#clojure – 2010-11-09

2010-11-09

00:04	replaca	Raynes: are you here?
00:05	Raynes	replaca: No.
00:05	Raynes	Yes, I'm here. :)
00:05	replaca	Raynes: got your pull req on autodoc, but I'm kind of moving in a different direction with autodoc
00:06	Raynes	I didn't send a pull request. :o
00:06	replaca	Raynes: I'm on the verge of pulling the lein plugin out into a separate plugin and here's why:
00:06	Raynes	replaca: Oh, my message.
00:06	replaca	is my senile memory messing with me?
00:06	replaca	yeah, that :)
00:06	Raynes	replaca: That's completely cool. I'm fine with cake-autodoc being separate. I just wanted to check with you to make sure I was being consistent.
00:07	replaca	when you have leinigen, autodoc and a project all running with possibly different versions of clojure
00:07	replaca	you get version headaches out the wazzoo
00:07	replaca	we had it with 1.2 and we'll have it with 1.3
00:08	replaca	so a very light external plugin that depends on autodoc and invokes it with args based on the project params seems like the way to go
00:09	replaca	I hope to get to that soon for leiningen and you could just copy me, but there are a lot of things I hope to get to soon :)
00:09	replaca	Raynes: that's all!
00:10	Raynes	replaca: Awesome.
01:06	lenw	hi all
01:06	scottj	new library for webapps, feedback welcome: https://github.com/scottjad/slice/blob/master/src/slice/example.clj
01:09	lenw	scottj: that looks very neat
01:42	quizme	how do you do a hard reload, where it reloads the file and recompiles?
01:42	scottj	C-c C-k in slime
01:42	lenw	(use 'x :reload)
01:42	quizme	I'm using (load) but it doesn't seem to be recompiling
01:42	quizme	lenw i'll try that
01:42	quizme	thnx
01:45	quizme	when i reload I'm getting this error: java.lang.IllegalStateException: flatten already refers to: #'clojure.contrib.seq/flatten in namespace: triangle.core (core.clj:1)
01:45	quizme	so it's preventing it from reloading
01:47	quizme	is there a way to ignore that warning?
01:47	quizme	error ?
01:50	lenw	are you not getting that error the first time ?
01:56	quizme	lenw: it's a warning the first time
01:59	LauJensen	Morning all
02:01	ppppaul	moo
02:03	Derander	lenw: I have the same issue as quizme w/ my setup
02:04	lenw	i am not really sure about whats going on there guys - what is the warning that you get the first time - perhaps that the thing to look at ?
02:05	quizme	WARNING: flatten already refers to: #'clojure.core/flatten in namespace: triangle.core, being replaced by: #'clojure.contrib.seq/flatten
02:11	lenw	what does your uses bit look like in your code ?
02:13	quizme	lenw: http://pastie.org/1283802
02:21	lenw	ok i can reproduce the error now
02:23	quizme	lenw: great, we're starting a movement.
02:28	lenw	http://clojuredocs.org/clojure_contrib/clojure.contrib.seq/flatten ?
02:29	lenw	use this rather http://clojuredocs.org/clojure_core/clojure.core/flatten
02:50	quizme	lenw: I took out the reference to clojure.contrib.seq and it seems to have removed that warning.
02:50	quizme	lenw: hope it doesn't break anything though...
02:51	lenw	i dont think that the behaviour of flatten has changed
02:55	quizme	lenw: ok awesome thanks.
03:20	raek	awesome! someone on the mail list used Clojure to solve a Myst puzzle... :D
04:01	npoektop	hi! i've just read a post about how to make server pick up changes without restarting repl: http://blog.darevay.com/2010/11/compojure-the-repl-and-vars/. There is an example what happens if we do not use var, but i do not get it. Why it prints hi, not hey, when the function greet is redefined?
04:04	raek	in code, the vars are used. in data, the var value is used.
04:04	raek	(fn [] (foo)) ;; <--- this will always call the most recent version of foo
04:05	raek	(some-fn foo) ;; <-- here, the value of foo is looked up, and the function that the var foo currently has is passed as the arg
04:09	raek	so, in the example, partial receives the current value of greet (the function that prints "Hi!")
04:12	raek	when a function is compiled (like the call function in the example) the functions is uses are not looked up, but references to the vars are remembered instead.
04:15	npoektop	raek: i thought there may be only one version of a function. It's a bit confusing. Do i get it right? When i do defn, it just binds an anonymous function to symbol, and if i redefine it, it just rebinds. So there really may be versions of functions.
04:16	npoektop	rebinds to a new anonymous function
04:16	raek	yes, when you reevaluate the defn, you rebind the var to a new function object
04:17	npoektop	reak: cool. thanks
04:17	raek	the old function object remains unchanged
04:19	raek	I usually start the web server using something like this:
04:19	raek	(def web-server (atom nil))
04:20	raek	(defn start [] (swap! #(do (assert (nil? %)) (run-jetty #'the-handler {:port 8080, :join? false}))))
04:20	raek	sorry, "... (swap! web-server #(do..."
04:21	jowag	if I have a function (defn my-fun [a b c] (foo a) (bar b) (baz c)), what is the best way to document what kind of arguments it accepts? In this case one needs to look into source or doc for foo, bar and baz to know the types of arguments, so some kind of documenting the arguments in my-fun would be helpful
04:22	jowag	some partial solutions could be - choose a good argument name, - document it briefly in the docstring, - add a comment explaining arguments, - use metadata/type hints
04:23	npoektop	jowag: http://tech.puredanger.com/2010/02/08/clojure-1-whats-up-doc/
04:26	jowag	npoektop: so you recommend to use docstring to document the argument types?
04:29	npoektop	jowag: sorry. i posted before i read your second line.
04:30	raek	i would write that in the docstring, along with the explanations of what the args are for
04:31	raek	you should always have docstrings for public fns
04:35	raek	ideally, the fn and namespace docstrings should contain everything the user needs to know in order to use the fn
04:53	jowag	I agree that docstring should contain everything user needs
04:53	jowag	but sometimes I find myself writing too large docstring if I want to describe fully what the parameters should be
04:56	jowag	in one case my function expects map with specified keys present and docstring is than a bit verbose
04:57	Raynes	I wrote excellent documentation for my secret-but-not-so-secret-that-you-cant-find-it-on-github-if-you-look-for-it project. I'm proud of myself.
05:00	jowag	I came from strongly typed land and the feeling that the only way how to elucidate arguments is in the docstring makes me a bit uneasy
05:06	jrp	hm, when I try to run a simple hello world function from a repl inside vim, i get '#<user$hello user$hello@3b26456a>' or similar when i try to run it
05:06	jrp	whats going on here?
05:10	jjido	jrp: where's your code?
05:11	jrp	jjido: In the vim window below it. I send it to the repl with vimclojure I think...
05:14	jjido_	jrp: please post the code on github or pastebin
05:15	jrp	jjido_: (defn hello [] println ("Word"))
05:17	jjido	ok, then the output is the result of your code. You define a function.
05:17	jjido	try: (defn hello [] (println "Word")) (hello)
05:18	jrp	hm ok, that works. Why cant I call it via the repl though once Ive defined it?
05:19	jjido	this is a function call: (hello)
05:19	jjido	the repl will forget the function when you restart it
05:20	jrp	but Im not restarting the repl?
05:23	bartj	I see a very weird behaviour with clojure.contrib.json
05:23	bartj	Here is the code sample: http://pastie.org/1283998
05:24	bartj	can anyone please take a look?
05:27	bartj	is this a bug?
05:28	bartj	, (type (json-str (Math/sqrt -1)))
05:28	clojurebot	java.lang.Exception: Unable to resolve symbol: json-str in this context
05:28	bartj	, (type (clojure.contrib.json.json-str (Math/sqrt -1)))
05:28	clojurebot	java.lang.ClassNotFoundException: clojure.contrib.json.json-str
05:29	bartj	ie. (type (json-str (Math/sqrt -1))) = (type (json-str "abc"))
05:30	bartj	but, when I do a read-json on this: (read-json (json-str (Math/sqrt -1))) it fails!
07:19	lenw	how do i see the queue's behind an agent ?
07:55	cemerick	lenw: it's in a package-private field `aq` in the agent
07:56	Raynes	cemerick: Mornin'!
07:56	cemerick	There's roughly no good reason to look at the queue, though.
07:56	cemerick	Raynes: Morning :-)
07:57	jcromartie	hi
07:58	Raynes	jcromartie: Hi.
07:58	jcromartie	I have to confess... Smalltalk (Pharo) is winning me over.
07:58	Raynes	I played with Pharo a while back.
07:58	jcromartie	It's like the opposite of Clojure.
07:58	jcromartie	It is basically just a toolkit for mutable state.
07:58	Raynes	I do declare that I cannot tolerate trying to work in an 'environment'.
07:59	jcromartie	heh, yeah it's weird
07:59	jcromartie	but I wonder if something like Seaside wouldn't be possible in Clojure?
07:59	Raynes	My environment is whatever I want it to be at any particular moment in time, not what somebody decided beforehand.
07:59	jcromartie	hey, at least it's consistent right?
07:59	Raynes	That it is.
08:00	Raynes	No "Help, I can't get SLIME working"
08:00	jcromartie	Seaside is just hands-down the best web development framework I've ever seen (and I've evaluated a lot)
08:00	jcromartie	what does Clojure have in terms of continuations?
08:00	Raynes	cemerick: I own raynes.me now. I'm getting all fogus_-like.
08:01	jcromartie	could fns be used to build continuations?
08:02	cemerick	Raynes: nice :-)
08:02	cemerick	BTW, what is "Raynes", anyway?
08:02	Raynes	amalloy asked me that earlier.
08:03	Raynes	I started using it as a username after playing the bloodrayne games several years ago. It started as Rayne, and then became 'Raynes' when I started running into (everything) where Rayne is already taken.
08:03	Raynes	I am become Raynes.
08:08	Raynes	cemerick: I remember people calling me Raynes at the conj. It always shocked me back to reality about where I was, because it was so bizarre.
08:20	Raynes	stuartsierra: I tawt I taw a blue haired Clojurian. I did! I did!
08:21	stuartsierra	What's up, doc?
08:21	stuartsierra	(I forget what the cat says.)
08:21	Raynes	That's bugs bunny.
08:29	cemerick	stuartsierra: I think you were looking for…sufferin' succotash! :-D
08:29	cemerick	of course, severe lisps don't translate in irc
08:30	stuartsierra	I only program in Severe Lisps.
08:30	stuartsierra	Other lisps are too milquetoast for me.
08:31	cemerick	"milquetoast" is a fantastic word
08:31	Raynes	$dict milquetoast
08:31	sexpbot	Raynes: noun: a timid, unassertive man or boy fearful of confrontation and easily manipulated and dominated.
08:31	stuartsierra	That's going to be the name of my next project then.
08:33	cemerick	the etymology of it is pretty amusing
08:36	cninja	Hi all! Im new to clojure. As part of teaching myself the language, I was looking at clojure.core and it seems to me the implementation of "distinct" is overly complex. I wrote https://gist.github.com/669075 that seems to be faster. What is it missing?
08:37	scottj_	jcromartie: see clj-cont and weblocks for what's possible.
08:37	Raynes	$source distinct
08:37	sexpbot	distinct is http://is.gd/gRO3Q
08:40	cninja	I thought maybe my faster version held onto the head of the column, but I'm not familar enough with the language to test for that sort of thing.
08:45	chouser	cninja: yours isn't lazy after the first step
08:46	Raynes	chouser: You're a ninja.
08:46	chouser	no, c is.
08:46	chouser	har har
08:46	cninja	ha ha
08:46	Raynes	I actually didn't make that association when I saw his nickname.
08:46	cninja	How did you test for that?
08:46	Raynes	I guess that's a sign that I've been awake too long.
08:47	chouser	cninja: still working on a test, but from your code it's clear enough to make a bold assertion on IRC. :-)
08:48	Raynes	chouser: Write me a map function that works like map, but when it comes across a collection, it maps into that collection and always returns the same data structures. It must not blow the stack. I want it on my desk by 9.
