On Jan 18, 2011, at 7:21 PM, Chas Emerick wrote: > You might find my minor sugar for defining records whose slots have default > values: > > http://cemerick.com/2010/08/02/defrecord-slot-defaults/
Nice -- I do think it'd be good for something like this to be built in. > Also worth noting is the further enhancement by Michael Fogus (linked to in > the comments on the above post) that adds support for kwargs to the factory > function for the defined record. Oh yes, keyword args are another nice thing about struct-map that I'd forgotten about but which have also been important to me in some contexts, e.g. when there are lots of args. > I also like a few aspects of what structmaps provide vs. regular maps (in > particular, documentation of intent), but records have too many far more > significant advantages and are too easy to add sugar to to reclaim the minor > bits structmaps offer to not use IMO. If the above conveniences get baked in then I'll probably agree. But FWIW I just looked over some old code and see that records also made things more complicated in my application because my defstruct/defrecord is produced by a macro expansion and that was simpler to arrange with defstruct. At least how I was able to see how to do it. Perhaps some will think this is some form of macro abuse, but in the context in which I conduct my runs it does the job. I'll bet some of you will also see much better ways for me to do the record version :-) Also BTW I saw no clear performance improvement (relative to the overall runtime of my system); some numbers on that are also below. For the morbidly curious (only) I'll include some details below. Thanks, -Lee ;; Some details from my code at https://github.com/lspector/Clojush ;; Here's a var that has a value that may be used in a couple of places including the struct/record definition: (def push-types '(:exec :integer :float :code :boolean :auxiliary :tag :zip)) ;; Here's the way I define things as struct-maps: (defmacro define-push-state-structure [] `(defstruct push-state ~@push-types)) (define-push-state-structure) (defn make-push-state "Returns an empty push state." [] (struct-map push-state)) ;; To switch to records I ended up doing something with a few more complications in a couple of places: (defn keyword->symbol [kwd] "Returns the symbol obtained by removing the : from a keyword." (read-string (name kwd))) (defmacro define-push-state-record-type [] "Defines the pushstate record type. The odd trick with read-string was a hack to avoid namespace qualification on the pushstate symbol." `(defrecord ~(read-string "pushstate") [~@(map keyword->symbol push-types)])) (define-push-state-record-type) (defmacro make-push-state "Returns an empty push state." [] `(pushstate. ~@(map (fn [_] nil) push-types))) ;; My performance tests are big stochastic messes but still they give a basic picture. ;; For each condition I start with a fresh JVM launch and then run the same test 3 times. ;; here's what I get with structmaps: ;1:3 clojush=> (time (stress-test 100000)) ;:no-errors-found-in-stress-test ;"Elapsed time: 59194.123 msecs" ; more runs: ;"Elapsed time: 52389.016 msecs" ;"Elapsed time: 54706.127 msecs" ;; here's what I get with records: ;1:3 clojush=> (time (stress-test 100000)) ;:no-errors-found-in-stress-test ;"Elapsed time: 57673.781 msecs" ; more runs: ;"Elapsed time: 57682.971 msecs" ;"Elapsed time: 55129.614 msecs" -- 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
