Hi!

On Thu, Jan 28, 2010 at 8:39 AM, Michał Marczyk
<michal.marc...@gmail.com> wrote:
> 2010/1/28 Konrad Hinsen <konrad.hin...@fastmail.net>:
>> The Clojure solution for your problem is an agent, which makes the access
>> thread-safe:
>>
>> user> (def data-source (agent (cycle [1 2 3])))
>> #'user/data-source
>> user> (defn get-some-data [] (let [v (first @data-source)] (send data-source
>> rest) v))
>
> Wouldn't this make it possible for two threads to obtain the same
> value of (first @data-source), then send two rest messages to the
> agent?
>
> A ref would not have this problem, though:
>
> (def data-source (ref the_sequence))
> (defn get-some-data [] (dosync (let [v (first @data-source)] (alter
> data-source rest) v)))

Right, or you could also use an atom:
(def data-source (atom (cons nil the_sequence)))
(defn get-some-data [] (first (swap! data-source rest)))

But this retains the last read value until next call to get-some-data.
I recently realized that one can work around this problem:
(def data-source (atom (cons nil the_sequence)))
(defn get-some-data []
  (let [s (swap! data-source rest)]
    (swap! data-source #(if (identical? s %) (cons nil (rest %)) %))
    (first s)))

This make me think that one could optimize Atom/swap by avoiding the
CAS and the validation (but not the watches) when (identical?
new-value old-value).

Christophe

-- 
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
Note that posts from new members are moderated - please be patient with your 
first post.
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

Reply via email to