08:48	cninja	chouser: I figured I missed something. It didn't make since that I just rewrote a core function and made it twice as fast
08:48	chouser	cninja: :-)
08:49	chouser	& (take 2 (distinct (map #(do (prn :took %) %) (list 1 2 3 1 2 3 4))))
08:49	sexpbot	⟹ (:took 1:took 21 2)
08:49	chouser	cninja: now, try that with distinct2
08:50	cninja	k
08:50	stuartsierra	Raynes: clojure.walk/postwalk
08:51	Raynes	&(doc clojure.walk/postwalk)
08:51	sexpbot	⟹ "([f form]); Performs a depth-first, post-order traversal of form. Calls f on each sub-form, uses f's return value in place of the original. Recognizes all Clojure data structures except sorted-map-by. Consumes seqs as with doall."
08:55	Raynes	stuartsierra: You're my bestest buddy ever.
08:59	Raynes	My life is so much easier now.
09:00	chouser	contrib generic has multimethod-based solutions
09:00	chouser	((fn gmap [f coll] (clojure.contrib.generic.collection/into (empty coll) (map #(if (coll? %) (gmap f %) (f %)) coll))) inc '[1 2 3 [4 5 6] 7 8 (9 10 11)])
09:01	chouser	hm, that's not quite right
09:01	chouser	(12 11 10)
09:03	fliebel	Oh, I really want a function-need identification graph.
09:05	Raynes	I find it amusing that Leiningen takes so much memory that I cannot run the 'deps' task with sexpbot running at the same time on my VPS.
09:07	Raynes	Not that it's Leiningen's fault. sexpbot isn't exactly lightweight. My VPS only has a gig of guaranteed RAM with 2gb burst.
09:07	Raynes	I should be moving him to a higher end box soon.
09:09	fliebel	Raynes: How hard would it be for me to implement pingpong timing? So me "Raynes: ping" you "fliebel: pong" sexpbot "Raynes's responds time id 5 sec"
09:10	Raynes	fliebel: Isn't that irrelevant when you can just use CTCP PING?
09:10	Raynes	* Ping reply from fliebel: 0.84 second(s)
09:10	jjido	irc messaging ping
09:11	fliebel	Raynes: It's not about my computers response time, but just to see how long it takes someone to respond ;)
09:11	Raynes	Oh, I misread what you said.
09:11	Raynes	Meh, shouldn't be very difficult.
09:12	Raynes	I need to write a plugin tutorial.
09:13	Raynes	People seem to get by pretty well just by looking at existing plugins.
09:13	fliebel	Raynes: That would be awesome! I looked at one and it didn't seem that hard, but there where a few magic bits.
09:14	Raynes	I'd be happy to explain those magic bits. In PM or #clojure-casual, though. Don't want to fill up #clojure with sexpbotspam. :p
09:15	fliebel	Raynes: I didn't know about casual...
09:15	Raynes	It's a very, very quiet channel.
09:16	jjido	#clojure is casual enough
09:16	Raynes	jjido: That's exactly right. Most of the time it is.
09:16	fliebel	Raynes: I'm going to look at a plugin and I'll be there with questions in a moment.
09:16	Raynes	I like to reserve #clojure-causal for totally non-clojure related stuff, or just long discussions about anything that just doesn't seem quite appropriate for #clojure.
09:19	romainp	Hi, what are a few good sources of clojure code to use as inspiration? (not domain-specific, just to get the feel of it)
09:20	belun	clojuredocs.org
09:20	belun	http://clojuredocs.org/
09:22	belun	i got a q too :)
09:22	belun	if there is let
09:22	belun	and letfn
09:22	belun	is there a mxin function
09:22	belun	?
09:22	belun	that can bind vars and functions ?
09:24	chouser	belun: no, but you can use let to bind fns. letfn is only needed when the fns are mutually-recursive, a generally pretty rare use case.
09:24	stuartsierra	~seen dysinger
09:24	clojurebot	Cool story bro.
09:24	belun	what does muttualy recursive mean ?
09:25	stuartsierra	$seen dysinger
09:25	sexpbot	dysinger was last seen quitting 21 hours, 27 minutes ago.
09:25	joegallo	A calls B, B calls A, A calls B, etc...
09:25	chouser	belun: the functions call each other, so at least one of them is calling a function that hasn't yet been defined.
09:26	jjido	belun: if you don't need mutually recursive let is fine
09:26	belun	that's the only usecase for let fn ? i was thinking to use it to give names to inline functions
09:26	belun	for readability...
09:27	chouser	,(letfn [(odd [x] (if (zero? x) false (even (dec x)))) (even [x] (if (zero? x) true (odd (dec x))))] (odd 5))
09:27	clojurebot	true
09:27	chouser	belun: you can give any fn a name for readability
09:28	chouser	& (let [cool (fn cool [a b] (+ a b))] (cool 4 5))
09:28	sexpbot	⟹ 9
09:29	belun	ya but this way cool stays in the namespace...
09:30	jjido	belun: no it does not go in global nakedness
09:30	jjido	lol iPhone autocorrect
09:33	chouser	neither let nor letfn nor fn change any namespace
09:36	jjido_	was I kicked? it was an autocorrect mistake
09:36	chouser	you weren't kicked
09:36	Raynes	Nobody kicked you.
09:36	Raynes	We <3 you.
09:37	jjido_	lol
09:45	belun	fn does not put functions into namespace ? what does then ?
09:45	chouser	dean: def-things. defn, defmacro, etc
09:45	chouser	belun: ^^^ (sorry dean)
09:47	Raynes	'fn' creates a function. A function is just an object. def things give these things names. For example, (def x (fn [])) is the same as (defn x []). These don't do anything magical, they just create functions with fn and bind them to a name.
10:04	fliebel	Where in Joy of Clojure are reify, defprotcol and deftype covered?
10:06	fliebel	ah, 9.3
10:06	Raynes	1.1.4, 9.3
10:11	fliebel	&(= (array-map 1 2 3 4) (hash-map 1 2 3 4)) ; what is the difference?
10:11	sexpbot	⟹ true
10:12	cemerick	fliebel: you should use the latter :-)
10:13	chouser	really?
10:13	jjido	no
10:13	cemerick	or, actually (into {} …)
10:13	chouser	& (hash-map 3 4 1 2)
10:13	sexpbot	⟹ {1 2, 3 4}
10:13	chouser	& (array-map 3 4 1 2)
10:13	sexpbot	⟹ {3 4, 1 2}
10:13	jjido	that is a map indexed by integers, is that what you want?
10:14	cemerick	chouser: you use array-map explicitly on occasion?
10:14	fliebel	jjido: Not particularly… Just to lazy to type :a 1 :b 2
10:14	chouser	cemerick: absolutely
10:15	fliebel	cemerick: I used it in a case where I was doing some strange apply thing that did not work with hash-map, not knowing what I was doing exactly.
10:16	fliebel	$(type {})
10:16	fliebel	&(type {})
10:16	sexpbot	⟹ clojure.lang.PersistentArrayMap
10:16	jjido	&(get (hash-map 1 2 3 4) 3)
10:16	sexpbot	⟹ 4
10:16	cemerick	chouser: not (into {} …)?
10:17	chouser	cemerick: array-map guarantees order. I do on occasion use that.
10:17	cemerick	huh
10:17	cemerick	I'd be too paranoid about the resulting map getting too big at some point, and tripping over into PHM
10:18	fliebel	chouser: Is that the only difference?
10:18	chouser	cemerick: right, you can't conj onto it and assume it'll still be an array-map
10:19	chouser	but, array-map always returns an array map, so as long as you use only that particular value, you're good to go.
10:22	fliebel	I still don't get the difference though, could you show me an example where it matters if I'm using one or the other?
10:22	chouser	I did.
10:23	chouser	& [(hash-map 3 4 1 2) (array-map 3 4 1 2)]
10:23	sexpbot	⟹ [{1 2, 3 4} {3 4, 1 2}]
10:24	fliebel	chouser: But why does that matter? get is going to return the same thing for both, right? or are you thinking about looping over it?
10:24	jjido	if you map or loop over it the order differs
10:25	cemerick	fliebel: also FWIW, equality isn't a very good indication of the natures of respective data structures. e.g.:
10:25	cemerick	&(= {1 2 3 4} (doto (java.util.HashMap.) (.put 1 2) (.put 3 4)))
10:25	sexpbot	⟹ true
10:25	Raynes	My heart stops every time something gets evaluated.
10:25	Raynes	And then I smile.
10:35	bartj	In a map like this: (def x {:a [1 2 3] :b [] :c [4 5] :d []})
10:35	bartj	Is there a better way to remove :b and :d (ie. they have empty vectors/collections)
10:35	bartj	than this: (select-keys x (remove nil? (map #(when (> (count (x %)) 0) %) (keys x))))
10:37	jjido	empty? tells you if the collection is empty
10:38	Raynes	&(into {} (filter (comp seq second) {:a [1 2 3] :b [] :c [4 5] :d []}))
10:38	sexpbot	⟹ {:a [1 2 3], :c [4 5]}
10:39	jjido	what does seq do here?
10:40	jjido	,(seq [])
10:40	clojurebot	nil
10:41	Raynes	&(seq [])
10:41	sexpbot	⟹ nil
10:41	clojurebot	⟹ "Returns the metadata of obj, returns nil if there is no metadata."
10:41	Raynes	If seq is called on an empty collection, it returns nil.
10:41	jjido	and nil is interpreted as false... almost obfuscation
10:41	mfex	&(into {} (remove (comp empty? val) {:a [1 2 3] :b [] :c [4 5] :d []}))
10:41	sexpbot	⟹ {:a [1 2 3], :c [4 5]}
10:42	Raynes	Right, val works here as well. I should have used val.
10:42	Raynes	Vals have more fun.
10:43	bartj	mfex, Raynes thanks!
10:44	bartj	&(remove (comp empty? val) {:a [1 2 3] :b []))
10:44	sexpbot	java.lang.Exception: Unmatched delimiter: )
10:44	bartj	&(remove (comp empty? val) {:a [1 2 3] :b [])
10:44	sexpbot	java.lang.Exception: Unmatched delimiter: )
10:44	Raynes	Missing a }
10:44	bartj	&(remove (comp empty? val) {:a [1 2 3] :b []})
10:44	sexpbot	⟹ ([:a [1 2 3]])
10:46	bartj	for a moment, I thought that the [:a [1 2 3]] was a vector!
10:46	Raynes	Well, it pretty much is.
10:46	Raynes	&(type (remove (comp empty? val) {:a [1 2 3] :b []}))
10:46	sexpbot	⟹ clojure.lang.LazySeq
10:46	jjido	bartj: yeah that is how clojure handles entries
10:46	Raynes	&(type (first (remove (comp empty? val) {:a [1 2 3] :b []})))
10:46	sexpbot	⟹ clojure.lang.MapEntry
10:47	Raynes	A MapEntry is similar to a vector.
10:48	bartj	cool
10:48	jjido	,(first {:a [1 2 3]})
10:48	clojurebot	[:a [1 2 3]]
10:48	bartj	, (into {} [[:a 1] [:b 2]])
10:48	clojurebot	{:a 1, :b 2}
10:50	bartj	cool, thanks everyone
10:53	fliebel	hrm, I thought there was a chapter like "when to use reify", but looking at it, those are for atom, agent, ref and these. I guess I'll just have to think hard about my problem ang do some hamocking afterwards.
10:56	jjido	the hammock joke is still making rounds ;)
10:58	bartj	jjido, care to explain?
11:00	belun	i think it;s from teh rich preentation
11:00	belun	http://www.livescribe.com/cgi-bin/WebObjects/LDApp.woa/wa/MLSOverviewPage?sid=R4XrwCpdzj5m
11:00	fliebel	bartj: Something to do with Clojure-conj. Rich gave a talk about problem solving, I think, and he said something about hamock time.
11:00	fliebel	to slow again.
11:02	belun	:)
11:03	LauJensen	Most dev communities are plauged by insomnia, long hours of coding. Looks like Rich solved that for #clojure :)
11:04	bartj	fliebel, belun thanks!
11:09	fliebel	ouch, this hurts… Python has only 68 core functions, Clojure over 400.
