#clojure log - Aug 06 2011

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

0:02 jcromartie: technomancy: thanks!

0:03 technomancy: next version you'll be able to just to M-x clojure-jack-in, but there's a bug right now that means sometimes it'll mess up if you don't kill *swank* first

0:42 fmw: I guess this wasn't a good time to ask a question

0:42 I'll try again later :)

0:45 jcromartie: fmw: sup?

0:47 fmw: jcromartie: "<fmw> are there problems with doseq in clojurescript? I'm suppose I'm just

0:47 using it incorrectly: https://gist.github.com/1128963 ? I'm getting

0:47 notified that something in core.js is undefined (see the provided paste)"

0:48 I tried applying the patch in CLJS-39 that fixes an issue with doseq scoping, but no luck

0:49 jcromartie: sorry :( no idea... kind of strange that it says "<big long string of code> is undefined"

0:50 fmw: jcromartie: yes, basically it is doing (function() [...]).call()

0:50 jcromartie: but the compiler must be doing something strange to yield that undefined message

0:50 fmw: but it says the thing its doing .call on is undefined

0:52 https://gist.github.com/1129025 there is the javascript function that is undefined (starting at line 11)

0:53 according to firebug that whole function is undefined

0:53 jcromartie: ah

0:54 basically: (function () {})().call(null,o);

0:56 fmw: jcromartie: indeed

0:56 I'm going to try and find the relevant cljs code now

0:57 and see if I can do (let-fn [fn [...]] (my-fn o)) or something to that extend

0:57 so it gets compiled differently

1:01 amalloy: fmw: letfn isn't implemented yet

1:01 (or ever? i think it will be eventually but i don't know)

1:04 fmw: amalloy: ah, thanks

1:05 amalloy: anyway, I will just (let [my-fn (fn [] ...)] ...) in that case

1:06 I can't really find where to do that for cljs.core, though (see https://gist.github.com/1129025 for the function I'm trying to let-fn on line 11)

1:09 https://github.com/clojure/clojurescript/blob/master/src/cljs/cljs/core.cljs#L126 seems to be how that JavaScript output I pasted is generated, but I can't figure out where to change it

1:09 the implementation, that is, to work around the undefined bug

1:12 amalloy: fmw: the line you linked to doesn't even have any code. it's just a protocol definition

1:12 the generated code was from the (or) macro

1:13 &(macroexpand '(or x y))

1:15 fmw: amalloy: I wonder where that is called, though.

1:15 amalloy: it seems to figure out the type of protocol and call the seq implementantation for it

1:17 amalloy: so maybe I'm looking in the wrong file and I should check where defprotocol is implemented, to find that (or) macro

1:18 amalloy: ah, https://github.com/clojure/clojurescript/blob/master/src/clj/cljs/core.clj#L113 I think that the or macro is there

1:22 yes, its that method I linked that produces the function that is undefined according to both chrome and firefox (https://gist.github.com/1129025)

1:23 s/method/defprotocol implementation

1:23 lazybot: <fmw> yes, its that defprotocol implementation I linked that produces the function that is undefined according to both chrome and firefox (https://gist.github.com/1129025)

2:27 jcromartie: is there a good way to do something with a timeout?

2:27 like, don't let a function run more than some amount of time?

2:28 tomoj: you don't happen to need async i/o or event driven stuff, do you? :)

2:28 lamina has very cool timeouts but wouldn't be worth including just for that

2:33 jcromartie: no

2:34 although what would be even better would be a time-bound lazy sequnce

5:34 zoldar: hello, I'm trying to use ztellman's gloss library for serializing/deserializing data. My attempt mostly follows the basic header usage example given on the wiki page but encoding attempt fails: http://paste.lisp.org/display/123831 . What am I missing?

5:44 fantazo: hi, how do you actually import in clojure all classes of a package? like in java java.awt.*?

5:47 zoldar: fantazo, afaik there's no way to do this

5:47 fantazo: ok, thank you

6:19 hmm. in scheme you normally do the simple counter example with (define (make-counter) (let ((x 0)) (lambda () (set! x (+ x 1)) x))) as clojure works with non-mutable structures in default, how do you something trivial like that? I know, this is actually not functional style.

