----- Original Message ----- From: "Lorenzo Caminiti" <lorcaminiti@gmail.com> To: <boost@lists.boost.org> Sent: Friday, April 30, 2010 7:39 PM Subject: Re: [boost] [contract] member initializers (was "diff n1962")
On Mon, Apr 19, 2010 at 12:48 PM, vicente.botet <vicente.botet@wanadoo.fr> wrote:
I think all the stuff needed to implement it is more confusing than useful. At the end what is the real problem with base-init > member-init > pre?
I have never experienced a real problem with base-init > member-init > pre.
I will play with this feature a bit more in the next development round of the library. However, currently I am thinking to: 1) Leave base-init > member-init > pre. 2) Document 1) over pre > base-init > member-init as a limitation of the library. 3) Document that eventually the library could support base-init > pre
member-init (so if someone finds a real use case for it, I might consider to implement it in the future).
This seams reasonable _____________________ Vicente Juan Botet Escribá http://viboes.blogspot.com/