11:10	jjido	fliebel: I did notice clojure has many functions in core
11:10	Raynes	fliebel: clojure.core covers a very wide range of tasks. The entire sequence library is stuffed in there, for example.
11:10	fliebel	jjido: I'll be doing a project about this after the hamock.
11:11	fliebel	Raynes: I know. That is my whole point. Pythons functions are not very useful on their own at all, so you need to look in the corret module, which makes searching a lot easier.
11:11	fliebel	*correct
11:12	Raynes	I tend to agree, but I very much like having so much available to me all the time.
11:13	jjido	Raynes: it would not do much harm to do one more (use) at the top
11:13	fliebel	Raynes: Wait and see, it'll be awesome :) But for now: hammock time.
11:13	Raynes	jjido: Yeah, but not necessarily just one.
11:13	jjido	agreed
11:15	LauJensen	tomoj: Hows that screencast coming along?
11:16	scottj_	uses are a pain. I'd hate to start every file of with uses clojure.agents clojure.seq clojure.files clojure.files.extra clojure.futures
11:17	Raynes	scottj_: Exactly.
11:21	bartj	Is this a bug in clojure.contrib.json or am I just using it inappropriately? - http://pastie.org/1283998
11:21	bartj	I mean with the read-json function
11:21	bartj	can anyone please have a look at the paste ?
11:27	vibrant	ook, so i'm defining a macro in namespace a which produces code using functions from that namespace
11:27	vibrant	and when i try to call the macro in namespace b, it says that it can't resolve a given symbol (that function name)
11:27	vibrant	is there a correct way to do it?
11:28	LauJensen	bartj: IIRC clj-json on github is the preferred way to go. Should be 2.5x faster IIRC
11:28	jjido	vibrant: don't use a macro?
11:28	bartj	LauJensen, thanks a bunch, but what is IIRC ?
11:28	LauJensen	If I Remember Correctly
11:29	bartj	LauJensen, will check it out; but, the issue is not of speed, I think there is a bug?
11:30	bartj	LauJensen, to put it lightly, what goes in, doesn't come out the same
11:30	LauJensen	bartj: Not sure. I never use JSON.
11:31	bartj	LauJensen, may I ask what do you use?
11:32	LauJensen	bartj: Well it depends a lot of what Im doing. If its just data to/from a browser I use the regular {key: val} notation. I've seen people use JSON is non-web contexts as well. Would never do that.
11:33	bartj	LauJensen, interesting...if you had to use Redis, and had to store mutliple values in a hash map, would you still not serialize the values to json?
11:33	LauJensen	Ask me again once Ive implemented my first Redis solution
11:34	LauJensen	afk
11:35	bartj	forget Redis
11:35	bartj	if you had to serialize mutliple values in a hash key ?
11:36	vibrant	why is there no sequential binding clause?
11:36	vibrant	or is there?
11:37	scottj_	vibrant, do you really mean binding or actually let?
11:37	vibrant	let is sequential i mean binding
11:38	scottj_	vibrant: re macro, are the functions fully qualified after expansion?
11:38	vibrant	scottj; nope.
11:39	scottj_	vibrant: bc you're quoting them?
11:39	vibrant	http://pastebin.ca/1986444
11:40	vibrant	cmd-* functions reside in another namespace which is :used in the one in which the macro is
11:41	scottj_	maybe wrap (symbol...) in resolve
11:41	Derander	so many functions to learn in clojure
11:41	Derander	fun fun fun
11:41	opqdonut	you must mean fn fn fn
11:45	vibrant	scottj; now it says 'can't call nil'
11:46	scottj_	update pastebin
11:48	Derander	When should I use "use" vs "require"?
11:49	scottj_	the bot must have an answer for that faq
11:49	Derander	http://stackoverflow.com/questions/871997/use-vs-require-in-clojure <-- I should have just googled.
11:56	tonyl	morning
11:58	amalloy	morning for real, tonyl
11:59	tonyl	amalloy: i'm in a conference, early morning. now some free time
12:00	amalloy	tonyl: fun. maybe erc?
12:00	fliebel	Is there something like exclusive or other than the bitwise one?
12:01	tonyl	amalloy: don't know emacs :( i'm using irssi
12:01	technomancy	they have gui IRC clients now?
12:01	amalloy	awww, no emacs! your life is so sad
12:02	amalloy	technomancy: the irc guis all collect identification data so that the command-line folks will have no trouble wiping us out when the revolution comes
12:02	tonyl	amalloy: I am learning it... slowly
12:02	hsuh	hi guys... so i have this hairy macro which isnt working (doesnt it always start like this?)... http://pastie.org/1284771
12:03	hsuh	the "argmap" can't be used like that, i tried using argmap# but that didn't work too, although it would give another error
12:03	hsuh	i'm wondering if i'm on the right direction or macros like this will just bring pain to me instead of making the rest of the code simpler
12:03	clojurebot	Barking spiders!
12:04	amalloy	argmap# is the right first step
12:04	amalloy	after that, time to see what the real error is
12:05	hsuh	why? one would think that since argmap is quoted there would be no complaint?
12:05	scottj_	fliebel: don't think so. like this? (defn xor [x y] (= x (not y)))
12:05	amalloy	`(let [argmap 1])
12:05	amalloy	&`(let [argmap 1])
12:05	sexpbot	⟹ (clojure.core/let [clojure.core/argmap 1])
12:05	amalloy	&(let [clojure.core/argmap 1])
12:05	sexpbot	java.lang.Exception: Can't let qualified name: clojure.core/argmap
12:05	fliebel	scottj_: Yea, think so...
12:05	amalloy	hsuh: ^^
12:06	amalloy	`(syntax-quote) is not the same as '(quote)
12:06	hsuh	amalloy: what are you saying that clojure is different from elisp ?
12:07	amalloy	hsuh: say it ain't so! :P
12:07	hsuh	ok with argmap# i have a much more interesting error, now at "runtime"
12:07	amalloy	it also looks like ~(args-names args) is a problem
12:08	amalloy	it will be saying you can't call the first element of args-names as a function
12:08	hsuh	yeah something like that, java.lang.String cannot be cast to clojure.lang.IFn
12:08	hsuh	nice clojure evaluator you got there
12:09	amalloy	?
12:09	hsuh	in your head.. because i totally missed that...
12:09	amalloy	haha
12:09	amalloy	well
12:09	amalloy	i had the exact same issue last night
12:09	amalloy	so...
12:09	amalloy	&(let [args '("foo" "bar")] `(blah ~(args)))
12:09	sexpbot	java.lang.ClassCastException: clojure.lang.PersistentList cannot be cast to clojure.lang.IFn
12:09	amalloy	&(let [args '("foo" "bar")] `(blah ~args))
12:09	sexpbot	⟹ (clojure.core/blah ("foo" "bar"))
12:10	amalloy	hsuh: see what ~args is expanding to?
12:10	amalloy	&(let [args '("foo" "bar")] `(blah ~(vec args)))
12:10	sexpbot	⟹ (clojure.core/blah ["foo" "bar"])
12:10	hsuh	gimme two days
12:10	amalloy	this is more like what you meant
12:21	amalloy	hsuh: working now?
12:22	hsuh	amalloy: err.. almost
12:24	amalloy	oh. you want ~@body
12:25	hsuh	ok, but the error is still on that like, the args-name thing
12:25	amalloy	and you changed it to use (vec args-names)? i don't see that in the paste
12:25	hsuh	i didn't changed the paste, i actually moved 'vec' inside args-names
12:27	Chousuke	what's your macroexpansion at the moment?
12:29	hsuh	i posted a little simpler version
12:29	hsuh	http://pastie.org/1284828
12:29	Tordmor	~pasge
12:29	clojurebot	I don't understand.
12:29	Tordmor	~paste
12:29	clojurebot	lisppaste8, url
12:30	Tordmor	hm
12:30	hsuh	while i get amalloy example, i can't translate it to what is wrong there
12:30	hsuh	oh and the error is, when trying to eval (test) i get java.lang.String cannot be cast to clojure.lang.IFn (like before...)
12:30	Chousuke	hsuh: I don't understand at all what you're trying to do
12:31	Tordmor	anybody knows why this won't quit, after printing my packages list? http://pastie.org/1284834
12:31	Chousuke	but in (map name (vec args)) the vec is completely redundant
12:31	hsuh	hm
12:32	Chousuke	args-names is still returning a list
12:32	Chousuke	or a seq
12:32	Chousuke	which means a function call
12:32	hsuh	ok now it works...
12:32	hsuh	what do you mean by "whcih means a function call" ?
12:32	Chousuke	a list form is a function call.
12:33	Chousuke	your macro emits a list, therefore it's evaluated as a function call
12:33	hsuh	of course, now it makes sense
12:33	hsuh	tks amalloy, Chousuke
12:34	hsuh	its kinda like '[1 2 3] can be understood as [1 2 3], but '(1 2 3) cant be seen as (1 2 3), right?
12:34	Chousuke	hsuh: hm, well
12:35	Chousuke	,'(1 2 3) :P
12:35	clojurebot	(1 2 3)
12:35	Chousuke	the thing is, when you evaluate (1 2 3) it's treated as a function call
12:35	amalloy	sure: &\|[1 2 3]\|& is okay, but &\|(1 2 3)\|& is no good, and the latter is what you were running after macroexpansion
12:35	sexpbot	[1 2 3] ⟹ [1 2 3]
12:35	sexpbot	(1 2 3) java.lang.ClassCastException: java.lang.Integer cannot be cast to clojure.lang.IFn
12:36	Chousuke	note though that there's a difference between '[foo bar] and [foo bar] :P
12:36	Chousuke	the former is a vector if symbols, and the latter is a vector of whatever foo and bar evaluate to
12:37	hsuh	hm, true... forget what i said
12:40	Chousuke	macros become a lot easier after you understand the whole "code is data" thing
12:41	Chousuke	and the evaluation semantics of the different data elements used.
12:42	hsuh	amalloy: strangely it only works with ~body instead of ~@body
12:42	Chousuke	hsuh: that's because you only accept a single expression as the body
12:43	Chousuke	You should destructure decl as [args & body]
12:43	hsuh	beautiful
12:43	amalloy	also if won't let you splice
12:44	amalloy	i mean, it will, but won't do what you want
12:44	amalloy	you should probably use when here
12:44	Chousuke	or you can just do (.. [name args & body] ...)
12:45	Chousuke	or you can just use do
12:45	Chousuke	for the if
12:46	hsuh	Chousuke: there is an else part to that if, is just that you are seeing the second version i think... let me post the working version
12:47	hsuh	does seeing a "with-meta" on the macroexpansion means anything special about my code?
12:48	Chousuke	I'm not sure
12:48	Chousuke	do you intend for the macro to expand to code that attaches metadata to something?
12:48	hsuh	no
12:48	hsuh	i'm just afraid of reflection because i read its eevil
12:48	Chousuke	then it's probably wrong .)
12:48	amalloy	hsuh: metadata has nothing to do with reflection
12:49	hsuh	i also dont knew that too of course
12:49	Chousuke	amalloy: :tag metadata is used to avoid it.
12:49	hsuh	pastie is slow, or i am
12:49	Chousuke	for symbols
12:49	Chousuke	and lists
12:49	hsuh	here is the working thing: http://pastie.org/1284771
12:49	amalloy	Chousuke: okay, fair enough
12:49	hsuh	the idea is that defn* will help me write dozens of similar functions for a webapp
12:49	hsuh	without repeating code... is that a fair use of macros?
12:50	amalloy	that's pretty much what macros are for
12:51	Chousuke	seems fine.
12:51	Chousuke	I'm not seeing with-metas either :P
12:52	hsuh	so i'm not using macroexpand right
12:52	hsuh	check the withMeta on http://pastie.org/1284771... am i doing something wrong?
12:53	hsuh	oh i'm on slime.. that might be it ?
12:54	Chousuke	hsuh: that's defn, not your macro :)
12:54	hsuh	but i want to expand an usage of the macro, or the macro? :)
12:54	Chousuke	anyway, it's fine
12:54	Chousuke	your macro expands to a defn which expands to something .withMeta
12:55	hsuh	heh
12:55	jfields	do dosyncs retry infinitely?