6:25 zoldar: fantazo: (let [entity-counter (atom 0)] (defn gen-entity-id [] (swap! entity-counter inc))

6:29 fantazo: is ok, but doesn't yet do what the scheme version does.

6:33 but this version does what I want: (defn make-counter [] (let [x (atom 0)] (fn [] (swap! x inc))))

6:36 zoldar: fantazo, yup, sorry, mindless copy-paste from my old code

6:37 fantazo: clojure is interesting, it helps to get over the non-functional coding habits.

6:37 or atleast seeing them as that.

6:37 zoldar: added benefit: clojure version of that counter if thread-safe

6:41 fantazo: with clojure java is finally useable.

9:29 pyr: when i define symbols in macros

9:29 with the symbol# idiom

9:29 they only live within the context of the expanded code, i can do no manipulation on them, right ?

9:31 for instance if one of these holds a vector, I can't do (fn argvector# ~@body)

9:35 raek: pyr: a macro is basically function that returns code in the form of clojure data structures

9:35 a symbol is just a variable name

9:35 clojurebot: c'est bon!

9:36 raek: when the returned code is compiled, there might be a variable with that name, and only when that code is actually executed it hols a value

9:36 thorwil: pyr: in other words, you can do "manipulations" on them

9:40 Vinzent: but (fn argvector# ~@body) won't work anyway

9:43 pyr: Vinzent: yep

9:43 got it

9:43 thorwil: Vinzent: because of the @, or is there an additional problem?

9:43 pyr: thorwil: because (fn somevec (do-something)) doesn't work

9:44 thorwil: and because there's no way to replace argvector# with it's value

9:44 Vinzent: yep

9:44 pyr: alright then, no worries

9:45 malkomalko: is there a more idiomatic way to add up the values in deeply nested maps then reduce combined with assoc/merge?

9:45 raek: pyr: yes, it doesn't have a value at that point. but your macro can emit code that does the operation on the value

9:45 paul__: hey

9:47 (clojure.contrib.str-utils/re-split #"\s+" "sdf sdf sdf")

9:47 i get an error with the above code :(

9:48 my swank auto completion can't find contrib.str-utils

9:48 raek: paul__: first, there is clojure.string since clojure 1.2

9:48 Vinzent: and re-split is in core

9:48 raek: paul__: have you required the namespace?

9:48 it isn't loaded until you have require'd or use'd it

9:48 paul__: i thought that i didn't need to do that when i write out the whole path

9:50 dnolen: paul__: the library needs to be loaded first, it's not loaded by default, and that won't automatically load it.

9:50 ,(require '[clojure.string :as string])

9:50 clojurebot: nil

9:50 dnolen: ,(string/split "sdf sdf sdf" #"\s+")

9:50 clojurebot: ["sdf" "sdf" "sdf"]

9:52 paul__: thanks

9:52 i wanted re-split because it has reverse args

9:55 dnolen: ,(let [flip (fn [f] (fn [& args] (apply f (reverse args))))] ((flip string/split) #"\s+" "sdf sdf sdf"))

9:55 clojurebot: ["sdf" "sdf" "sdf"]

9:56 raek: skynda

9:57 eh, wrong channel

9:57 paul__: cool

10:36 ejackson: hmm... Google Closure doesn't seem to have a datagrid widget.

10:37 an emoji picker, yes.... datagrid, sorry.

10:49 abedra: technomancy, does @hiredman's RT.load fix solve your issues?

10:49 I am reviewing the patch right now

10:50 I'm trying to collect patches and get them checked in so we can cut a beta2 soon

10:51 everything looks reasonable in the patch, but I just wanted to make sure it resolves the issue you were having

11:14 triyo: Is there a filter-like function that can actually return two sequences, the difference based on predicate?

11:18 thorwil: triyo: you mean one list with matches and one with non-matches?

11:18 triyo: oops, I mean a vector with two elements. [pass-pred-test-seq, fail-pred-test-seq]

11:19 thorwil: I see there is something simple in clojure.contrib called seperate

11:20 source -> [ (filter f s), (filter (complement f) s) ]

11:23 thorwil: sure, though the whole filtering work is done twice

11:23 triyo: Hmm, thats quadratic time right?

