Hi Stuart, Thanks for checking that out for me! Sorry for not realizing in the first place.
I of course would be happy to submit a patch. Should I submit that here or over on the assembla page? On Jul 13, 9:10 am, Stuart Halloway <stuart.hallo...@gmail.com> wrote: > Hi Cam, > > Your tests aren't testing the interesting part without a doall. > > That said, my quick tests with doall show your approach faring even better. > :-) Also, I think what my-flatten does with Java arrays is intuitive (and the > current flatten not so much). > > A patch that preserves the semantics of the existing flatten (except for > working with Java arrays) would be welcome. > > Thanks! > Stu > > > > > Another flatten thread! Sorry.. > > > Hello all, before I realized there was a flatten in the master branch > > (and before I looked at contrib) I wrote this pretty standard code: > > > (defn my-flatten [coll] > > (lazy-seq > > (when-let [coll (seq coll)] > > (let [x (first coll)] > > (if (sequential? x) > > (concat (my-flatten x) (my-flatten (next coll))) > > (cons x (my-flatten (next coll)))))))) > > > (There's very similar versions on the boards. I'm not claiming this is > > anything amazing or unique.) > > > It's not as elegant as what's in core, but in my micro benchmarks (ran > > on my laptop; 2.26 core 2 and 4gb ram) it seems to perform a bit > > better, _especially_ in the already flattened case. It behaves just > > like core/flatten except that it doesn't return an empty list when > > passed a map or set, it just returns whatever you gave it but with the > > top level converted to a seq. I'm pretty much a clojure noob, so are > > there any hidden detractors of this implementation as opposed to the > > version introduced in 1.2? > > > Also, quick note, if you swap the call to sequential? with seqable? > > from contrib/core, it flattens maps and sets like you'd expect as > > well. > > Here is how it looks > > user=> (my-flatten #{1 {2 3} 4 [5 6 7 #{8 {9 10}}]}) > > (1 2 3 4 5 6 7 9 10 8) > > > And for the micro-benchmarks (using "sequential?"): > > > user=> (time (dotimes [_ 1e7] (flatten [1 2 3 4]))) > > "Elapsed time: 14,661.592 msecs" > > nil > > > user=> (time (dotimes [_ 1e7] (my-flatten [1 2 3 4]))) > > "Elapsed time: 922.268 msecs" > > nil > > > user=> (time (dotimes [_ 1e7] (flatten [1 [2 [3 [4 [5 [6 [7 [8] > > [[[9]]] 10 [11] 12 [13 14 [15]]]]]]]]]))) > > "Elapsed time: 18,147.959 msecs" > > nil > > > user=> (time (dotimes [_ 1e7] (my-flatten [1 [2 [3 [4 [5 [6 [7 [8] > > [[[9]]] 10 [11] 12 [13 14 [15]]]]]]]]]))) > > "Elapsed time: 6,088.914 msecs" > > nil > > > user=> (time (dotimes [_ 1e7] (flatten [[1 2 3 4 5 6 7 8 9 10]]))) > > "Elapsed time: 11,696.693 msecs" > > nil > > > user=> (time (dotimes [_ 1e7] (my-flatten [[1 2 3 4 5 6 7 8 9 10]]))) > > "Elapsed time: 1,533.983 msecs" > > nil > > > Thoughts? > > > -- > > 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 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