12:55	amalloy	jfields: no
12:55	amalloy	but they try quite a few times
12:55	Chousuke	10000, or something :P
12:55	jfields	amalloy, any docs on the web? I'm looking in the Joy of Clojure and not seeing any details
12:55	hsuh	Chousuke: hm, but my test function is almost mempty..
12:56	amalloy	jfields: that's funny, JoC is where i remember learning this. i'll take a look
12:56	Chousuke	hsuh: yes? the expansion looks correct to mee
12:56	Chousuke	-e
12:56	hsuh	Chousuke: what about the withMeta?
12:56	Chousuke	hsuh: that's correct too
12:57	Chousuke	hsuh: it's just adding metadata to the resulting function
12:57	hsuh	oh
12:57	Chousuke	hsuh: it's what defn does
12:57	hsuh	ohhh
12:57	hsuh	ok ok
12:58	amalloy	jfields: check out 11.1.4
12:59	amalloy	"First, there are transaction restart limits..."
12:59	hsuh	to be sincere though, if your patience allows, i still don't get why i can't use argmap without gensym... since that is just "literal text" that should appear on the expanded form
12:59	jfields	amalloy, thanks
12:59	amalloy	hsuh: no it isn't
13:00	Chousuke	hsuh: it's not text :)
13:00	Chousuke	it's a symbol
13:00	Chousuke	hsuh: and it might conflict with local bindings
13:00	amalloy	if you used exactly the same name every time, there would be problems with nesting. someone else might use that same name, or you might step on a local
13:00	amalloy	or you might even expend your macro recursively, stepping on your own local
13:01	Chousuke	what are you going to do if someone does (let [argmap foo] (defn* test [a b c] (somestuff argmap)))?
13:01	hsuh	see it know
13:02	hsuh	now
13:02	amalloy	hsuh: clojure makes it very hard to make that mistake on accident, but there are occasions when you actually want to do it
13:02	amalloy	see the google group for an example
13:06	hsuh	using a macro like this i have to make sure that functions used by the macro (the expanded defn) are acessible where the macro is used... is there any alternative to this or just live with it ?
13:07	kotarak	hsuh: use a foo.internal namespace which is :use'd in foo (where your macro lives) or @#'foo/shout-at-the-private-var long enough.
13:14	hsuh	kotarak: tks
13:35	kotarak	Someone ever used expect on Windows? Does anyone still use expect anyway?
13:54	fogus_	pdlogan: Are you the author of the blog http://patricklogan.blogspot.com ?
13:55	pdlogan	fogus_: yep
13:55	fogus_	pdlogan: One of my favorite blogs of all time. I miss it.
13:56	pdlogan	hmm... thanks. I've been thinking about posting a new one re: some clojure / web stuff I've been dinking with.
13:57	fogus_	Well, here's me hoping you do.
13:57	pdlogan	(n.b. my views on clojure have changed a good bit since I was blogging there)
13:58	fogus_	for better or worse. ;-)
13:59	pdlogan	(better)
14:03	pppaul	(inc better)
14:03	sexpbot	⟹ 1
14:03	fliebel	When I wrote this function, only God and I knew what I was doing. Now only God knows. https://github.com/pepijndevos/clj-fneeid/blob/master/src/fneeid.clj
14:04	pppaul	weird
14:04	pppaul	(doc resfn)
14:04	clojurebot	It's greek to me.
14:06	amalloy	fliebel: having a truth parameter is bizarre, since you just use (not) anyway
14:07	amalloy	which always returns either true or false
14:07	fliebel	amalloy: How is it bizarre?
14:08	fliebel	amalloy: What I'm doing is that if truth is true, I "filter" while otherwise I "remove"
14:08	amalloy	ugh. just make your users pass in (complement f) instead of f; it's more readable on both sides of the function call
14:09	fliebel	amalloy: Doesn't work the same… function is determined on the outer fn, while truth is defined on the inner fn.
14:12	fliebel	amalloy: Depending on how you call the inner fn it gives you one or the other half of input, which is passed around through either positive or negative.
14:12	pppaul	,(clojure.contrib.math/round (rand 10))
14:12	clojurebot	java.lang.ClassNotFoundException: clojure.contrib.math
14:13	amalloy	&(rand-int 10)
14:13	sexpbot	⟹ 5
14:13	kotarak	,(require 'clojure.contrib.math)
14:13	clojurebot	java.io.FileNotFoundException: Could not locate clojure/contrib/math__init.class or clojure/contrib/math.clj on classpath:
14:16	amalloy	fliebel: i'm not going to try to understand why you want your function to do this. but you could make it more readable, i think, by using (filter (comp #{truth} boolean function)), since you seem to just be using (not) to coerce things to booleans
14:18	fliebel	amalloy: I'm doing sort of a conditional complement, so I could do (if truth fuction (complement function))
14:19	amalloy	yep, you could do that too
14:19	amalloy	which is probably nicer since it doesn't involve booleans where they don't belong :P
14:22	pppaul	,(repeatedly 40 #(rand-int 11))
14:22	clojurebot	(10 6 1 4 10 6 6 4 5 4 ...)
14:22	amalloy	&(repeatedly 40 #(rand-int 11)) :)
14:22	sexpbot	⟹ (3 7 7 5 6 4 1 9 5 5 10 10 10 8 2 2 10 8 1 4 3 7 2 0 10 1 2 3 5 4 9 3 10 4 0 4 8 5 0 7)
14:27	pppaul	&(repeatedly 400 #(rand-int 11))
14:27	sexpbot	⟹ (4 8 10 6 10 5 8 3 3 5 8 9 4 9 5 8 1 3 6 8 2 10 10 8 3 6 9 8 9 5 5 5 2 9 1 10 8 0 7 2 0 2 8 4 1 8 0 0 4 1 10 7 10 6 6 8 4 9 4 10 10 4 1 6 1 6 1 10 6 1 6 3 9 2 9 6 6 10 10 0 10 6 7 9 1 10 9 3 2 1 7 8 8 0 2 8 6 7 9 4 7 8 10 3 5 8 8 2 1 3 6 4 1 6 4 0 4 3 1 6 2 10 0 8 1 2 10 0 9 1 2 10 4 4 10 9 4 9 3 6 ... http://gist.github.com/669654
14:28	pppaul	sexbot is the way to go!
14:28	bobo_	im impressed!
14:28	ohpauleez	also funny that you called him sexbot, I wonder how many times that will happen
14:29	pppaul	sexbot should know his nick-name
14:30	ohpauleez	haha
14:30	amalloy	ohpauleez: it will happen a lot :P
14:31	amalloy	i have to type sex<TAB> to get into my repo; i never use the p, so if i ever need to use it in a non-tab-completion context...
14:31	pppaul	it would be cool if clojurebot has a lisp
14:32	amalloy	$sed -pppaul s/s/th/
14:32	sexpbot	<pppaul> it would be cool if clojurebot hath a lithp
14:32	amalloy	:P
14:32	pppaul	ahaha
14:35	pppaul	&(= '(1 2 3 4) '(4 3 2 1))
14:35	sexpbot	⟹ false
14:36	pppaul	i guess i have to sort to compare for equality if order doesn't matter for me?
14:37	kotarak	clojurebot: botsnack
14:37	clojurebot	thanks; that was delicious. (nom nom nom)
14:37	amalloy	pppaul: use sets
14:37	fliebel	How can I make partial retain the meta of the original fn?
14:37	pppaul	a set wouldn't work because i need duplicate entries
14:37	amalloy	&(= #{1 2 3 4} #{3 4 1 2})
14:37	sexpbot	⟹ true
14:38	pppaul	&(= #{1 3 3 3} #{3 3 1 1})
14:38	sexpbot	java.lang.IllegalArgumentException: Duplicate key: 3
14:38	kotarak	fliebel: (with-meta (partial f arg) (meta f))?
14:38	amalloy	pppaul: you could sort, or use every?/some
14:38	ohpauleez	pppaul: You can use sets, or map a set operation across both lists
14:38	fliebel	kotarak: Yea, that sort of stuff.
14:38	ohpauleez	damnit amalloy, stop jumping on my suggestions
14:38	ohpauleez	:)
14:38	kotarak	fliebel: well, that was the answer
14:38	amalloy	ohpauleez: mine isn't actually any good
14:38	pppaul	(doc every)
14:38	clojurebot	Titim gan éirí ort.
14:38	fliebel	kotarak: Thanks :)
14:39	amalloy	oh! i know the answer pppaul!
14:39	pppaul	!
14:39	pppaul	i think i need to sort
14:39	amalloy	nonono
14:39	pppaul	^_~
14:39	ohpauleez	I think you can use all, or another set operation
14:39	amalloy	&(= (map frequencies [[1 2 3 4 3] [3 4 1 3 2]])
14:39	ohpauleez	on the lists
14:39	sexpbot	java.lang.Exception: EOF while reading
14:39	amalloy	&(= (map frequencies [[1 2 3 4 3] [3 4 1 3 2]]))
14:39	sexpbot	⟹ true
14:39	pppaul	wooah
14:39	pppaul	&(doc frequencies)
14:39	sexpbot	⟹ "([coll]); Returns a map from distinct items in coll to the number of times they appear."
14:40	amalloy	gonna be much faster than sort
14:40	pppaul	sure that isn't sorting?
14:40	amalloy	although haha i typed it wrong
14:40	amalloy	&(apply = (map frequencies [[1 2 3 4 3] [3 4 1 3 2]]))
14:40	sexpbot	⟹ true
14:40	amalloy	yeah it's not doing a full sort. it's hashing
14:41	ohpauleez	yeah, that's a good one actually
14:41	amalloy	&(apply = (map frequencies [[1 2 3 4 3] [3 4 1 4 2]]))
14:41	sexpbot	⟹ false
14:42	pppaul	$(freguencies [ 1 2 2 3 3 4 5 5])
14:42	pppaul	&(freguencies [ 1 2 2 3 3 4 5 5])
14:42	sexpbot	java.lang.Exception: Unable to resolve symbol: freguencies in this context
14:42	kotarak	(defn =-by [f & items] (apply = (map f items)))
14:42	pppaul	&(frequencies [ 1 2 2 3 3 4 5 5])
14:42	sexpbot	⟹ {1 1, 2 2, 3 2, 4 1, 5 2}
14:42	pppaul	um, that is a sort
14:43	amalloy	pppaul: no it isn't. it's a hashmap
14:43	pppaul	it's like count-sort
14:43	amalloy	&(frequencies [ 1 2 2 3 3 4 5 55 1231])
14:43	sexpbot	⟹ {1 1, 2 2, 3 2, 4 1, 5 1, 55 1, 1231 1}
14:43	pppaul	or bucket-sort
14:43	amalloy	bucket-sort, if you like. but that's faster than actually sorting the list
14:43	kotarak	&(frequencies [1 2 3 4 2 1231 55])
14:43	sexpbot	⟹ {1 1, 2 2, 3 1, 4 1, 1231 1, 55 1}
14:43	pppaul	i know ^_^
14:43	amalloy	ah. thanks kotarak
14:44	pppaul	oh
14:44	amalloy	there, see? no sorting pppaul, just hashing :P
14:44	kotarak	small maps are array maps.
14:44	amalloy	oh yeah
14:44	LauJensen	large maps are sorted with rand-elt
14:44	amalloy	&(frequencies (range 2000))
14:44	sexpbot	⟹ {0 1, 1024 1, 32 1, 1056 1, 64 1, 1088 1, 96 1, 1120 1, 128 1, 1152 1, 160 1, 1184 1, 192 1, 1216 1, 224 1, 1248 1, 256 1, 1280 1, 288 1, 1312 1, 320 1, 1344 1, 352 1, 1376 1, 384 1, 1408 1, 416 1, 1440 1, 448 1, 1472 1, 480 1, 1504 1, 512 1, 1536 1, 544 1, 1568 1, 576 1, 1600 1, 608 1, 1632 1, 640 ...
