#clojure – 2011-02-02

2011-02-02

02:00	amalloy	ieure: i ran into an issue with php-repl when trying to use it for anything complex
02:01	amalloy	eg, foreach (range(0,10) as $x) {echo $x;}
02:01	amalloy	doesn't work because you're doing eval('return foreach (...) {...};')
02:02	amalloy	i can fix this by writing function f() {...} followed by f(), but yuck
02:05	chouser	the javascript generated by clojurescript has a somewhat startling number of anonymous inner functions.
02:05	chouser	and now, to bed.
02:07	amalloy	chouser: startling? i'd expect one for every lexical scope at least
02:08	ieure	amalloy, I haven’t touched PHP or that code in a couple years. If it’s a big issue for you, I’ll lean into it, though.
02:08	amalloy	ieure: don't bother yourself about it yet. i'll fork it on github and take a look; if it's unexpectedly complicated i'll let you know
02:24	fliebel	morning
02:26	hoeck	morning
02:31	amalloy	morning fliebel
02:33	fliebel	amalloy: did tonyl finish his code yet?
02:33	ieure	amalloy, sounds good.
02:33	amalloy	fliebel: if so he didn't tell me
02:33	fliebel	okay
02:36	fliebel	did you show him the php code?
02:38	amalloy	fliebel: no, didn't think of it
03:04	amalloy	but now it's bedtime for me, so you'll have to link him if he shows up, fliebel :P
03:06	fliebel	amalloy_: good night :)
04:19	Licenser	aloa
04:39	mduerksen	,(let [v2scale (fn [s [a b :as v]] (assoc v 0 (* a s) 1 (* b s)))] (time (dotimes [n 1000000] (v2scale 0.2 [2 2]))))
04:39	clojurebot	"Elapsed time: 1895.93 msecs"
04:40	mduerksen	this is the fastest fn i could up with. any suggestions?
04:42	mduerksen	unchecked-math isn't a great improvement in this case, it's share in the overall execution time is quite small
05:46	MrHus	, (= 1 -1)
05:46	clojurebot	false
06:58	clgv	I am curious if anyone tried to build a Clojure DSL for Latex to get rid of some in inadequateness of it...
07:40	bartj	I have a bunch of extremely ugly code to write
07:41	bartj	but, I think Clojure will have an "elegant" way of writing it
07:41	bartj	here is the task:
07:41	bartj	values = v1, v2, v3 and there are fields f1, f2, f3
07:42	bartj	I need to concatenate the fields f1-v1, f1-v2, f1-v3 if f1 exists
07:42	bartj	ditto with f2, f3
07:42	bartj	my normal way of doing it is:
07:42	bartj	(for [v values] (if f1 (str f1 "-" v)) (if f2 (str f2 "-" v)) ....
07:43	bartj	I was wondering if there is a better way ?
07:46	clgv	(map #(when %1 (str %1 "-" %2)) fields values)
07:46	clgv	does that work for you?
07:48	clgv	&(map #(when %1 (str %1 "-" %2)) ["f1" "f2" "f3"] [1.0 2.0 3.0])
07:48	sexpbot	⟹ ("f1-1.0" "f2-2.0" "f3-3.0")
07:48	clgv	as test
07:49	ogonzalez	bartj, using for: (for [f f-values v values :when f] (str f "-" v))
07:49	ogonzalez	I think it's right
07:51	ogonzalez	&(for [f ["f1" nil "f2"] v ["v1" "v2"] :when f] (str f "-" v))
07:51	sexpbot	⟹ ("f1-v1" "f1-v2" "f2-v1" "f2-v2")
08:39	bartj	ogonzalez, yep, will try that
08:39	bartj	ogonzalez, thanks!
09:28	pppaul	how do i add a classpath to clojure while it's running?
09:28	dnolen	pppaul: no reliable way to do that.
09:29	pppaul	oh
09:29	pdk	http://stackoverflow.com/questions/402330/is-it-possible-to-add-to-classpath-dynamically-in-java
09:29	pdk	java isn't really built for this but look into URLClassLoader it seems
09:30	pdk	remember java comes from the mindset of the old c style write compile run debug loop
09:30	pppaul	hmmm
09:30	AWizzArd	Instead of directly the URLClassLoader one could also have a look at the DynamicClassLoader that ships with Clojure.
09:31	AWizzArd	This IS an URLClassLoader too.