11:23 hehe

11:23 ;)

11:23 n^3

11:23 oops

11:23 I mean n^2

11:24 O(n^2)

11:24 thorwil: if you can make sense of (source filter), you could do a trivial modification to get your result, i guess

11:25 triyo: yup, I'll come back to efficiency question after I finish the implementation.

11:25 that seems the most logical to me.

11:26 Question is also what size of seq I'll be dealing with

11:26 thorwil: yes, premature optimization and all that ...

11:31 abedra: triyo, remember that contrib is deprecated

11:32 and won't be supported on 1.3

11:32 this isn't a big deal in your case since the function you are working on is pretty simple.

11:32 but it is worth remembering

11:33 triyo: abedra: yup, I know, thanks for the reminder. I did a search on clojuredocs.org and came across the implementation. I just wanted to look at the source.

11:33 abedra: triyo, ok good deal

11:33 triyo: "he function you are working on is pretty simple." -> exctlly

11:44 malkomalko: maybe I'm missing something, what's the best way to test if something is a number?

11:45 Vinzent: ,number?

11:45 clojurebot: #<core$number_QMARK_ clojure.core$number_QMARK_@70cf21>

11:45 Vinzent: ,(number? 5)

11:45 clojurebot: true

11:47 malkomalko: ahh, that's not on the api cheatsheet

11:50 Vinzent: hm, indeed (and here http://clojuredocs.org/quickref/clojure%20core too)

11:56 abedra: malkomalko, The best way to learn is to crack open core.clj

11:56 put a marker in the code and take a look at one or two functions a day and try them out

11:57 malkomalko: makes sense

11:57 abedra: pretty soon you will have a great understanding of the language

11:57 malkomalko: the most difficult part of clojure is realizing how much is already there for you

11:57 abedra: it

11:57 yes

11:57 malkomalko: and trying to re-implement things that already exist

11:57 abedra: yes

11:57 that is why I encourage people to look inside

11:57 the language is evolving

11:57 cheat sheets really won't cut it

11:58 malkomalko: take for example http://clojure.github.com/clojure-contrib/map-utils-api.html deep-merge-with

11:58 abedra: the source is really easy to examine

11:58 malkomalko: love that function, just found it, exactly what I needed

11:58 abedra: malkomalko, but that is clojure-contrib

11:58 not clojure

11:58 malkomalko: I know

11:58 abedra: and contrib is being deprecated

11:58 malkomalko: in 1.3?

11:58 clojurebot: I don't understand.

11:58 abedra: yes

11:59 in 1.3

11:59 malkomalko: what are they planning to do with those libraries? require things in as needed instead of having a clojure.contrib?

12:00 abedra: github.com/clojure

12:00 libraries from inside of contrib have been pulled in as their own projects

12:00 chouser: and/or will be

12:00 <-- culprit

12:01 malkomalko: is the plan to have all ported over? or are some going to not make the cut?

12:01 abedra: malkomalko, a lot of libraries went unmaintained

12:01 no some will not make it

12:01 malkomalko: that's understandable

12:01 abedra: those that have gone stale

12:01 malkomalko, here's the current plan http://dev.clojure.org/display/design/Contrib+Library+Names

12:01 malkomalko: well chouser wrote map-utils so I hope deep-merge-with makes it in :)

12:02 abedra: malkomalko, that's up to @chouser :)

12:02 malkomalko: that's very helpful

12:02 chouser: I wrote map utils?

12:02 malkomalko: well, either your name is on it

12:03 abedra: malkomalko, the idea is that any library from contrib can get it's own project as long as there is a maintainer

12:03 it does not have to be the original author

12:03 it does, however, fall under the same CA policies as Clojure

12:03 malkomalko: alright

12:03 is there a good post detailing the gameplan for 1.2 -> 1.3?

