libstdc++ not conforming with --enable-concepts-check?

[Paolo Carlini]

Matt Austern wrote:

> Yes.  I put them in a long time ago, before we had the language rule 
> saying that default arguments were instantiated separately from the 
> function.  (i.e. if you supply both arguments then the default isn't 
> instantiated)  At the time this was the only way to make sure that you 
> could have containers of elements that didn't have default 
> constructors.  It was important at the time, but it's now irrelevant.

Thanks again: these days, besides helping improving the lib ;) I'm 
learning a lot!

Paolo.