I couldn't resist writing some more macros!
(defmacro << [k-or-ks f]
`(fn [state#]
(update-in state#
~(if (vector? k-or-ks)
k-or-ks
(vector k-or-ks))
~f)))
(defn >>* [state form]
(condp = (first form)
'>>> `(fn [~state] ~@(rest form) ~state)
'<<< `(fn [~state] ~@(rest form))
form))
(defmacro >> [state & forms]
(reduce #(list (if (#{'<< 'fn 'fn*
'clojure.core/fn
'clojure.core/fn*} (first %2))
%2
`(fn [s#] ~%2 s#)) %)
(conj (map (partial >>* state) forms)
state)))
With those the resulting code looks pretty clean and remain purely
functional:
(next-phase [state]
(>> state
(stop-timer)
(<< :phases #(conj (vec (rest %)) (first %)))
(>>> (log :info "change phase to %s" (-> state :phases first key)))
(<<< (start-phase state))))
Now, I'd need to find better names!
On Sunday, April 15, 2012 11:25:21 PM UTC-4, Nicolas Buduroi wrote:
>
> I'm working on a turn-based game and I'm looking for a good way to manage
> states. In the game each turn is composed of multiple phases. I started by
> using atoms for the phases field (this is a sequence of functions) in a
> record and realized that it wouldn't be ideal to keep track of states in
> the case where I'd need to keep a snapshot of every phases. Here's the
> original code I had:
>
> (defrecord Game [phases]
> (next-phase [this]
> (stop-timer)
> (swap! phases #(conj (vec (rest %)) (first %)))
> (log :info "change phase to %s" (key (first @phases)))
> (start-phase this))
>
> I then started to think that this would be a good opportunity to use a
> state monad. I've tried to reimplement the above code using the algo.monads
> library but the result was less than satisfactory (probably due to my own
> shortcoming), here's the monadic version:
>
> (defrecord Game [phases]
>
> (next-phase [this]
> (->
> ((domonad state-m
> [_ (fn [s] (stop-timer) [s s])
> _ (update-state
> (fn [s]
> (update-in s [:phases]
> #(conj (vec (rest %)) (first %)))))
> _ (fn [s]
> (log :info "change phase to %s" (key (first (:phases s)))) [s
> s])]
> nil)
> state)
> second
> start-phase))
>
> As my code probably doesn't need the full power of the state monad, I
> tried to write a lighter-weight version using the following macro:
>
> (defmacro >> [& state-and-forms]
> (reduce #(list (if ('#{fn fn*} (first %2))
> %2
> `(fn [s#] ~%2 s#)) %)
> state-and-forms))
>
> Which let me write:
>
> (next-phase [state]
> (>> state
> (stop-timer)
> (fn [s] (update-in s [:phases] #(conj (vec (rest %)) (first %))))
> #(do (log :info "change phase to %s" (key (first (:phases %)))) %)
> #(start-phase %)))
>
> With some more helper macro this version looks promising. In the end I
> wonder if there's some Clojure feature I'm overlooking or if I should
> rethink the whole solution? Is there a better way to accomplish this?
>
>
>
--
You received this message because you are subscribed to the Google
Groups "Clojure" group.
To post to this group, send email to [email protected]
Note that posts from new members are moderated - please be patient with your
first post.
To unsubscribe from this group, send email to
[email protected]
For more options, visit this group at
http://groups.google.com/group/clojure?hl=en