12:03 abedra: the link i just posted

12:03 it gives the current strategy

12:03 chouser: malkomalko: quite possible I did and have forgotten. :-P

12:04 abedra: for contrib

12:04 malkomalko: not for contrib libs

12:04 for core

12:04 abedra: there has been lots of mailing list posts

12:04 malkomalko: yah yah, no worries :P

12:04 ok, I'll check the google group

12:04 abedra: quite a bit has changed

12:04 and there's a beta1 that you can grab

12:04 or if you want to just run off of master

12:05 http://dev.clojure.org/display/doc/Maven+Settings+and+Repositories

12:05 if you add the sonatype repository to your project.clj file or pom.xml you can pull in 1.3.0-master-SNAPSHOT

12:05 which should become beta2 soon

12:06 just a few more patches to get in

12:06 malkomalko: very helpful.. I'll go look around to figure out the bigger changes

12:06 abedra: cool

12:07 numerics, defrecord, :dynamic, bindings, all kinds of things have been modified

12:07 malkomalko: and very good idea to just read core.clj

12:08 abedra: yes

12:08 chouser, I just put up a patch to make BigInt math much faster for long sized values inside of BigInts

12:08 chouser: cool

12:10 malkomalko: also, the joy of clojure rocks

12:10 so, thanks for that :)

12:11 abedra: yes, that book is awesome

12:13 malkomalko: reading the tests seem very handy too, no brainer way to pick things up

12:14 chouser: ah, sure! glad you like it.

12:15 abedra: chouser, I hope you guys get the Jolt!

12:17 ejackson: does anybody know how to determine which events a Closure Library widget supports ?

12:21 chouser: huh, Jolt? Hm.

12:21 jsoftw: .

12:22 lucian: ejackson: don't the closure docs say such things

12:22 ?

12:22 ejackson: lucian: to which docs do you refer ?

12:23 lucian: ejackson: it has no docs

12:23 ?

12:23 bah

12:23 ejackson: it does

12:23 for instance: http://closure-library.googlecode.com/svn/docs/class_goog_ui_BidiInput.html

12:23 thought you might have others

12:24 but these don't give me the events it can produce

12:24 lucian: hmm

12:25 i don't know of a way to introspect event listeners in js

12:25 uh, rather which events one can listen to

12:38 momox: how do i read (filter (fn [x] (every? #(% x) filters)) coll) ? I am particularly confused on the #(% x) portion...

12:42 ejackson: it depends on what filters contains

12:43 momox: ejackson: it contains a sequence of functions

12:43 filters [#(> % 1) #(< % 5)]]

12:43 ejackson: ok

12:44 in that case it filters coll, returning a seq for which applying every predicate in filters to the elements is true

12:45 the (every? ...) bit runs each function in filters against x, and return true if each function application returns true

12:45 the outer filter will return only those x

12:46 in coll for which that inner operation is true

12:47 momox: but i don't know what #(% x) expands into...i know x will be the element in coll but what is %?

12:48 ejackson: oh, % will be each function in filters in turn

12:48 you see, x gets bound to an element of coll

12:48 then every? runs the #(% x) function for each element of filters, and in each case % is bound to one of those functions

12:49 the problem you might have is that you've got % meaning two things, apparently :)

12:50 as far as every? is concerned its a function, but that function is itself written in terms of #(> % 1), so in the inner case the % will be bound to the argument THAT function was passed, being x.

12:50 foncusing indeed

12:52 gotta hop, cheers all.

13:33 pyr: is there a way to update a var

13:33 as in, change it's meta data for instance

13:34 Vinzent: alter-var-root to update value, and something similar for meta :)

13:35 pyr: alter-var-root!

13:35 thanks

13:50 arohner: pyr: there's also vary-meta for metadata specifically

14:00 pyr: ok, alter-meta! was the one i was looking for

14:00 ultimately

14:09 crazyFox: Hi. Ive got a question about (lazy) seqs. Anybody there to answerß

14:14 Vinzent: crazyFox, don't ask to ask, just ask :)

14:14 crazyFox: alright. coming up...

14:17 im trying to do (take 10 (cons 2 primes)) in the repl. wheres primes is an infinite lazy seq. The thing is: the expression doesnt evaluate. No result shows up. No new prompt shows up. What am i doing wrongß

14:19 Vinzent: ,(take 10 (cons 2 (range)))

14:19 clojurebot: (2 0 1 2 3 ...)

14:19 Vinzent: crazyFox, probably the problem is in definition of primes

14:20 crazyFox: its the one from clojure.contrib.lazy-seqs. actually i just tried again. it worked...

14:21 omg. it was a paren mistake. bugged me yesterday already. i just realize now. ok all is fine. dont be bothered.

14:24 Vinzent: crazyFox, :) you should consider using paredit

14:25 crazyFox: i know. still have to set up emacs or some other decent editor with built-in repl.

