IMHO, for what you describe, i.e. "I'd like to view an atom's changes in
state as a lazy sequence." there is no need to change the IAtom or
add-watch function.
Watchers are called for their side-effects, so if you want to consume their
output as a "lazy sequence", you have to do that transformation step in the
watcher function that you attach to the atom/ref/agent.
Having said that, as changes to an atom happen asynchronously, I don't
think that exposing it as lazy-seq is very efficient way of consuming the
values.
However, with the advent of core.async, there is now a very elegant way of
consuming them via a channel and that channel can have a transducer stack
attached (clojure 1.7+), so something along the lines of
(def atom-to-watch (atom 0))
(defn make-watcher [out-ch-with-transducer]
(fn [x]
(async/put! out-ch-with-transducer x))))
(def ch (async/chan 100 your-transducer-stack))
(add-watch atom-to-watch (make-watcher ch))
;; somewhere else in your code you can consume the values
(go
(loop [v (async/<! ch)]
...
))
Obviously you need to make sure that the values get consumed on the channel
or use a dropping or sliding buffer.
hope this helps,
Las
2014-10-27 5:39 GMT+00:00 Mike Thompson <m.l.thompson...@gmail.com>:
>
> I've been reading about transducers with interest.
>
> The official docs at http://clojure.org/transducers say "*Because
> transducers are decoupled from input or output sources, they can be used in
> many different processes - collections, streams, channels, observables,
> etc.* "
>
> At this stage, there is direct support for collections and core.async. But,
> given the comment above about "observables", this got me to wondering
> about another scenario ...
>
> I'd like to view an atom's changes in state as a lazy sequence. The first
> state of the atom is like the first item in the seq, the next time the atom
> changes state, its new state is the second item in the seq is generated,
> etc. An infinite seq of new states. (I want to do this in the context of
> clojurescript and a GUI library called reagent. Think FRP).
>
> Now imagine that I could attach a transducer to this seq of state changes,
> so that I end up with a modified seq of state changes.
>
> Now, I can figure out how to do this by gluing together an atom watcher
> with a core.sync channel. All doable with a small bit of work, I guess.
>
> But I wondered. Given the claim above about transducers and how they can
> be used in the case of observables, should this atom-as-seq-with-transducer
> be easier? Should the IAtom interface or the add-watcher function be
> changed to make this process easier?
>
> Just a thought.
>
> --
> Mike
>
>
> --
> 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
> ---
> You received this message because you are subscribed to the Google Groups
> "Clojure" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to clojure+unsubscr...@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.
>
--
László Török
Checkout justonemorepoint.com - Know your true value
--
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
---
You received this message because you are subscribed to the Google Groups
"Clojure" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to clojure+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