14:44	amalloy	&(frequencies (range 200))
14:44	sexpbot	⟹ {0 1, 32 1, 64 1, 96 1, 128 1, 160 1, 192 1, 1 1, 33 1, 65 1, 97 1, 129 1, 161 1, 193 1, 2 1, 34 1, 66 1, 98 1, 130 1, 162 1, 194 1, 3 1, 35 1, 67 1, 99 1, 131 1, 163 1, 195 1, 4 1, 36 1, 68 1, 100 1, 132 1, 164 1, 196 1, 5 1, 37 1, 69 1, 101 1, 133 1, 165 1, 197 1, 6 1, 38 1, 70 1, 102 1, 134 1, 16... http://gist.github.com/669681
14:45	pppaul	cool
14:45	LauJensen	&(set (range 32))
14:45	sexpbot	⟹ #{0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31}
14:45	amalloy	it's "sorted" by hashcode, because it's a hashmap
14:45	LauJensen	&(set (range 33))
14:45	sexpbot	⟹ #{0 32 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31}
14:45	kotarak	&(frequencies (range 8))
14:45	sexpbot	⟹ {0 1, 1 1, 2 1, 3 1, 4 1, 5 1, 6 1, 7 1}
14:45	amalloy	LauJensen: that one never gets old :)
14:45	kotarak	&(frequencies (range 9))
14:45	sexpbot	⟹ {0 1, 1 1, 2 1, 3 1, 4 1, 5 1, 6 1, 7 1, 8 1}
14:45	pppaul	&(frequencies (repeatedly 10000 (rand-int 11)))
14:45	sexpbot	java.lang.ClassCastException: java.lang.Integer cannot be cast to clojure.lang.IFn
14:45	LauJensen	amalloy: that and hardcoded small ints and booleans
14:45	kotarak	&(frequencies (range 10))
14:45	sexpbot	⟹ {0 1, 1 1, 2 1, 3 1, 4 1, 5 1, 6 1, 7 1, 8 1, 9 1}
14:46	kotarak	bleh
14:46	amalloy	?
14:46	kotarak	8 entries is the limit IIRC.
14:46	kotarak	for array maps
14:46	amalloy	kotarak: small integers hash to themselves, so even a hashmap will look sorted
14:47	amalloy	&(frequencies (range 50 60))
14:47	sexpbot	⟹ {50 1, 51 1, 52 1, 53 1, 54 1, 55 1, 56 1, 57 1, 58 1, 59 1}
14:47	pppaul	&(frequencies (repeatedly 1000 (rand-int 11)))
14:47	sexpbot	java.lang.ClassCastException: java.lang.Integer cannot be cast to clojure.lang.IFn
14:47	pppaul	what am i doing wrong?
14:47	amalloy	&(frequencies (repeatedly 1000 #(rand-int 11)))
14:47	sexpbot	⟹ {0 105, 1 101, 2 85, 3 94, 4 71, 5 80, 6 83, 7 91, 8 99, 9 93, 10 98}
14:47	pppaul	oh yeah
14:48	pppaul	silly me
14:48	amalloy	sexpbot: you don't like 4?
14:48	amalloy	(inc 4)
14:48	sexpbot	⟹ 1
14:48	amalloy	giggle
14:49	pppaul	&(frequencies (repeatedly 100000 #(rand-int 11)))
14:49	sexpbot	⟹ {0 9216, 1 9098, 2 8977, 3 9098, 4 9005, 5 9053, 6 8979, 7 9227, 8 9041, 9 9151, 10 9155}
14:50	pppaul	so, in your solution, what's the diff of (=... and (apply =...
14:50	amalloy	&(= (range 2))
14:50	sexpbot	⟹ true
14:50	amalloy	&(apply = (range 2))
14:50	sexpbot	⟹ false
14:51	amalloy	1: is (range 2) equal to itself? yes
14:51	amalloy	2: is 0 equal to 1? no
14:52	micahmartin	What's the better name for a rspec port to clojure: 'specl' or 'speclj' (both pronounced 'speckle')?
14:52	pppaul	&(map frequencies '((1 2) (2 1)))
14:52	sexpbot	⟹ ({1 1, 2 1} {2 1, 1 1})
14:53	pppaul	&(=(map frequencies '((1 2) (2 1))))
14:53	sexpbot	⟹ true
14:53	pppaul	&(apply = (map frequencies '((1 2) (2 1))))
14:53	sexpbot	⟹ true
14:53	pppaul	i understand the reason to use apply, since we are getting back a seq
14:53	amalloy	&(=(map frequencies '((1 2) (212321))))
14:53	sexpbot	⟹ true
14:54	pppaul	why is (=... working?
14:54	amalloy	it's not
14:54	amalloy	it will always return true
14:54	pppaul	oh
14:54	pppaul	&(= 5)
14:54	sexpbot	⟹ true
14:54	pppaul	i see
14:56	amalloy	&(=)
14:56	sexpbot	java.lang.IllegalArgumentException: Wrong number of args (0) passed to: core$-EQ-
14:56	amalloy	aw
14:56	amalloy	&(+)
14:56	sexpbot	⟹ 0
15:03	pppaul	&(replicate 100 '10)
15:03	sexpbot	⟹ (10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10... http://gist.github.com/669721
15:03	pppaul	better than cycle!
15:07	fliebel	It is working! https://github.com/pepijndevos/clj-fneeid/blob/master/src/fneeid.clj Line 17 and 18 can be other questions, which can contain other questions in turn. This way you can get down to the function you need.
15:10	pppaul	my input is (1 + 2 + 3 + 4) i want my output to be (1 2 3 4) (+ + +) i've looked over the seq functions and thought (partition would do what i wanted, but i don't think so... is there something like (take with steps?
15:11	fliebel	pppaul: take-nth?
15:12	pppaul	thanks
15:13	fliebel	&((juxt #(take-nth 2 %) #(take-nth 2 (next %)) [1 2 3 4 5 6])
15:13	sexpbot	java.lang.Exception: EOF while reading
15:13	pppaul	(take-nth 2 data) & (take-nth 2 (rest data))
15:14	kotarak	&((juxt #(take-nth 2 %) #(take-nth 2 (next %))) '(1 + 2 + 3 + 4))
15:14	sexpbot	⟹ [(1 2 3 4) (+ + +)]
15:14	kotarak	&separate
15:14	sexpbot	java.lang.Exception: Unable to resolve symbol: separate in this context
15:14	pppaul	is there a reason for using next instead of rest?
15:14	fliebel	pppaul: Not unless you care about total lazyness
15:14	kotarak	pppaul: next is faster in case you know you'll need the result anyway
15:15	kotarak	pppaul: rest is required to be 100% lazy
15:15	pppaul	i thought they were almost the same cept one returned nil
15:16	kotarak	pppaul: which implies that it can't be 100% lazy.
15:16	kotarak	pppaul: it has to realise the sequence to judge whether to return nil or not.
15:16	amalloy	$source next
15:16	sexpbot	next is http://is.gd/gSsNS
15:16	amalloy	hm. i didn't realize next was faster; i just assumed it was (comp seq rest). thanks kotarak
15:17	kotarak	amalloy: using the result of next may be faster
15:17	pppaul	the details of the source is pretty hidden
15:17	pppaul	one calls more, one calls next...
15:18	kotarak	amalloy: whether next is (comp seq rest) depends on what you have. eg. for lazy-seq yes. For a seq on a vector rest is rather (or (next s) ()).
15:19	amalloy	pppaul: you can look up the java source for more and next, though that tends to be hard to track down too :P
15:19	amalloy	kotarak: huh. i guess that makes sense
15:19	kotarak	look in src/jvm/clojure/lang/ASeq.java
15:19	fliebel	What questions do you ask yourself to get to the function you need? I'm trying to come up with useful questions for https://github.com/pepijndevos/clj-fneeid/blob/master/src/fneeid.clj
15:20	kotarak	amalloy: it depends on your underlying thing being lazy or not. datastructures are not lazy.
15:20	amalloy	right
15:21	pppaul	^_^
15:21	amalloy	fliebel: what are you trying to do? write a decision tree that asks the user "are you doing x? what about y? in that case, maybe you should use pmap"?
15:22	fliebel	amalloy: Exactly :)
15:22	amalloy	you could do it like those online 20-question games
15:23	amalloy	hm, i can't find one easily
15:23	amalloy	aha! fliebel: http://www.guessthename.com/
15:26	fliebel	amalloy: That's the plan. :)
15:26	amalloy	fliebel: my point is, they don't build this ahead of time
15:27	fliebel	amalloy: huh?
15:27	amalloy	they crowdsource it: if you think of an entry they can't guess, then they prompt you for a question that would disambiguate
15:27	amalloy	and add that to the question database
15:27	amalloy	http://www.guessthename.com/movie/?pos=20387n
15:28	amalloy	so rather than building a tree yourself by hand, write a program that will ask you questions and build a tree itself
15:31	fliebel	amalloy: Then it'd have to be a web app or something.
15:32	amalloy	fliebel: my plan was for you to run it yourself at the repl to build up the db :P
15:33	amalloy	the idea being that that's easier than hard-coding the tree
15:33	fliebel	amalloy: And another problem is that their system only rules out one show per question, if I'm correct.
15:34	amalloy	no, although it does have kinda a poor database
15:36	amalloy	imagine you had a tree that looked like: (is it a sequence? {yes: (do you want a sequence as a result? {yes: map, no: reduce}), no: inc})
15:36	amalloy	then answering no to the first question eliminates both map and reduce
15:36	fliebel	amalloy: How? I tell them at the end that I was thinking about b and now a and give them a rule to define that. So depending on my answer they can rule out a or b, but there is no way to tell if c matches my question or not.
15:37	fliebel	amalloy: They haven't a tree where you can kill branches.
15:37	pppaul	&(drop 3 [1 + 2 + 3 + 4])
15:37	sexpbot	⟹ (#<core$_PLUS_ clojure.core$_PLUS_@8b8791> 3 #<core$_PLUS_ clojure.core$_PLUS_@8b8791> 4)
15:37	amalloy	fliebel: yeah, actual crowd-sourcing tends to lead to not-very-good data
15:38	KirinDave	So you all can see what a lousy clojure coder I am. ;)
15:38	amalloy	i was just proposing a way to make your development more interactive and intuitive by having the program ask you for questions. if you want to do it another way, feel free
15:38	pppaul	&(drop 3 [1 '+ 2 '+ 3 '+ 4])
15:38	sexpbot	⟹ (+ 3 + 4)
15:39	fliebel	amalloy: You have a point there.
15:39	pppaul	wisdom in crowds
15:39	pppaul	or was is crows?
15:41	amalloy	KirinDave: how can we see what a lousy coder you are? i don't see any other messages from you
15:41	KirinDave	amalloy: Oh. There is this twitter thing. But...
15:41	KirinDave	https://github.com/BankSimple/Clothesline
15:43	KirinDave	amalloy: I guess it's customary to fire off an ANN to the mailing list, huh?
15:44	amalloy	KirinDave: i dunno. i've seen some of those and it certainly wouldn't be spam, but you can invent your own customs
15:44	amalloy	but some of us don't follow twitter
15:45	KirinDave	Better people than me don't use twitter. ;)
15:45	kzar	How come 08 is an invalid number? ,08
15:45	kzar	(well I get that it's because it starts with 0 but why does that make it invalid?)
15:46	kzar	&08
15:46	sexpbot	java.lang.NumberFormatException: Invalid number: 08
15:46	amalloy	KirinDave: i don't use twitter either :P
15:46	amalloy	&010
15:46	sexpbot	⟹ 8
15:46	tlockney	kzar: because it's read as octal when you put the 0 in front
15:46	KirinDave	kzar: Here is a hit
15:46	KirinDave	,07
15:47	clojurebot	7
15:47	kzar	Oh gotya, thanks
15:47	KirinDave	,09
15:47	clojurebot	Invalid number: 09
15:47	KirinDave	(fail)
15:47	fliebel	&(read-string (str "0." "08" "e2"))
15:47	sexpbot	⟹ 8.0
15:48	kzar	Oh cool there was an easy way to strip the leading 0 in my code
15:49	amalloy	kzar: &(apply str (drop-while #{\0} "0008"))
15:49	amalloy	kzar: &\|(apply str (drop-while #{\0} "0008"))\|&
15:49	sexpbot	⟹ "8"
15:50	amalloy	hmmm. maybe i should get sexpbot to recognize nick: &expr
15:50	fliebel	amalloy: Good one :)
15:51	kzar	amalloy: and then get it to check the nick and if it's kzar fuck with me by showing the wrong results
15:51	amalloy	kzar: that code is already there
15:51	kzar	heh
15:52	amalloy	kzar: &\|(= (map range [10 20]))\|&
15:52	sexpbot	⟹ true
15:52	amalloy	see?