14:32 zvrba: Vinzent: i looked at paredit and it introduced a bunch of hard to remember keyboard shortcuts :(

14:35 Vinzent: zvrba, I use only 5: C-M-i,k for killing sexps, C-M-j,l for navigating and C-M-u for going up level

14:36 (M-j,l,i,k for moving cursor)

14:36 very handy

14:50 leeda: Is this the best way to write this: (last (take-while #(>= x %) cutoffs)) ? (cutoffs is a vector of integers in strictly increasing order)

14:54 actually what I want is: (- (count (take-while #(>= 15 %) cutoffs)) 1)

14:54 (the index in the vector)

15:04 crazyFox: leeda: so, u want the index of the largest number <= 15 in a vector with strictly increasing order, is that rightß

15:05 leeda: crazyFox: yeah. (15 actually would be any integer)

15:09 crazyFox: leeda: i would do (dec (take-while #(>= limit %) yourvec)). im not too firm with the clojure libraries though. still a newcomer... :)

15:09 leeda: i forgot the count...

15:10 leeda: crazyFox: oh yeah ok. so same thing but just use dec instead of #(- % 1), good idea

15:10 MasseR: Didn't read the entire conversation, but I just learned about 'indexed' which might be of use

15:10 crazyFox: leeda: yes

15:11 MasseR: http://clojuredocs.org/clojure_contrib/clojure.contrib.seq/indexed

15:11 (indexed "foo") would return [1 \f 2 \o 3 \o]

15:11 leeda: hm i think that would just make it more complicted

15:18 jsnikeris: Is it idiomatic clojure to throw exceptions?

15:18 thorwil: leeda: just for fun: (reduce #(if (>= 20 %2) (inc %1) %1) 0 [1 5 9 12 15 20 34])

15:19 leeda: thorwil: nice

15:20 jsnikeris: For example, I have a function that attempts to save a blog entry. If a similar entry already exists, I don't want to overwrite it, and I want to communicate to the caller why the entry wasn't saved. Would this be the right place to throw an exception?

15:21 Vinzent: MasseR, vtw, there is map-indexed and keep-indexed in core

15:21 *btw

15:22 jsnikeris, I think if you don't have to ineroperate, you should prefer more flexible exception system over plain java exceptions

15:22 jsnikeris: Vinzent: What would you suggest?

15:23 zvrba: jsnikeris: why not return an error code?

15:23 crazyFox: leeda: there is faster solutions though, but much harder to write

15:24 leeda: crazyFox: yeah i'm sure.

15:25 Vinzent: jsnikeris, I can't remember the name... try to search scigiraldi try+ on github. also, raek have written such lib. There is error-kit in the 1.2 contrib.

15:29 jsnikeris: Vinzent: looks like there is clojure.contrib.condition by Stephen C. Gilardi

15:30 Vinzent: jsnikeris, yeah, and also c.c.condition. (first two libs was for 1.3)

15:33 crazyFox: is clojuredocs.org linked from clojure.org? i didnt know about it up to now. its neat! better than the doc on github

15:44 mjg123: If I have a list of functions (f1 f2 f3... ) and a list of values (v1 v2 v3...) what's a good way to create a list like ((f1 v1) (f2 v2) (f3 v3) ... ) ?

15:45 as in, to apply each function to its corresponding value?

15:45 crazyFox: did anybody try jswat-4.5 with clojure-1.2.1? Does it work?

15:48 MasseR: mjg123: Maybe (map #(%1 %2) funs vals)

15:48 or (map apply ..

15:50 (Both are untested)

15:52 mjg123: MasseR, the first does me fine, thanks!

15:53 thorwil: ,(reduce #(if (integer? %2) [(conj (first %1) %2) (second %1)] [(first %1) (conj (second %1) %2)]) [[] []] [1 "a" 2 "b" 3 4])

15:53 clojurebot: [[1 2 3 4] ["a" "b"]]

15:53 thorwil: too bad triyo isn't around anymore ^^

15:58 ,(partition 2 (interleave ["a" "b"] [1 2]))

15:58 clojurebot: (("a" 1) ("b" 2))

15:58 thorwil: ^ mjg123

16:03 crazyFox: ,(map #(%1 %2) '(+ - * /) (range 1 5))

16:03 clojurebot: (nil nil nil nil)

16:03 crazyFox: why does it not give results i expect?

