Re: GSOC | Extending Common Lisp support

[Jamison Hope <jrh at theptrgroup dot com>]

Sorry I didn't get back to you last night, was too tired when I got  
home.

On Apr 24, 2012, at 6:57 PM, Charles Turner wrote:

> On 24 April 2012 23:50, Jamison Hope <jrh@theptrgroup.com> wrote:
> > If you're creating a new let via Compilation#letStart() and going the
> > letVariable()/letEnter()/letDone() route, then you don't need to  
> > explicitly
> > do the lexical.push(). I had just mentioned that in case you were  
> > going
> > to reuse an existing let for which letEnter() had already been  
> > called --
> > in that case, you'd need to call let.add(decl) as well as  
> > lexical.push(decl)
> > in order to simulate the letVariable()/letEnter() calls.
>
> I did briefly consider reusing an existing LET, but wasn't sure how to
> get a handle on the most recent LetExp, in some case (like after a
> DEFUN), not such expression would exist IIUC.

I *think* that you don't necessarily need the most recent LetExp, just
the nearest enclosing ScopeExp, which should be current_scope (or maybe
current_scope.outer.. not sure how the "LetExp defs" stuff in  
rewrite_body
affects it). LambdaExp extends ScopeExp, too, so you should be OK there.

So my thought was to do something like this (in processDeclArgs):

Object var = x.get(1);
decl = this.lexical.get(var);
if (current_scope.decls does not contain decl) // must iterate over  
the decls linked list to determine this
{
  // create a new Declaration(decl.getSymbol(), exp2Type(x)),
  // set its value to be the same as decl's, and insert it into
  // current_scope (and call lexical.push(), so that any subsequent
  // this.lexical.get(var) would find the correct Declaration)
}
else // decl did in fact come from the surrounding ScopeExp, be it a  
lambda or let or whatever
{
  // do what you originally had:
  decl.setType()
  decl.setFlag()
}

-Jamie

--
Jamison Hope
The PTR Group
www.theptrgroup.com