15:53	kzar	amalloy: I get it, you want apply =
15:53	amalloy	chuckle yes. but for a second it looks like sexpbot is lying to confuse you
15:53	fliebel	amalloy: I think that made it upstream into Clojure, it gives the same result for me on my own repl :S
15:53	amalloy	hahaha
15:53	kzar	amalloy: (apply = (map range [10 20]))
15:53	kzar	oops
15:53	kzar	amalloy: \|&(apply = (map range [10 20]))
15:53	kzar	amalloy: &\|(apply = (map range [10 20]))
15:54	fliebel	kzar: traing one as well
15:54	amalloy	kzar: &\|'[some expr]\|&
15:54	sexpbot	⟹ [some expr]
15:54	amalloy	it needs to be on both ends
15:54	kzar	amalloy: &\|(apply = (map range [10 20]))\|&
15:54	sexpbot	⟹ false
15:54	kzar	oh yea
15:54	amalloy	and yes, i know. this was a topic of discussion in #clojure an hour or two ago
15:54	fliebel	that is what i... tried to say :(
15:55	amalloy	kzar: btw, if you can think of a better way to unambiguously delimit exprs than the &\| nonsense, let me know
15:55	kzar	amalloy: How about this, if you type a sexp with no special characters the bot doesn't eval but it notices. Then if you type eval it evals it
15:55	amalloy	i will happily add it
15:56	amalloy	kzar: too hard, and also vague. everything you type is a sexp
15:56	amalloy	how does he know "sexp" wasn't something i might want to eval later? it's a valid symbol
15:57	kzar	amalloy: Well for that feature it would just worry about things in parenthesis I guess but good point
15:57	amalloy	&[(inc 1) (dec 1)]
15:57	sexpbot	⟹ [2 0]
15:57	amalloy	anyway
15:58	amalloy	but if you can think of something within a single message that's easier to type than &\| i'd love to hear it
15:58	amalloy	maybe #=expr=#
15:59	jarpiain	,`=#
15:59	clojurebot	=__147__auto__
15:59	amalloy	ah
15:59	amalloy	jarpiain: =# is legal clojure, but so is \|&. they're both things nobody is likely to type though :P
15:59	jjido	amalloy: does it have to have a closing tag?
16:00	kzar	Yea I mean if there's a end of message before the closing tag couldn't it just take the hint ?
16:00	amalloy	jjido: without it, you can't do &\|'clever\|& stuff like &\|(inc 1)\|&
16:00	sexpbot	'clever ⟹ clever
16:00	sexpbot	(inc 1) ⟹ 2
16:00	amalloy	but you're right, i could let it implicitly close at EOM
16:02	amalloy	though atm only Raynes has permission to restart sexpbot, and he's taking a nap
16:03	jjido	how about &\|('clever) &\|((inc 1))
16:04	amalloy	jjido: yeah, i can do that too
16:05	amalloy	i think i'll add ## as well, with similar shorthand forms
16:09	amalloy	jjido: ##10 ##11
16:09	buttered-toast	10 ⟹ 10
16:09	buttered-toast	11 ⟹ 11
16:10	amalloy	good? any other suggestions?
16:11	jjido	amalloy: good. I suggest skipping ## if it appears inside a matched pair of double quotes
16:11	buttered-toast	java.lang.Exception: Unable to resolve symbol: if in this context
16:11	amalloy	jjido: sadly that's kinda too hard :P
16:11	amalloy	regexes aren't that clever
16:11	jjido	(println "## hello")
16:11	buttered-toast	java.lang.Exception: Unable to resolve symbol: hello in this context
16:11	amalloy	yes, i know what you mean
16:12	amalloy	but such is life. that's why Raynes picked the super-obscure &\|'syntax\|& - to avoid any accidental matches
16:12	sexpbot	⟹ syntax
16:12	buttered-toast	⟹ syntax
16:12	amalloy	haha sorry bots
16:12	pppaul	i want to (eval (+ 1 2)) from seq (1 + 2) named 'data' i tried (eval (list (second data) (first data) (nth 3 data))) but got a cast error from lazySeq to number
16:13	amalloy	&\|(nth [1 2 3] 2)\|& &\|(nth 2 [1 2 3])\|&
16:13	sexpbot	(nth [1 2 3] 2) ⟹ 3
16:13	sexpbot	(nth 2 [1 2 3]) java.lang.ClassCastException: clojure.lang.PersistentVector cannot be cast to java.lang.Number
16:14	pppaul	i think my problem is (class +) != (class (second data))
16:14	amalloy	no pppaul, see my eval above for what your problem is
16:14	amalloy	you have two: params in wrong order; (nth n 0) == (first n)
16:15	pppaul	oh
16:15	amalloy	i think you also need to quote the + if it's not already quoted
16:15	pppaul	(eval (list (second data) (first data) (nth data 3)))
16:15	pppaul	java.lang.Exception: Unable to resolve symbol: 1 in this context (NO_SOURCE_FILE:159)
16:15	pppaul	the error changed
16:16	pppaul	&(class +)
16:16	sexpbot	⟹ clojure.core$_PLUS_
16:16	pppaul	my '+' is a symbol, though
16:16	jjido	,(re-seq #"(([^\"]\"){2})##" "(println \"## hello\")")
16:16	clojurebot	(["##" nil nil])
16:16	pppaul	&(class (quote +))
16:16	sexpbot	⟹ clojure.lang.Symbol
16:17	amalloy	&(list* (map #(nth '(1 + 2) %) [1 0 2]))
16:17	sexpbot	⟹ (+ 1 2)
16:17	pppaul	&(class (unquote (quote +)))
16:17	sexpbot	java.lang.IllegalStateException: Var clojure.core/unquote is unbound.
16:17	amalloy	jjido: and what about \\"? that really is the end of a quote
16:17	amalloy	and \\\" isn't
16:18	amalloy	it's probably possible but i think failing when ## is part of the string is an acceptable compromise for getting easy multi-eval
16:19	jjido	yes, at worse there will be more messages in the channel
16:19	amalloy	right, or someone who actually wants to write ##(println "##foo") will have to use & instead :P
16:19	amalloy	pppaul: see my list* for how i would rearrange your list
16:20	pppaul	yeah, that's what i'm doing, your's is shorter
16:20	amalloy	and eval isn't working?
16:22	cs02rm0	does anyone know why lein would complain at _everything_ (e.g. lein deps) that it is not a task?
16:22	amalloy	cs02rm0: lein self-install?
16:22	cs02rm0	amalloy: it complains that isn't a task too
16:23	amalloy	clojurebot: lein?
16:23	clojurebot	lein is http://github.com/technomancy/leiningen
16:23	pppaul	i was inputting wrong data, i changed it and debugging now...
16:23	amalloy	cs02rm0: you could get a fresh copy of lein from https://github.com/technomancy/leiningen/raw/master/bin/lein and see if that helps
16:23	raek	cs02rm0: did it work before, or is this after a fresh install?
16:24	technomancy	cs02rm0: there's a bug in 1.3 where it would say that if you ran tasks that required a project.clj in a directory without one.
16:24	cs02rm0	raek: fresh. OSX.
16:24	cs02rm0	technomancy: this is 1.2.0
16:24	amalloy	cs02rm0: fresh from where? last i heard the macports version of lein wasn't very good
16:25	cs02rm0	amalloy: ports. i'll give it a go from github. thanks
16:27	technomancy	cs02rm0: that bug may be present there too
16:29	romain_p	Hi everyone, http://clojure.pastebin.com/D4AjBJ80
16:29	romain_p	how do I make this function return an empty vector if the input is an empty seq?
16:29	hsuh	shouldn't you be using map ?
16:30	amalloy	hsuh: i would, but for is fine too
16:30	hsuh	fun, wasnt aware of its existence
16:31	amalloy	&(let [x #{}] (if (seq x) [] (conj x 10)))
16:31	sexpbot	⟹ #{10}
16:31	amalloy	&(let [x #{}] (if-not (seq x) [] (conj x 10)))
16:31	sexpbot	⟹ []
16:31	amalloy	romain_p: this what you want?
16:32	romain_p	amalloy: probably, thanks
16:34	cs02rm0	superb - the github version of lein seems to work flawlessly, many thanks.
16:36	amalloy	cs02rm0: enjoy
16:36	amalloy	technomancy: just pinging you for ^^
16:37	technomancy	great!
16:38	kotarak	,(or (seq nil) [])
16:38	clojurebot	[]
16:39	jccc	how do i best accomplish something like (insert [1 2 3] 0 4) -> [1 4 2 3]
16:40	kotarak	jccc by using finger trees maybe. vectors are not designed for such things.
16:40	jccc	yeah i was hoping not to have to pull in contrib
16:41	kotarak	jccc: if you need to: (defn insert [vektor idx value] (into (subvec vektor 0 idx) [value] (subvec vektor idx))) or something like that
16:43	jccc	kotarak: yeah i was thinking i would have to do something like that. guess ill try to figure out finger trees. thnx
16:49	raek	instead of _insterting_ elements at the right places, would i be possible to make a function that _selects_ the elements in the right order and returns them as a sequence?
16:50	raek	jccc: ^
16:50	kotarak	raek: (concat (subvec v 0 idx) [value] (subvec v idx))
16:50	kotarak	But you loose random access.
16:52	raek	I was thinking about something like the difference between insertion sort and selection sort
16:52	pppaul	woot! got my recursive infix function working
16:57	raek	...since instertion into the middle is costly, but making lazy sequences is cheap
16:57	jccc	raek/kotarak: i think just vectorizing kotaraks solution is fine for now. not sure of its performance
16:57	rata_	hi all
16:57	hsuh	KirinDave: i'm trying to understand how clothesline/webmachine differ from ring in how you have to write your services
16:58	kotarak	jcc: O(N) in vector length
16:58	jccc	kotarak: yeah where like assoc on a vector is O(1)
16:59	kotarak	jccc: yes, but assoc != insertion
16:59	kotarak	honestly it's O(log_32 N)
17:00	jccc	kot: no i know. i wish it could though :). oh true
17:00	LauJensen	kotarak: enough chatting, start blogging :)
17:00	kotarak	LauJensen: argh.
17:00	kotarak	;)
17:01	kotarak	LauJensen: currently mulling about a C program.
17:13	kzar	How would I do something like this?
17:13	kzar	&(map + [[1 2 3] [1 2 3]])
17:13	sexpbot	java.lang.ClassCastException: Cannot cast clojure.lang.PersistentVector to java.lang.Number
17:13	amalloy	&(apply map + [[1 2 3] [1 2 3]])
17:13	sexpbot	⟹ (2 4 6)
17:14	kzar	amalloy: How come map gets evaled there?
17:14	amalloy	&(doc apply)
17:14	sexpbot	⟹ "([f args* argseq]); Applies fn f to the argument list formed by prepending args to argseq."
17:15	amalloy	kzar: so it turns into (map + [1 2 3] [1 2 3])
17:15	kzar	&(apply map + - * / [[1 2 3] [1 2 3]])
17:15	sexpbot	java.lang.IllegalArgumentException: Don't know how to create ISeq from: clojure.core$_
17:15	amalloy	&(apply map (juxt + - * /) [[1 2 3] [1 2 3]])
17:15	sexpbot	⟹ ([2 0 1 1] [4 0 4 1] [6 0 9 1])
17:16	kzar	amalloy: I don't really understand how you can have one extra function but not two?
17:16	amalloy	kzar: you don't have any extra functions
17:16	amalloy	apply is a way to tell clojure that the last argument should not be treated as a single list-argument, but rather expanded, and each element treated as its own argument
17:17	amalloy	so apply mucks around with the last argument, then gets out of the way and calls map
17:17	kzar	amalloy: OK so (apply map + [[1 2] [1 2]]) only expands the last argument so it's something like (apply map + [1 2] [1 2]) which is then something like (map + [1 2] [1 2])?