16:04 Vinzent: ,(map #(%1 %2) [+ - * /] (range 1 5))

16:04 clojurebot: (1 -2 3 1/4)

16:04 crazyFox: so whats wrong with the quoted list?

16:04 Vinzent: ,(#('+ %) 1)

16:04 clojurebot: nil

16:05 Vinzent: ,('+ {'+ :ok})

16:05 crazyFox: ok. makes sense. ty

16:05 clojurebot: :ok

16:06 Vinzent: crazyFox, so symbols acts like keywords when treated like a functions

16:06 pdk: (key map) is a common idiom for value lookup in maps

16:07 it's not a matter of the symbol being in function position, it's that it's quoted and the second argument is a map

16:07 crazyFox: yeah. knew that. its cool

16:09 lucian: ,(take 10 (iterate + 1))

16:09 clojurebot: (1 1 1 1 1 ...)

16:10 crazyFox: ,(take 10 (iterate inc 1))

16:10 clojurebot: (1 2 3 4 5 ...)

16:11 Vinzent: pdk, but it's quoted = it's a symbol in that case :)

16:11 crazyFox: ,(+ 1)

16:11 clojurebot: 1

16:11 pdk: exactly

16:11 lobotomy: it would be kind of cool if (=) worked

16:12 crazyFox: ,(take 10 (iterate #(+ 1 %) 1))

16:12 clojurebot: (1 2 3 4 5 ...)

16:12 Vinzent: also, only keywords and symbols work that way?

16:12 lucian: crazyFox: hmm

16:13 Vinzent: lobotomy, hm, should it return nil or true?

16:14 lobotomy: dunno :)

16:19 crazyFox: my Firefox just crashed. Its also a crazy fox... :)

16:51 lucashansen: hello world

16:55 I am just starting with Clojure. What is the best way to use it with vim?

16:56 crazyFox: did u have a look at http://dev.clojure.org/display/doc/Getting+Started+with+Vim already?

16:57 markskil1eck: lucashansen: :! emacs

16:58 ;D

16:58 lucashansen: I'd seen that plugin but I wasn't sure if it was the complete experience. Whenever I look up lisp on vim, people tell me to use emacs ha ha

16:59 markskil1eck: I love Vim, so I do. But it felt as though I was fighting Vim to get some Lisp support, so I gave up. Emacs offered a better immediate experience.

16:59 Slimv was... alright.

17:23 crazyFox: did anybody try clooj? (https://github.com/arthuredelstein/clooj) Is it worthwhile?

17:25 micrypt: crazyFox: Some screenshots would've been nice.

17:27 crazyFox: try this http://dev.clojure.org/display/doc/getting+started+with+Clooj

17:31 wastrel: clooj fancy

17:39 crazyFox: so u tried it?

17:47 jsnikeris: How do I access a class created with :gen-class? For example, I want to check if this class is being thrown: http://tinyurl.com/3bvsd3s. So I have (require '[clojure.contrib.condition :as cond]) and I want to write something like: (is (thrown? cond/Condition (fn-where-it-might-be-thrown)))

17:48 Do I always have to write out the whole ns: clojure.contrib.condition.Condition

17:55 amalloy: jsnikeris: you can't require/as the class, because it's a class, not a function. but you can import it, just like any other class

17:58 jsnikeris: amalloy: ahh, import. I knew I was missing something. Thanks

17:59 zoldar: when using leinigen, can I point to a certain git repository as a dependency ?

18:00 arohner: zoldar: kind of. read the 'checkouts' section of the lein faq

18:01 jsnikeris: amalloy: so I added, (:import [clojure.contrib.condition.Condition]) to my ns declaration. I expected to be able to reference a bare Condition in my code, but I'm getting an Unable to resolve classname: Condition exception. Am I missing something?

18:01 amalloy: jsnikeris: you probably don't want the []s

18:02 zoldar: arohner, thanks, this will do

18:03 jsnikeris: amalloy: that did it. Thanks

19:29 zmaril: What's the function that lets me iterate over two lists at the same time element by element? Not for, but something that iterates index by index

19:29 crazyFox: interleave ?

19:30 zmaril: Not interleave. Can you use map on two collections at once?