09:31	AWizzArd	https://github.com/clojure/clojure/blob/master/src/jvm/clojure/lang/DynamicClassLoader.java
09:41	edw	Any cake users here?
09:42	edw	I'm trying to add /usr/local/share/java to my java.library path and am putitng an entry into $PROJECT/.cake/config with no success.
09:44	dnolen	edw: what does you entry look like?
09:45	edw	Line 1: "project.library.path = /usr/local/share/java"
09:45	edw	Line 2 is the same, but cake.library.path.
09:45	edw	Do I also need to add a "native-path" entry to my project.clj?
09:46	dnolen	edw: if you're on 0.6 I think the syntax changed, jvm.opts = -Xms1024M -Xmx2048M -Dfoo=bar
09:47	nurv101	Question here: is it possible to create a lisp closure in clojure?
09:47	edw	Ah. I'll check that out. Are they additive i.e. can I put that in $HOME/.cake/config and it'll be concatenated with the projects .cake/config file?
09:48	dnolen	nurv101: yes
09:48	raek	nurv101: what exactly do you mean by "lisp closure"?
09:48	nurv101	dnolen: how?
09:48	nurv101	raek: (let ((x 0)) (defun foo() (incf x)))
09:48	dnolen	edw: no they are separate.
09:49	edw	(let [x 12] (fn [] x))
09:49	raek	(defn create-counter [a (atom 0)] (fn [] (swap! a inc)))
09:49	edw	Ahh. Okay.
09:49	raek	nurv101: ^
09:49	nurv101	raek: that is not a clojure
09:50	raek	clojure won't allow you to change the local introduced by let
09:50	nurv101	raek: closure i mean
09:50	raek	if that was your question
09:50	nurv101	raek: i know
09:50	raek	every time you call create-counter, you will get a function with its own atom
09:50	nurv101	the problem is that even local-vars can't be change
09:51	nurv101	that way
09:51	raek	in clojure, you have to have some form of mutable cell that does the change part
09:51	raek	i.e. an atom, agent, ref, or var
09:52	raek	all these have their own approach to what mutation means in a concurrent context
09:53	edw	But the answer to your question is that yes, clojure has lexical scoping aka closures.
09:54	raek	yes. functions close over their free variables.
09:54	edw	Yes.
09:55	nurv101	lexical scope isn't a clojure
09:55	nurv101	closure
09:57	mefesto	Good morning everyone.
09:58	mefesto	If i wanted to send some clojure data over the network using pr and read -- is there a preferred content-type for this?
09:58	mefesto	application/clojure-source or something?
09:59	edw	nurv101: A closure is an artifact of lexical scoping combined with the ability to declare a procedure in the scope of another.
10:00	raek	not that I'm aware of. I used "text/x-clojure-source" to mimic "text/x-java-source"
10:01	mefesto	raek: works for me, thanks :)
10:01	raek	the "x-" means that it hasn't been registered
10:12	akopa	A bit rough, but here's an clr interop example adapted from RDNZL for Common Lisp: https://gist.github.com/807795
10:23	edw	Has anyone used zmq and cake on osx? My forehead is bloody from beating it against my monitor.
10:29	__name__	hi
10:31	raek	if __name__ == "__main__": print "__name__: hi"
10:32	__name__	Could any of you be ask kind as to express criticism about https://gist.github.com/806541, I'd like to know where I can improve.
10:53	semperos	I call a function A which returns a vector; I need to call a function B which takes the _contents_ of the vector, like splicing; what's the proper way to handle this?
10:54	mefesto	semperos: (apply B [1 2 3])
10:54	__name__	semperos: do you mean (apply + [1 2 3])
10:54	mefesto	semperos: is the same as calling: (B 1 2 3)
10:55	semperos	makes sense
10:55	semperos	thanks, simplest things escape me sometimes...
10:56	pdk	you can also put extra args in front of the vector
10:56	pdk	so (apply + 1 2 3 [4 5 6]) is the same as (+ 1 2 3 4 5 6)
10:57	__name__	i wish apply and partial had shorter names
11:11	edw	In cake, "ext-dependencies", what are these things? I'm currently going through the source code, but if cake's project file format is documented anywhere, that would be great.
12:06	edw	Is there a safe way to eval strings? Basically a JSON-y reader procedure?
