I see what you are doing with the validations and defining them in the
init-modelcall.
I think that's a good idea actually and like it better than my solution to
putting the
callbacks in the model namespace.
A couple of small things that I noticed when starting to play around with
clj-record. In your
example you have:
(ns com.example.user)
(clj-record.core/init-model)
but when I do that I get the error:
java.lang.ClassNotFoundException: clj-record.core
When I add a require it clears it up like:
(ns com.example.user
(:require [clj-record.core :as cr.core]))
(cr.core/init-model)
I know this is picky, but in your clj-record.core/init-model function you
have:
(first (reverse (re-split #"\." (name (ns-name *ns*)))))
and could be changed to:
(last (re-split #"\." (name (ns-name *ns*))))
I look forward to seeing clj-record progress. Thanks John.
On Mon, Dec 29, 2008 at 10:20 PM, John D. Hume <duelin.mark...@gmail.com>wrote:
>
> On Mon, Dec 29, 2008 at 7:15 PM, Brian Doyle <brianpdo...@gmail.com>
> wrote:
> > I noticed that in the init-model macro you are creating a 'defn table []
> > ...' function in the
> > model namespace and was wondering why you didn't just make it a def
> instead,
> > since
> > it doesn't take any args?
>
> That didn't occur to me. I do like the idea of all the functionality
> being available directly on clj-record.core with model name as a first
> argument in addition to being defined in the model namespace, so I
> guess that's one reason. But the inference of table name from model
> name should only be the default, so maybe it would be better to just
> def it there and let clj-record.core look it up there if needed,
> somewhat as you recommended doing for callbacks...
>
> > I have an idea for you with the callbacks like, before-create,
> after-create,
> > etc. In the
> > all-models-metadata in the core.clj you could store the namespace in the
> > hash
> > associated with the model-name on setup. Then in the create function
> > defined in
> > the core.clj you could call the before-create function defined in the
> model
> > something
> > like:
> >
> > (let-if [bc ('before-create (ns-publics <model-namespace>)] (bc
> > attributes#))
>
> That's interesting. It's basically the opposite of the way I did the
> validation stuff, where you pass in validation functions and those get
> stored away in metadata and aren't directly accessible. (Here
> "metadata" is not used in the clojure.org/metadata sense of the word.
> I've thought about using clojure metadata to store model info, but I
> don't know what symbol or collection I'd attach it to.) I think I
> prefer the validation approach, because those functions have a certain
> job to do and shouldn't be sitting there cluttering the model
> namespace. (On the other hand, it is awfully convenient to be able to
> test those functions directly.)
>
> Thanks again for looking. If you or anyone else has further comments,
> I'd like to hear them.
> git clone git://github.com/duelinmarkers/clj-record.git
>
> -hume.
>
> --
> http://elhumidor.blogspot.com/
>
> >
>
--~--~---------~--~----~------------~-------~--~----~
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
-~----------~----~----~----~------~----~------~--~---