17:18	amalloy	yep. i said that, in fact, with my "so it turns into..."
17:19	kzar	amalloy: yea thanks, I got that it worked but I didn't understand the process
17:19	kzar	amalloy: so my dodgy example would be (map + - * / [1 2 3] [1 2 3]) which ovbiously isn't going to work
17:19	amalloy	exactly
17:19	kzar	sweet I understand :)
17:20	amalloy	but you can get results basically like that by wrapping + - * / all into a single function, like i did with juxt
17:20	kzar	amalloy: Yea I get you
17:21	kzar	to start with I thought it was some special case or something
17:22	KirinDave	hsuh: I don't understand
17:22	KirinDave	hsuh: Clothesline uses ring.
17:22	amalloy	KirinDave: i think he's asking what clothesline does for him that ring doesn't
17:22	KirinDave	Oh
17:22	KirinDave	hsuh: It provides (or, it will provide) correct http 1.1 behavior by default.
17:23	hsuh	KirinDave: yep, kinda.. because from what it seems to me now, using ring, because you get the http request as a map, you can do anything with it easily..
17:23	KirinDave	hsuh: Well, you won't be altering the request.
17:23	KirinDave	hsuh: WebMachine's benefit is not in golfing low code count on a simple http service.
17:24	KirinDave	hsuh: It's more about providing a very clean way to link your code with a http handler and not have to worry about crazy edge cases in protocol.
17:24	KirinDave	You get correct behavior by default.
17:24	KirinDave	It also lets you do things like handle incoming content types much more easily.
17:25	ohpauleez	KirinDave: re: "...but none of them are specifically oriented towards designing APIs with correct HTTP 1.1 behavior"
17:25	KirinDave	ohpauleez: Did I miss one>
17:25	ohpauleez	In the past I have found Moustache to fill this spot for me, quite well
17:25	KirinDave	I know of coda's jersey for Scala...
17:25	technomancy	it does begs the question why not just fix ring...
17:26	ohpauleez	or enhancing Ring
17:26	hiredman	technomancy: or at least implement as a wrapper for ring
17:26	KirinDave	Right now it produced handler functions that are suitable for use with ring.
17:26	KirinDave	There are some produce-sever functions for convenience there, but they all just use ring.
17:26	hiredman	well then
17:27	KirinDave	I mean, Ring is plenty good at what it does. It's like Rack. Putting tools for making restful APIs into Ring seems outside the scope of that project. The same argument would work against compojure (w.r.t serving generated content).
17:28	KirinDave	hsuh: Does that answer the question?
17:29	hsuh	KirinDave: oh yeah, i'll keep an eye as well :) sooner or later someone will use it for something neat and mind will be blown.. or something
17:30	KirinDave	Maybe.
17:32	KirinDave	ohpauleez: You know, I have been considering making some mustache-like syntax for generating clothesline handlers.
17:32	KirinDave	ohpauleez: Currently Clothesline is awkward.
17:32	ohpauleez	KirinDave: I've been doing the same for some stuff I'm working on in Aleph
17:32	KirinDave	It'd be nice to get a familiar and helpful layer on top for actually delivering things.
17:32	ohpauleez	I find it extremely powerful and elegant
17:33	KirinDave	Could I see a snippet of moustache code for doing some restful stuff?
17:35	KirinDave	ohpauleez: Do you have a git project I could peer at?
17:36	ohpauleez	Currently none of that is sitting in git, but when I hurray up and finish net-ns, it'll be in there
17:36	ohpauleez	let me link you
17:37	ohpauleez	KirinDave: https://github.com/ohpauleez/net-ns A whole lot of nothing but scratches right now.
17:37	KirinDave	I'm curious how you'd do things like handle http headers properly
17:37	KirinDave	e.g., if-modified-since
17:37	KirinDave	(Not that clothesline does that just yet)
17:43	nickik	does anybody know why form-dot-clj isn't working on the appengine?
18:09	pppaul	someone want to help me with a regex problem?
18:11	amalloy	pppaul: don't use regexes? :)
18:11	pppaul	but they are sexy
18:11	amalloy	clojurebot: ask?
18:11	clojurebot	Gabh mo leithscéal?
18:11	amalloy	clojurebot: anyone?
18:11	clojurebot	Please do not ask if anyone uses, knows, is good with, can help you with <some program or library>. Instead, ask your real question and someone will answer if they can help.
18:11	pppaul	^(?:\d[\\+])+\d$ what does the '$' mean?
18:11	amalloy	clojurebot: regex?
18:11	clojurebot	Sometimes people have a problem, and decide to solve it with regular expressions. Now they have two problems.
18:11	amalloy	bah
18:11	amalloy	anyway pppaul it means only match at end of input
18:11	replaca	pppaul: you said that the other night. I think you need to go review "sexy" as a concept :)
18:12	pppaul	regex is kinky
18:12	amalloy	pppaul: are you trying to match either a \ or a +? that's kinda a weird thing to want
18:12	replaca	pppaul: in the sense that you might get all tangled up and not be able to get out, yeah
18:12	pppaul	^_^
18:12	amalloy	replaca: haha awesome
18:13	pppaul	i want to match "1+2+3+4" and not "12+2+3+4"...
18:13	amalloy	you want...only single digits?
18:13	pppaul	it is weird, but it's for the 24 game, so it's got to be that way
18:13	amalloy	ah
18:14	amalloy	[\\+] is not what you want. \+ is fine, and so is [+], but what you have will match 1\2\3+4
18:14	pppaul	my tests pass with \\+. i'll try with \+
18:15	amalloy	yes, it matches everything you want, and also some things you don't want
18:15	amalloy	is my point
18:15	replaca	pppaul: you don't need any \ there. inside the [] you don't need to escape +
18:16	amalloy	&(re-find #"^(?:\d[\\+])+\d$" 1\2\3\4)
18:16	sexpbot	java.lang.IllegalArgumentException: Wrong number of args (5) passed to: core$re-find
18:16	amalloy	&(re-find #"^(?:\d[\\+])+\d$" "1\2\3\4")
18:16	sexpbot	⟹ nil
18:16	pppaul	ok, i'll try not escaping...
18:16	amalloy	pppaul: don't "try" it; it's guaranteed to work :P
18:17	replaca	pppaul: either escape or [] will work, both doesn't really make sense
18:17	pppaul	java.util.regex.PatternSyntaxException: Illegal character range near index 10
18:17	pppaul	^(?:\d[+-\*])+\d$
18:17	pppaul	^
18:17	pppaul	java.lang.Exception: Unable to resolve symbol: string in this context (NO_SOURCE_FILE:0)
18:17	pppaul	java.lang.Exception: Unmatched delimiter: )
18:17	pppaul	java.lang.Exception: Unmatched delimiter: ]
18:17	pppaul	user=> java.lang.Exception: Unable to resolve symbol: match in this context (NO_SOURCE_FILE:683)
18:17	pppaul	java.lang.Exception: Unmatched delimiter: )
18:17	pppaul	java.lang.Exception: Unmatched delimiter: )
18:17	pppaul	didn't work
18:17	amalloy	pppaul: don't escape * either!
18:17	amalloy	inside []
18:17	pppaul	i wasn't
18:17	pppaul	i want to capture it
18:17	replaca	what's this for: \
18:18	replaca	?
18:18	amalloy	\ != /
18:18	pppaul	^(?:\d[+-/*])+\d$ works
18:18	replaca	amalloy: I think you made a new emoticon
18:18	amalloy	hahaha
18:18	joegallo	Ah, \ != /, I'm pretty sure that's an operator in perl.
18:18	replaca	joegallo: most things are
18:18	amalloy	my secret identity is Angry Regex Man
18:18	pppaul	lol
18:19	pppaul	love regex
18:19	pppaul	regex grows hair on my chest
18:19	amalloy	or maybe "I'm so excited by operators!"
18:37	amalloy	anyone have experience with somnium.congomongo? i'm trying to store a tree/nested-map like {:word "doctor" {:links {"phil" 1}}} {:word "phil" {:links {...}}}
18:37	amalloy	but it coerces all the map keys to keywords, and all the map values to strings. i'd be totally fine with only strings or only keywords, but i'm having some serious trouble getting either
18:38	amalloy	(because if i look up "doctor"s links i see that the key is :phil, not "phil", and so the map manipulation i do in clojure doesn't work too great)
18:50	Raynes	amalloy: Hahaha a nap.
18:52	amalloy	ish
18:55	Raynes	scottj: Huh?
18:55	Raynes	scottj: I used it in SLIME, and almost all of my development was in a cake REPL. I don't understand.
18:59	Raynes	Oh shit. You're right. How weird.
19:03	gertalot	hey all. I need to write a function that given a vector [a b c d] returns a vector [a a+b a+b+c a+b+c+d]. How would I do that?
19:04	amalloy	&(reductions + [1 2 3 4])
19:04	sexpbot	⟹ (1 3 6 10)
19:04	gertalot	ah!
19:04	gertalot	i love clojure :)
19:04	gertalot	thx
19:04	amalloy	welcome gertalot
19:05	pppaul	(inc gertalot)
19:05	sexpbot	⟹ 1
19:10	pppaul	commenting in emacs is messed up
19:11	amalloy	...blaming emacs for your problems tends to end with "gosh, i see now why the way it works is better than what i expected"
19:12	pppaul	i dono
19:12	pppaul	i'm putting in comments, and emacs is doing it's own thing
19:12	pppaul	then i try to format the comments and my code suddenly ends up in the middle of sentences i wrote in my comments
19:14	technomancy	pppaul: are you using M-;?
19:15	pppaul	i'm not sure
19:15	pppaul	i'm using ;
19:15	pppaul	let me try the meta key
19:16	pppaul	so, how do i change the setting where it word-wraps?
19:16	amalloy	M-x autofill-mode to toggle it entirely, C-x f to change the width
19:16	pppaul	doesn't C-x f open a new file?
19:17	amalloy	no, that's C-x C-f
19:17	pppaul	my emacs settings are diff
19:17	pppaul	C-x f does file stuff
19:18	pppaul	oh well
19:18	Raynes	technomancy: ping
19:19	Raynes	technomancy: What is SLIME's 'eval thread' and how can one kill it with Clojure code? It appears that I'm doing just that, and I have absolutely no clue how. In any case, when my code runs, C-x C-e and the like don't work any longer.
19:21	amalloy	Raynes: is it possible that starting the java security manager kills some daemon threads?
19:21	amalloy	ie, i bet it's swank, not slime
19:21	technomancy	I'm not sure how to do it from clojure. you can try M-x slime-list-threads
19:21	Raynes	It never did this from clj-sandbox, so it has to be something specific to my code.
19:23	Raynes	https://github.com/Raynes/clojail/blob/master/src/clojail/core.clj#L65 Is the relevant function if anyone has any epiphanies.
19:25	amalloy	Raynes: it happens when all you do is call (sandbox ...)?
19:25	Raynes	I believe so. Not sure if it happens then or when you call what sandbox creates.
19:25	Raynes	I'll test.
19:25	amalloy	if the former then it has to be the security manager, since that's the only code that actually executes
19:26	Raynes	It's the former then.
19:27	defn	Hey Raynes -- what's this new project I heard about on twitter?
19:27	Raynes	amalloy: Removing that code doesn't do anything to fix this.
19:27	amalloy	hm
19:27	Raynes	defn: This one. My awesome new sandbox problem that doesn't work properly in SLIME.
19:28	Raynes	Great introduction, eh?
19:28	Raynes	project*
19:28	Raynes	amalloy: Maybe it happens when I use, let's try that.
19:29	amalloy	hey technomancy: C-c M-p doesn't work if the ns form has ^{metadata}
19:30	Raynes	It actually appear SLIME isn't working properly for me at all, even fresh.
19:30	Raynes	...
19:31	Raynes	amalloy: Can you clone clojail and try this out yourself, pretty please?
19:31	amalloy	Raynes: just did
19:31	amalloy	in the slime repl, i evaluated (sandbox secure-tester)
19:31	amalloy	C-x C-e still works fine
19:31	defn	Raynes: How is this different from yours and Licenser's sandbox?