12:50	phenom_	im having a bit of trouble with cake ... I created a project.clj file correctly and added a single file in the test folder (chapter08/date-operations-spec.clj) but when I run "cake test" it complains (hapter08/date_operations_spec.clj on classpath, note the underscore's)
13:30	Scriptor	hey everyone, I have a question about the sequence implementation as described in http://blog.higher-order.net/2009/02/01/understanding-clojures-persistentvector-implementation/
13:32	Scriptor	say you have <32 elements, and you cons a new element on the vector
13:33	Scriptor	following that code, it seems like it finds the right and bottom-most array, and copies it
13:34	Scriptor	what exactly is the new root being created with such a cons?
13:46	amalloy	Scriptor: as i understand it, if you have fewer than 32 elements and cons another, it copies the whole thing
13:46	amalloy	well. conj, not cons
13:47	Scriptor	amalloy: by the whole thing, you mean just that set of <32 elements, right?
13:47	amalloy	Scriptor: yes, since that is the whole thing
13:47	Scriptor	isn't that still a little inefficient?
13:48	amalloy	edw: maybe check out clojail
13:48	Scriptor	ok, the whole thing could be the whole tree too, right? :)
13:48	amalloy	Scriptor: i don't understand. if you have less than 32 elements, the tree is a single node
13:49	Scriptor	amalloy: oh, I meant 32 elements in the rightmost and bottom-most set of elements
13:49	amalloy	edw: it's the safe-evaluation library Raynes and i have put together, and it's what sexpbot uses. see, for example, ##(inc (eval (read-string "2")))
13:49	sexpbot	java.lang.SecurityException: You tripped the alarm! eval is bad!
13:50	Scriptor	for example, looking at the diagram, if we imagined that red series of 1..32 was instead something like 1..14
13:50	Scriptor	er, 0..14
13:50	Scriptor	and we added another element to it
13:50	amalloy	then yes, those 32 elements are copied
13:51	amalloy	and that's inefficient but unavoidable, and not terribly inefficient
13:51	Scriptor	ah, I understand now, thanks!
13:51	amalloy	one malloc and 32 pointer copies?
13:52	Scriptor	right, it uses System.arraycopy
13:52	Scriptor	which I guess would be pretty efficient
13:52	amalloy	probably
13:52	amalloy	and iirc it has to make a copy of every node heading up to the root, but that will be a pretty small number
13:59	pdk	okasaki's book is a good primer on functional data structures scriptor
14:01	Scriptor	pdk: is his thesis paper a good substitute
14:01	Scriptor	?
14:01	Scriptor	That's what I've been trying to use
14:01	pdk	the book expands it some and adds haskell translations of the ML code samples
14:02	Scriptor	ooh, haskell translation would be nice
14:02	Scriptor	argh, still pricey though :/
14:03	ossareh	morning all
14:03	Scriptor	morning ossareh
14:04	pdk	oh wow it's on kindle now
14:04	Scriptor	yep, I have one too
14:05	Scriptor	I just wish I knew for certain I'd read it
14:05	Scriptor	would you say the book is more beginner-friendly?
14:05	ossareh	which book?
14:05	pdk	i haven't completed it but i doubt it'll be much less dense
14:05	pdk	aside from adding introductory bits
14:05	pdk	with stuff like the tree copying example
14:06	Scriptor	ossareh: purely functional data structures
14:07	ossareh	Scriptor: ah - nice - man that was a hard read
14:13	Scriptor	ossareh: I know! I'm honestly still trying to find a detailed enough explanation of amortization
14:13	__name__	I know I already posted it, but I think the message was not noticed (and I updated the code): Please tell me if you can see anything unidiomatic in https://gist.github.com/806541.
14:15	opqdonut	looks ok
14:22	shortlord	How can I import a simple .java file that is placed in the same directory as the clojure files?
14:23	shortlord	I have written a file called Foobar.java
14:23	shortlord	and used :import [Foobar]
14:23	__name__	opqdonut: Thanks
14:23	shortlord	but (Foobar/foo "test") still doesn't work
14:23	mefesto	shortlord: did you compile it and it's class is on the classpath?
14:25	shortlord	mefesto: I have compiled it using javac
14:25	mefesto	shortlord: are you using leiningen or cake for building this project?
14:25	shortlord	I am using cake
14:26	amalloy	__name__: i don't think we use lazy-cat anymore
14:26	__name__	amalloy: what else?
