[e-lang] Bug (0.8.35f): Scope oddities in ObjectExpr with 'extends'; Kernel-E change proposal

[Mark S. Miller]

Kevin Reid wrote:
>   ? e`def a extends (def b := 1) {}`.staticScope()
>   # value: <[] := ["E"] =~ ["a"] + var []>
> 
> 'b' is not visible outside.

That's what I'd intended.


>   ? e`def a implements (def b := 1) {}`.staticScope()
>   # value: <[] := [] =~ ["b", "a"] + var []>
> 
> 'b' is visible outside.

That's a bug. Curiously, in E-on-Java, the bug is only in the calculation of 
the staticScope. Actual execution hides b:

     ? def a implements (def b { to audit(_,_) :any { return true }}) {}
     # value: <a>

     ? b
     # problem: Failed: Undefined variable: b


> I think this could be fixed by a minor contortion of the ObjectExpr  
> expansion:

It can't be fixed by expansion, since Kernel-E must be a subset of E. Except 
for the awful Hilbert hotel renaming issue, a Kernel-E expression expanded by 
the E-to-Kernel-E expansion should expand to itself.

The semantics of a Kernel-E ObjectExpr should be that it exports no 
definitions into the surrounding scope. The calculation of staticScope should 
be fixed to reflect that.

-- 
Text by me above is hereby placed in the public domain

     Cheers,
     --MarkM