> 
> I noticed that the following statement is NOT an compilation error - as I'd
> suspect but evaluates to "2".
> 
> #(1 2) do: [:i | i].
> i
> 
> I'm aware of that this happens because of the way, block arguments are
> dealt with but still, it's a source of difficult to find errors.
> 


You're right.  Maybe when closure support is added, the compiler will
be rewritten, and at that point this scope violation will be
fixed.


Lex