexpr explicitly doesn't make any difference.
Maybe you are right. I said that because clang was requiring it but gcc-4.7 accept the default constructor as a constexpr.
I have checked it and you were right. "If that user-written default constructor would satisfy the requirements of a constexpr constructor (7.1.5), the implicitly-defined default constructor is constexpr" It seems that clang 3.2 has a bug here. Thanks, Vicente