19:30 crazyFox: ,(interleave [:a :b :c] [1 2 3])

19:30 clojurebot: (:a 1 :b 2 :c ...)

19:31 crazyFox: yes, if the supplied function can deal with two args

19:31 zmaril: Nice! Thanks

19:31 crazyFox: ,(map str "abc" "123")

19:31 clojurebot: ("a1" "b2" "c3")

19:32 amalloy: map

19:41 Pupeno: Apparently I became the clojars most valuable programmer: https://www.masterbranch.com/developer-achievements?uid=1566 lolo

19:47 amalloy: man, the recent attempts to make "social" programmer profiles that just scrape their data from github are pretty unimpressive

20:14 pschorf: I've defined a macro which is a wrapper around defrecord, but records defined using it cannot be imported to another ns

20:14 has anyone else seen this behavior?

20:19 amalloy: pschorf: then your macro is broken

20:24 zoldar: (apply hash-map (reduce concat (SomeRecord. a b c))) -> is there some more succint way of converting record to pure map?

20:27 amalloy: zoldar: if you want to do that, why are you using records at all?

20:27 (that said, (into {} (SomeRecord. a b c)) is nicer)

20:28 zoldar: amalloy, I'm using gloss library for data serialization and at the same time I do some java interop - gloss works with maps but doesn't play nicely with records

20:29 veered: Sorry for the noob question, but I can't seem to get the REPL to work on osx. It runs and evals ok, but I can't use the arrow keys. They come up as "^[[C" etc.

20:29 zoldar: always forget about "into" ..

20:29 pschorf: amalloy, it turns out it was due to a - in the lib name...for the import, i had to replace it with a _

20:30 veered, you need to use JLine or something similar

20:31 MicahElliott: veered: I just installed leiningen: https://github.com/technomancy/leiningen

20:31 and it nicely gives history editing, via realine AFAICT.

20:32 So I'm even doing vi-mode editing in repl. Very nice.

20:33 Which gets to my question...

20:33 It's pretty annoying to kill repl with ctrl-c.

20:33 I found this old thread: https://groups.google.com/d/topic/clojure/MuT4kwBFYRg/discussion

20:33 But it's not addressed. Anyone know how to enable ^C without killing repl?

20:37 amalloy: veered: or rlwrap: http://stackoverflow.com/questions/5639502/one-repl-to-bind-them-all

20:37 veered: MicahElliott: Thanks that did the trick. I can now work on finishing my first Clojure tutorial!

20:43 MicahElliott: amalloy: rlwrap is awesome! thx

20:55 amalloy: MicahElliott: you don't say what you mean by "enabling C-c", but if you mean interrupting a single command while leaving the repl running, that's not possible in general. java doesn't have a portable way to handle signals like SIGINT

20:57 MicahElliott: amalloy: Yes, that's exactly what I mean. Bummer.

21:56 jimblomo: is there a way to define a global variable in lein? eg dev-mode/debug?

22:06 Nanarpuss: is anybody using clojurescript?

22:10 jimblomo: a bit

22:12 Nanarpuss: I'm having a hell of a time getting my timer callbacks to work. I'd like to have one timer update the counter on 6 divs

22:13 jimblomo: you're using goog.Time/callOnce?

22:13 Timer

22:13 Nanarpuss: no (. timer (start))

22:15 it only updates the last div

22:21 jimblomo: so are you setting up a event listener?

22:21 Nanarpuss: yeah, hold on I might have something

23:21 jweiss: i'm having trouble using proxy to subclass a java class - getting 'no matching ctor found'. there are several ctors of the same arity so i assume i need a type hint. but that doesn't seem to help: http://paste.lisp.org/display/123845#1

23:22 tried adding ^ThreadFactory before the (reify ...) - no help

23:27 amalloy: $google javadoc threadpoolexecutor

23:27 lazybot: [ThreadPoolExecutor (Java 2 Platform SE 5.0)] http://download.oracle.com/javase/1,5.0/docs/api/java/util/concurrent/ThreadPoolExecutor.html

23:29 amalloy: jweiss: what version of clojure? could be int/long confusion maybe?

23:29 jweiss: amalloy: 1.2

23:29 what i am not clear on with type hints - do you need just enough hints to narrow the choices down to 1?

