[nem-en] Private members & type inference

[Michal Moskal]

On 1/10/07, Kamil Skalski <kamil.skalski at gmail.com> wrote:
> I generally think that inference for fields with initializers could be
> supported. At least compiler could try to type the expression using
> clean typer and if type is fully bound use it. It could even give
> error otherwise.
> Michal, maybe it could "just work" by using FakedTyper.TypeExpr  ?

Yes, this could work. There are some timing problems involved though
-- we need to know the type early and at this stage where we need to
know it, we don't know types of other global fields/methods, which
means typing can very easily fail. This can be hopefully fixed with
some kind of queue.

-- 
   Michał