#clojure log - Jul 13 2012

The Joy of Clojure
Main Clojure site
Google Group
IRC
List of all logged dates

21:55 wingy: does anyone here know how to run midje tests from the repl?

22:19 gfredericks: amalloy: correctly handling the introduction of new locals and shadowing takes a bit of concentration

22:23 andres-v: ma

22:33 duck1123: wingy: just eval that buffer

22:33 if you're using emacs, there's midje mode, but it doesn't work well for me

22:44 sjl: I don't suppose there's the equivalent of #(map (partial map %1) %2) lurking around the standard library somewhere?

22:45 i.e.: map2d

22:53 also, what's the idiomatic way to slice apart a two-dimensional vector into a smaller one? all I can find is a google groups post from four years ago

22:54 technomancy: sjl: for is 2-dimensional

22:54 n-dimensional, rather

22:56 gfredericks: ,(let [2dvec [[1 2 3] [4 5 6] [7 8 9]]] (for [a 2dvec b a] (+ b (apply max a))))

22:56 clojurebot: #<ExecutionException java.util.concurrent.ExecutionException: java.lang.NumberFormatException: Invalid number: 2dvec>

22:57 gfredericks: ,(let [twodvec [[1 2 3] [4 5 6] [7 8 9]]] (for [a twodvec b a] (+ b (apply max a))))

22:57 clojurebot: (4 5 6 10 11 ...)

22:59 sjl: one sec

23:01 this is what I'm looking for: https://www.refheap.com/paste/3605

23:02 as for the mapping one, I'd have to use a nested for to get a nested result, rigfht?

23:04 gfredericks: sjl: yeah I think so, about the nesting

23:05 as far as the shearing, you could do a bit better for efficiency, but I'm not sure about simplicity

23:06 though depending on what else you're doing you might not be gaining anything by using vectors over seqs

23:06 sjl: gfredericks: yeah, of course

23:06 gfredericks: what would you change to make it faster? Not sure I need the extra speed, but might be nice to have.

23:06 gfredericks: on line 4

23:06 your mapv

23:07 sjl: yep

23:07 gfredericks: will subvec every entry of the outer vector

23:07 when you only care about some of them

23:07 sjl: true

23:07 gfredericks: sjl: also know that subvec is only efficient in read-only mode

23:07 sjl: read-only mode?

23:08 gfredericks: if you want to make any updates to it you'll pay a linear cost for the subvec

23:08 sjl: hmm, is there something else better?

23:11 gfredericks: well using seqs is more natural for clojure

23:11 I guess there's not a subseq function that I know of, but it'd be a one-liner with drop and take

23:11 sjl: I may be able to get away using seqs in this particular case... I'll have to play around

23:11 thanks

23:12 gfredericks: np

23:12 duck1123: does anyone know what i need to do to make waltz less noisy? I like that it logs all the state changes, but i wouldn't want that in production

23:24 technomancy: https://github.com/kingtim/nrepl.el/issues <- hope I didn't go overboard

23:24 Raynes: technomancy: Is this guy taking over for you?

23:24 eggsby: heh

23:24 technomancy: Raynes: I handed off to him after spending 2 or 3 hours on it

23:24 Raynes: technomancy: Or did you just pick the most recent fork and throw issues at it?

23:25 It would have been so hilarious if you had done what i said.

23:25 technomancy: hehe

23:36 sjl: if anyone's bored and wants to review code, I feel like there's probably a better way to be doing a lot of this roguelike drawing code: https://github.com/sjl/caves/blob/master/src/caves/ui/drawing.clj

23:41 amalloy: sjl: you're passing around start-x and start-y separately to a lot of functions - why not be consistent and use pairs for that as well?

23:41 sjl: amalloy: yeah I was thinking about that. probably a good idea, you're right

23:42 amalloy: then you can do stuff like (defn get-viewport-coords-of [start coords] (map - coords start))

23:43 eggsby: hey sjl i've been enjoying reading your series btw, good stuff :)

23:43 technomancy: Raynes: so nrepl.el was ~100 lines when I handed it off; now it's around ~1000

23:43 sjl: eggsby: thanks!

23:44 technomancy: srsbsns

23:44 clojurebot: Huh?

23:44 amalloy: clojurebot: nrepl.el is srsbsns

23:44 clojurebot: Roger.

23:44 Raynes: I want it so bad.

23:44 technomancy: How useful is it right now?

23:45 technomancy: Raynes: useful!

23:45 Raynes: to see everything it's missing, see the issues link above =)

23:45 duck1123: On paper it looks good. I meant to try it, but decided to do some cljs instead

23:45 technomancy: but it's way way simpler than slime

23:46 I have a branch that implements reloading, and I did simple jump-to-definition, doc, and completion

23:47 Raynes: the best part is how easy the codebase is to work with

23:49 amalloy: technomancy: isn't that unidiomatic for elisp?

23:50 technomancy: amalloy: very =)

23:50 amalloy: have I ranted about vectors though?

23:50 in elisp?

23:50 they're bad beyond belief

23:51 amalloy: more than a few times, yes

23:51 technomancy: just making sure

23:51 amalloy: good call

23:51 technomancy: and hash tables; phew!

23:51 don't even bother

23:51 the only reason I bother with vectors is it makes interop with Clojure easier

23:52 amalloy: really?

23:52 technomancy: sure, you just send clojure forms across the wire and call read-from-string on 'em

23:52 as long as there aren't any maps or hash-tables in them

23:52 sorry

23:52 maps or sets

23:52 geez

23:52 or reader literals

23:53 amalloy: oh, so elisp'

23:53 s vectors are also [x y z]?

23:53 really?

23:53 technomancy: yeah, visually they're the same

23:53 but with implicit quoting

23:53 amalloy: right

23:53 technomancy: namespaced keywords even work in elisp, I was happy to see

23:54 if I port swank.core/break to nrepl I'm going to call it limit-break: http://p.hagelb.org/limit-break.jpg

23:54 amalloy: cool. can someone else take over listening to technomancy complain about elisp? i gotta go

23:54 technomancy: but I haven't even gotten to the part about mutable strings yet

23:54 come baaaaaaaaaaaaaaaaaack

Logging service provided by n01se.net