23:30 or do you need to hint everything? or something in between?

23:30 hinting the ThreadFactory in 6th position should narrow it to 1

23:31 ataggart: jweiss: it's all or nothing to resolve reflection

23:31 jweiss: ataggart: so i have to hint all the arguments?

23:31 ataggart: there are improvements pending for Reflector, but they've been pushed back another release

23:31 yarp

23:31 jweiss: ah, that would explain it, thanks

23:32 ataggart: not needing that is one of the many fixes I made for Reflector

23:32 but it's too big of a change to fit into 1.3.0

23:33 right now I'm trying to figure out why clojurescript is utterly broken

23:33 seancorfield: we're close to the next milestone build i'd hope? (of 1.3.0)

23:34 jweiss: ataggart: how do i hint an int or long?

23:34 ^int 2 gives me "Metadata can only be applied to IMetas"

23:35 ataggart: well, you only need to type hint variables, not literals

23:36 bortreb: I want to make something that behaves like vectors, but has the additional property of either being "up" or "down". What's the best way or going about this?

23:36 ataggart: so, it would be mor correct to say, all args need to be known to resolve reflection

23:36 either by typehinting or literal

23:37 jweiss: ataggart: expressions need to be hinted i assume

23:37 ataggart: if the expression does not have a known type, yes.

23:38 jweiss: ataggart: hm, i'm not sure - it's a call to reify.

23:38 should have a known type, the interface i'm reifying :)

23:38 ataggart: need more information

23:38 gist it

23:38 jweiss: http://paste.lisp.org/display/123845#1

23:39 not sure which i need to hint. tried the ^TimeUnit, ^BlockingQueue, and ^ThreadFactory, no help

23:40 amalloy: jweiss: reify may actually not have a known type, because you can reify multiple interfaces and the tagging system can only hold one class at a time

23:40 jweiss: amalloy: i see. not sure what's missing then, maybe forgot an import or something - checking

23:41 jimblomo: can i use clojure forms in my lein defproject? eg. an (if) form?

23:43 jweiss: aha - ataggart amalloy - (int 2) (int 2) (long 0) fixed it

23:43 thanks for your help!

23:43 * amalloy claims the prize for his int/long confusion guess

23:44 jweiss: good knowledge about type hinting there, not sure i've ever seen that explicitly stated in the docs

23:44 ataggart: odd. IIRC the reflector automatically coerces ints to/from longs

23:45 amalloy: ataggart: it's supposed to, but i wouldn't be surprised if there are edge cases

23:45 ataggart: ah wait, no, the reflector is dumb. it can't make widening conversions

23:45 that was the main driver for my changes

23:45 and 1.2.0 those literals are ints

23:46 * ataggart really wishes someone on core took my improvements seriously

23:49 amalloy: really? it only makes narrowing conversions? i guess that makes a little sense given 1.2's automatic widening of arithmetic

23:50 but widening is the conversion it *should* do, isn't it?

23:50 like, the direction that java goes

23:52 ataggart: amalloy: the method finding code does not account for widening conversions. The calling of methods in the copmiler *does* though.

23:53 amalloy: i don't understand the distinction

23:53 doesn't it have to find a method to call it?

23:54 ataggart: At compile-time Compiler asks Reflector for a Method object given a collection of args. If an arg is an int, and all the methods take longs, Reflector returns null.

23:54 *if there is more than one method for a given name

23:56 assuming a Method was returned, the Compiler emits bytecode to call it, and that's when it does the long/int autocoerce

23:56 so, in the case of only one named method that takes a long, you could call it with an int and it'd work.

23:56 amalloy: and if no method is found, then it emits bytecode to do runtime reflection, which does widen? i see

23:56 ataggart: yarp

23:57 though that only works for int/long and float/double

23:57 the other integral types don't get converted

23:57 sadly

23:57 amalloy: right

23:58 ataggart: but again, I fixed all this months ago

23:58 it's sitting in jira

23:58 amalloy: but short and byte are sorta a joke anyway. who uses them

23:58 short, anyway

23:58 ataggart: yeah, i remember

23:58 ataggart: people who need to work on byte arrays

23:58 https://github.com/ataggart/codec/blob/master/src/codec/base64.clj

Logging service provided by n01se.net