14:27	mefesto	shortlord: i haven't tried a mixed java + clojure source project using leiningen (or cake) so I'm not sure if this is necessary but did you specify a :java-source-path in your project.clj?
14:27	shortlord	mefesto: no, I didn't. Do you have any link about the :java syntax?
14:27	amalloy	i'm still working through what you're actually doing, but it seems like you're using ::p as a sentinel "non-data" value. i usually use (complement #{::p}) rather than #(not= ::p %) for that, but it's not 100% clear it's applicable to your case
14:28	amalloy	oh, you mean what instead of lazy-cat?
14:28	mefesto	shortlord: im going by this leiningen example: https://github.com/technomancy/leiningen/blob/stable/sample.project.clj
14:30	shortlord	mefesto: thx, I will look at that
14:30	__name__	amalloy: yes
14:30	amalloy	concat should be lazy enough for your purposes
14:30	__name__	so concat is lazy by default?
14:31	amalloy	$source concat
14:31	sexpbot	concat is http://is.gd/e8fROL
14:31	amalloy	the source is messy because of optimizations, but you can see it's wrapped with lazy-seq
14:32	__name__	where's the difference to lazy-cat then?
14:32	amalloy	lazy-cat is a relic of 1.1, i think
14:32	amalloy	ie, there's no reason to use it
14:32	__name__	amalloy: fair enough, thank you
14:33	brehaut	morning
14:33	__name__	good morning, brehaut
14:33	amalloy	hola brehaut
14:34	__name__	amalloy: the source to concat confuses me :)
14:34	brehaut	$source concat
14:34	sexpbot	concat is http://is.gd/e8fROL
14:36	brehaut	__name__: are you focusing just on the arity-2 version?
14:37	__name__	yes
14:38	amalloy	__name__: ignore the chunked half, too
14:38	brehaut	the guts of the algorithm is the else part of the inner most if; everything else can be handwaved as 'implementation details'
14:38	__name__	(cons (first s) (concat (rest s) y))
14:38	__name__	that?
14:38	brehaut	yup
14:39	__name__	$source cons
14:39	sexpbot	cons is http://is.gd/5brDTa
14:39	brehaut	ie a recursive call. it doesnt blow stack because concat is wrapped in a lazy-seq
14:39	brehaut	cons is the singlely linked list / seq equivalent of conj
14:40	__name__	fair enough, fp still sort of confuses me by times
14:40	amalloy	(defn my-cat [x y] (lazy-seq (let [s (seq x)] (if-not s y (cons (first s) (concat (rest s) y)))))) would be a simple version, and i think correct
14:41	amalloy	brehaut: did you do something weird when talking about wrapping? i see a \u0008 character right before "wrapped"
14:42	brehaut	i dont ithnk i did?
14:43	brehaut	amalloy: i thik your version is correct yes
14:43	__name__	amalloy: that makes sense
14:43	amalloy	__name__: oh good
14:43	__name__	thank you very much, amalloy
15:00	opqdonut	this is weird
15:00	opqdonut	(first (map prn [1 2 3])) prints all of the numbers
15:01	opqdonut	even though (map prn [1 2 3]) has class clojure.lang.LazySeq and chunked-seq? returns false for it
15:01	opqdonut	is there something I'm missing?
15:01	amalloy	opqdonut: chunked-seq? isn't very reliable in my experience
15:02	opqdonut	is there some way I can force a seq to be one-by-one
15:02	amalloy	it's definitely a chunking issue
15:02	opqdonut	yes, I figured as much
15:02	opqdonut	(this is 1.2.0)
15:02	amalloy	ummmmm, i'm gonna make chouser answer that question. he put some kind of unchunker in JoC
15:02	tomoj	&(class (next (map prn [1 2 3])))
15:02	sexpbot	⟹ 1 2 3 clojure.lang.ChunkedCons
15:03	amalloy	&(class (seq (map prn [1 2 3])))
15:03	sexpbot	⟹ 1 2 3 clojure.lang.ChunkedCons
15:03	opqdonut	indeed
15:03	amalloy	before the seq call, it can't know that the underlying lazy seq will be chunked
15:04	tomoj	is it the side-effects you're worried about?
15:04	opqdonut	I'm worried about consistency
15:04	opqdonut	I'm teaching a course on clojure...