19:31	Raynes	Yeah, I guess my SLIME is just all of a sudden broken.
19:31	technomancy	amalloy: true. IMO ns metadata is not a good idea, but I would take a patch to fix slime.
19:32	amalloy	Raynes: and now i've called the resulting sandbox: (s '1), and C-x C-e still works
19:32	Raynes	defn: Mine is arguably a bit cleaner, and it's simpler. Most of all, it's documented and not monolithic.
19:32	amalloy	defn: he's reimplementing it but in a less bad way
19:33	defn	:D
19:33	Raynes	defn: I'm trying to take clj-sandbox and just reimplement it in a more sane way.
19:33	defn	Cool.
19:33	Raynes	amalloy: I'm going to nix Emacs and restart it.
19:34	defn	I'd love to see a sandbox that is like Licenser's, but runs in an honest-to-god VM
19:34	defn	so you can be liberal about running everything under the sun
19:34	amalloy	Raynes: sounds good. really everything slime is working for me still: C-c C-m and friends work fine
19:34	defn	side effects and all
19:35	amalloy	defn: if you have a few GB of ram to host some VMs on we're accepting donations
19:35	Raynes	Yep, it's all working great now.
19:35	Raynes	I don't know what happened with scottj.
19:36	Raynes	I found a bug though.
19:38	amalloy	hah, what was it?
19:46	hsuh	can someone give me an insight on how to go from [["a" "b"] [1 2]] to [["a" 1] ["b" 2]] ? trivial in imperative way, but cant see it in clojure
19:46	amalloy	&(apply map vector [["a" "b"] [1 2]])
19:46	sexpbot	⟹ (["a" 1] ["b" 2])
19:46	amalloy	even more trivial in clojure :)
19:48	amalloy	hsuh: the apply converts it to (map vector [...]), which is effectively [(vector "a" 1) (vector "b" 2)]
20:04	hsuh	amalloy: that code is puzzling me a little bit because map accepts more than one collection
20:04	hsuh	&(map vector ["a" "b"] [1 2])
20:04	sexpbot	⟹ (["a" 1] ["b" 2])
20:04	hsuh	but if that works, why isnt this allowed?
20:04	hsuh	&(map (fn [x] (* x 2)) [1 2 3] [4 5 6])
20:04	sexpbot	java.lang.IllegalArgumentException: Wrong number of args (2) passed to: sandbox4138$eval6573$fn
20:04	amalloy	&(map (fn [x y] (* x y)) [1 2 3] [4 5 6])
20:04	sexpbot	⟹ (4 10 18)
20:05	hsuh	oh, its like clojure doesnt need map2
20:05	amalloy	&(doc map)
20:05	sexpbot	⟹ "([f coll] [f c1 c2] [f c1 c2 c3] [f c1 c2 c3 & colls]); Returns a lazy sequence consisting of the result of applying f to the set of first items of each coll, followed by applying f to the set of second items in each coll, until any one of the colls is exhausted. Any remaining items in other colls ... http://gist.github.com/670171
20:06	amalloy	i don't know what map2 is, but no, we don't need it :P
20:06	amalloy	&(apply map + (replicate 100 (range 10)))
20:06	sexpbot	⟹ (0 100 200 300 400 500 600 700 800 900)
20:06	amalloy	look, a map100 :)
20:07	hsuh	i think map2 memory came from ocaml, which probably explains its need to exist
20:07	hsuh	amalloy: thanks a lot
20:21	livingston	how do I type hint for a non-primitive array? e.g. I want to hit this method: [ 3] add : void (Iterable,Resource[])
20:22	amalloy	livingston: ISTR it's ^LResource or something close to that
20:23	amalloy	java has weird classnames for arrays
20:23	livingston	i'll give that a shot, thanks
20:23	amalloy	livingston: &\|(make-array Object 0)\|&
20:23	sexpbot	⟹ #<Object[] [Ljava.lang.Object;@b62f6>
20:24	amalloy	the [Ljava.lang.Object is the array's official classname
20:24	amalloy	so you probably have to do like ^{:tag "[Ljava.lang.Object"}
20:24	amalloy	anyway i'm off. good luck
20:25	livingston	amalloy: that will give me an empty array right? that's cool, that's what I happen to want (this time)
20:25	livingston	thanks
20:25	amalloy	yes it will
20:25	livingston	looks like that worked
20:33	pppaul	(ns rosettacode.24game
20:33	pppaul	(:require '[clojure.string :as str]))
20:33	pppaul	rosettacode.24game=> (str/capitalize "12312312")
20:33	pppaul	java.lang.Exception: No such namespace: str (NO_SOURCE_FILE:83)
20:33	pppaul	:(
20:34	pppaul	what am i doing wrong?
20:34	pppaul	^_^
20:34	technomancy	pppaul: no need to quote require clauses inside the ns form
20:35	pppaul	ok, i copied it off of the clojure API
20:35	technomancy	I find that ... unlikely?
20:35	technomancy	but if there's a typo in the official docs it should get fixed
20:35	pppaul	It is poor form to (:use clojure.string). Instead, use require
20:35	pppaul	with :as to specify a prefix, e.g.
20:35	pppaul	(ns your.namespace.here
20:35	pppaul	(:require '[clojure.string :as str]))
20:36	pppaul	works now, thanks :)
20:36	pppaul	should get the official docs edited
20:36	technomancy	eep; do you have a URL for that?
20:37	pppaul	http://clojure.github.com/clojure/clojure.string-api.html
20:37	technomancy	ok, good catch. sorry I doubted you. =\
20:37	pppaul	^_^
20:38	replaca	That whole comment doesn't even really belong there
20:38	replaca	I think it was Stuart reacting to the old (googlecode) autodoc for contrib which had use in it
20:41	pppaul	i use :use for clojure.test... is that bad?
20:41	pppaul	does it make sense at all?
20:42	technomancy	I think that's acceptable
20:42	pppaul	why is :use in the language?
20:42	technomancy	nothing wrong with :use if you also have :only to limit the scope of what you bring in
20:42	technomancy	it's a design bug that leaving out :only brings in everything
20:42	pppaul	i need a tutorial on ns stuff :(
20:42	Derander	can you :require :as?
20:43	technomancy	it's slated for fixing; everyone agrees it is needlessly complex
20:43	technomancy	pppaul: the best thing to do is read existing code
20:43	pppaul	i find it really tricky (been having lots of problems getting my stuff to work using (ns :require) form
20:43	Derander	yes, the whole use/require/import stuff is one of my biggest sticking points as a newbie
20:43	Derander	I'm still cargo-culting it
20:44	pppaul	i use clojuredocs, but the examples for ns stuff seem a bit weak
20:44	Derander	yes
20:44	Derander	same problem
20:47	replaca	Derander: yes, you can :require :as
20:47	pppaul	?
20:48	pppaul	ns rosettacode.24game
20:48	pppaul	(:require [clojure.string :as str])
20:48	pppaul	(:use (clojure.test))) rosettacode.24game=> java.lang.Exception: Unable to resolve symbol: deftest in this context (NO_SOURCE_FILE:158)
20:49	pppaul	so, i use :use, but deftest still isn't defined
20:50	pppaul	but, i created the ns before, and then tried to fix it by adding :use
20:50	pppaul	are namespaces allowed to change? or are they immutable?
20:51	tomoj	(:use clojure.test)
20:51	tomoj	not (:use (clojure.test))
20:51	amalloy	pppaul: tomoj is right
20:52	amalloy	symbols, vectors, and lists all behave differently in a :use entry
20:52	tomoj	vectors and list are the same, aren't they?
20:52	pppaul	from (doc ns) (ns foo.bar
20:52	pppaul	(:refer-clojure :exclude [ancestors printf])
20:52	pppaul	(:require (clojure.contrib sql sql.tests))
20:52	pppaul	(:use (my.lib this that))
20:52	pppaul	(:import (java.util Date Timer Random)
20:52	pppaul	(java.sql Connection Statement)))
20:52	tomoj	I mean, they're seqs
20:52	tomoj	(my.lib this that) means both my.lib.this and my.lib.that
20:52	amalloy	&(map class () [])
20:52	sexpbot	⟹ ()
20:53	technomancy	ugh; that's awful. one more thing to be fixed in CLJ-272 I guess.
20:53	amalloy	,(map class () [])
20:53	clojurebot	()
20:53	amalloy	oh hah
20:53	amalloy	&(map class [() []])
20:53	sexpbot	⟹ (clojure.lang.PersistentList$EmptyList clojure.lang.PersistentVector)
20:53	pppaul	wooo
20:53	amalloy	tomoj: maybe you're right. namespaces still confuse me :P
20:54	amalloy	&(macroexpand (ns foo (:require [clojure.contrib math] (clojure.contrib math))))
20:54	sexpbot	java.io.FileNotFoundException: Could not locate clojure/contrib/math__init.class or clojure/contrib/math.clj on classpath:
20:54	amalloy	&(macroexpand '(ns foo (:require [clojure.contrib math] (clojure.contrib math))))
20:54	pppaul	works now :D
20:54	sexpbot	⟹ (do (clojure.core/in-ns (quote foo)) ((fn* loading__4410__auto__ ([] (dot clojure.lang.Var (clojure.core/pushThreadBindings {clojure.lang.Compiler/LOADER (dot (dot loading__4410__auto__ getClass) getClassLoader)})) (try (clojure.core/refer (quote clojure.core)) (clojure.core/require (quote [clojure.... http://gist.github.com/670215
20:57	amalloy	tomoj: yeah, okay, [] and () are the same in ns. another of my delusions spoiled
20:57	pppaul	i would assume that [] gets converted to a seq
20:58	pppaul	it would have to if any seq funcs were applied to it
20:58	amalloy	pppaul: true as far as it goes. but rich could check whether it's a vector or a list before doing anything to it, and then do different things depending. apparently he doesn't
20:59	pppaul	yeah... it looks like that is sorta being done, since (clojure.contrib is not an op
20:59	tomoj	a list is a list :)
20:59	pppaul	least i don't think it is
21:00	amalloy	tomoj: ?
21:00	pppaul	it's an unevaluated list, which is different from an evaluated one
21:00	tomoj	there doesn't need to be any special handling to avoid treating clojure.contrib as an op
21:00	tomoj	it's just a list
21:01	pppaul	wouldn't the special handling be in the macro?
21:01	amalloy	pppaul: yes, but not necessarily in the expansion
21:01	tomoj	no need for special handling in the macro either
21:01	tomoj	the macro just receives a list and can process it like a seq
21:01	pppaul	the macro doesn't say to not evaluate the list?
21:02	amalloy	tomoj: i know. (i think) we're discussing the hypothetical situation where rich wants to treat lists and vecs differently
21:02	tomoj	yeah, hypothetical, not sorta being done
21:02	pppaul	anyway, it doesn't matter much... documentation matters more now :)
21:02	pppaul	newbs in trouble
21:03	amalloy	(defmacro is-list? [var] (if (list? var) () [])
21:03	amalloy	(is-list? [1 2 3]) ; expands to [] with no trace of the list/vec check
22:33	tomoj	&(keys [1 2 3])
22:33	sexpbot	java.lang.ClassCastException
22:33	tomoj	try that in a plain repl
22:34	tomoj	oh, nevermind
22:34	tomoj	I was trying to figure out why it prints a single '(', but that's seqiness
22:34	tomoj	for some reason, though, (keys [1 2 3]) causes slime to error
22:34	tomoj	something strange about the exception I guess?
22:45	amalloy	tomoj: keys is lazy
22:45	amalloy	&(let [x (keys [1 2 3])] 1)
22:45	sexpbot	⟹ 1
22:46	Raynes	,(.replace (.replace "x" "x" "y") "y" "x" "x")
22:46	clojurebot	java.lang.IllegalArgumentException: No matching method found: replace for class java.lang.String
22:46	amalloy	so (keys [1 2 3]) completes without error, and then when the repl tries to show it to you the exception finally comes up
22:46	tomoj	right
22:46	amalloy	Raynes: ?
22:47	Raynes	Too many arguments.
22:47	Raynes	Weird error message.