From: "David A. Black" <dblack@...>
Date: 2003-12-11T11:45:44+09:00
Subject: Re: Adjusting the Scope of Blocks

Hi --

On Wed, 10 Dec 2003, Dan Doel wrote:

> [ interesting Lisp remarks which I fully intend to go back ]
> [ and look at again :-)                                    ]
>
> Proc objects are the same way. Code + context.  #call works like a
> normal closure (same as calling the method in Scheme), while #?_eval
> replaces some of the context.

At the risk of repeating myself: yeah, but that's the part I don't
like :-)  Actually I think it is, for me, a lexical-scope thing.
Guy's example, or something like:

  obj.instance_eval { stuff }

where it's all visible and lexically scoped together, doesn't bug me
at all.  It only starts feeling "trick"-like when the fact that the
block is going to be instance_eval'd in obj's self context isn't
manifest.  

OK, trying to zero in on this, here's a case.  What if I *want* to use
a locally defined method in a block?  For example:

   def x; [1,2,3]; end
   obj.some_method { x }   # I want the block to evaluate to [1,2,3]
                           # if called or yielded to.  That's what I
                           # mean by 'x'.  

Now, if #some_method's class happens to have an #x method,
#some_method can sort of hijack my call to #x.  But I shouldn't know
that or have to know it.  If my "x" is going to be eval'd as a method
call other than the one I thought it was, I want to be in on the deal
and send it as an eval string.  

It reminds me a bit of things that depend on coordination (and
therefore coupling) of local variable names between methods.  It also
seems to be predicated on a somewhat arbitrary variable/method-call
split; if I'd done x=[1,2,3] above, instead of defining #x, 'x'
wouldn't be vulnerable.

> In some sense it is.  self.instance_eval is a no-op with respect to
> Strings passed in.  self.instance_eval is a no-op as long as you're
> just talking about a collection of code. It's only not a no-op when
> you take context into account.

(I think that makes it a sometimes-op :-)

> [...]
> >I may have passed over that thread... I'll look again.  
> >  
> >
> The subject at the top of the thread was: Re: Controlled Block
> Variables You actually replied to my first post about it basically
> saying you didn't like the idea of blocks-as-just-code. :) Then
> nothing after that so maybe you didn't read anything else.

Yeah, I looked up the thread and found that I had participated in it
and forgotten.  Old age, you know :-)


David

-- 
David A. Black
dblack@wobblini.net