Sharing in D

[Robert Fraser]

bearophile Wrote:
> Robert Fraser:
> > However, I'm worried that this may perhaps add even more complexity
> > to an already stretched type system.
> 
> If the type system is able to manage the transitive immutability, then it can manage this too, I think. The things that are going to stress the type system of D are different ones, I think; for example see the thread about nested templates, where the best solution proposed was to use an union to punch a hole in it.

I was talking UX wise. I don't want to have to think whether my variable 
is a shared(atomic(const(int)*)[]).