On Feb 24, 7:07 am, "bsmith.occs" <bsmith.o...@gmail.com> wrote:
> On Feb 23, 11:46 pm, Jeffrey Straszheim <straszheimjeff...@gmail.com>
> wrote:
>
> > Have you figured this out yet?
>
> > On Mon, Feb 23, 2009 at 4:58 PM, Jeffrey Straszheim <
>
> > straszheimjeff...@gmail.com> wrote:
> > > The identifier "fibl" is holding on to the head of the sequence.
>
> Yes, this works:
>
> (defn fibl []
> ((fn h [a b] (lazy-seq (cons a (h b (+ a b))))) 0 1))
>
> ;; (count (str (last (take 100000
> (fibl)))))
> ;; --> 20899
>
> by making fibl a function, the head of the sequence is no longer
> hanging off some global variable, never to be garbage collected.
>
> I think I now understand why it's holding on to the head. lazy-seq
> evaluates its body only once and then caches the result, which would
> be a cons with a number in the car and a lazy-seq in the cdr, which in
> turn... tada! full heap! macroexpand-1 gave me the clue I needed:
>
> => (macroexpand-1 '(lazy-seq (cons a (h b (+ a
> b)))))
>
> (new clojure.lang.LazySeq (fn* [] (cons a (h b (+ a b)))))
>
> Thanks for the pointer in the right direction!
>
and just for completeness, I should note that I've since discovered
that Programming Closure covers this very issue just a few pages
beyond where I had been reading. It's always nice when the author has
magical mind powers. :o)
// ben
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups
"Clojure" group.
To post to this group, send email to clojure@googlegroups.com
To unsubscribe from this group, send email to
clojure+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/clojure?hl=en
-~----------~----~----~----~------~----~------~--~---
