[Nickle]Mystic shallow copy semantics solved

[Keith Packard]

Around 21 o'clock on Feb 19, Keith Packard wrote:
> 
> When 'b' is undeclared, the assignment 'b = a' makes a shallow copy of 'a',
> when 'b' is declared as an array, the copy is deep.
> 
> Oops.

And fixed in CVS now -- the type of copy depended on the declaration of 
the lvalue instead of the type of the object itself, now the object type 
is all that matters.

There is still a hack here -- objects declared 'double' always have their 
values stored in 'double' form.  'double's are the only imprecise datatype 
supported currently, and I don't have any special operators for dealing 
with imprecise numbers yet.

I'll poke through the scheme spec and try to remember how they did this 
stuff.

-keith