15:04	tomoj	I recommend teaching them not to pass side-effecting functions to map :)
15:05	opqdonut	that I am doing, I'd just like to demonstrate laziness
15:06	amalloy	&(first (iterate #(do (prn %) (inc %)) 0))
15:06	sexpbot	⟹ 0
15:06	amalloy	&(second (iterate #(do (prn %) (inc %)) 0))
15:06	sexpbot	⟹ 0 1
15:07	opqdonut	&(first (map prn (range)))
15:07	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 nil
15:07	opqdonut	&(first (map prn (repeat 50 "!")))
15:07	sexpbot	⟹ "!" nil
15:08	tomoj	could be dangerous
15:10	tomoj	imagine defining the ackermann function and using (range), taking a few
15:10	shortlord	does anyone know how to compile .java files using cake?
15:10	opqdonut	tomoj: indeed
15:12	tomoj	I hardly remember ever having actual problems from it
15:26	LauJensen	shortlord: AFAIK the cake javac plugin is loaded by default, so it should just pick up the java files and compile them, try asking in #cake.clj if you have any problems
15:26	amalloy	fwiw opqdonut i've never been impacted by chunking
15:26	opqdonut	well me neither
15:26	amalloy	it's something that sounds appalling but winds up not really mattering
15:26	opqdonut	that's why I was kinda surprised
15:28	opqdonut	for educational reasons I'd like a dechunkifier
15:28	opqdonut	(defn my-seq [s] (map first (iterate next s))) seems to work at least
15:29	amalloy	opqdonut: really? i don't think you can unchunk a seq that's already chunked; you have to prevent it happening in the first place
15:30	opqdonut	that's what I meant, of course
15:30	shortlord	LauJensen: thx for the suggestion, javac does not seem to be loaded with cake by default, but I managed to get it working with lein javac
15:30	opqdonut	so (first (map prn (my-seq [1 2 3]))) prints only 1
15:31	shortlord	is there something like :reload-all for imports?
15:32	amalloy	shortlord: i don't think so. java doesn't like classes to change at runtime
15:33	shortlord	amalloy: I'd like to reimport them in clojure without restarting the REPL though
15:35	shortlord	amalloy: is there really no way of doing it? it should not be too different from how clojure does it itself, right?
15:35	shortlord	it works as soons as I restart them REPL and then import the file again, but always closing and restarting the REPL kills the dynamic workflow
15:36	brehaut	shortlord: but for a dynamic workflow you infrequently need to reload more than one ns
15:36	opqdonut	you can google "java class reloading" and read all the problems :)
15:37	amalloy	i don't know. clojure does it, i think, by re-reading the .clj files and generating new classes with different (?) names
15:40	ossareh	hm, if that were the case permgen would run out - and i've not seen that to be the case in any of my super long running repl sessions (often weeks at a time between restarts)
15:41	raek	hrm. when was support for docstrings in 'def' added?
15:42	opqdonut	&(doc def)
15:42	sexpbot	java.lang.SecurityException: You tripped the alarm! def is bad!
15:42	opqdonut	:D
15:42	opqdonut	okay then
15:43	arohner	ossareh: permgen is GC'd when the classloader that created the class is GC'd
15:43	raek	http://clojure.org/special_forms says (def symbol doc-string? init?)
15:43	arohner	ossareh: err. classes can be GC'd after their classloader is GC'd
15:44	opqdonut	in practice getting classes to GC can be bothersome
15:44	arohner	clojure gets around the permgen issue by creating a new classloader each time it creates classes
15:44	brehaut	raek: presumably thats in 1.3 cause 1.2 erros if i actually try it
15:44	ossareh	arohner: ah, I didn't know that - cool. Cheers.
15:44	opqdonut	when I last meddled in classloaders I think I ended up creating one per loaded class
15:45	opqdonut	(the program was generating a ton of one-off classes)
15:45	ossareh	opqdonut: right - I had a similar issue with a simple web app hosted in jboss, a few deploys and I guess something was keeping class headers around.
15:58	raek	what is the preferred way of testing private variables with clojure.test? @#'foo? something else?
16:05	pjstadig	amalloy: the compiler generates new class names using a global serial number in clojure.lang.RT
16:07	pjstadig	raek: fogus wrote a macro http://blog.fogus.me/2010/09/03/monkeying-with-clojures-deftest/
16:07	pjstadig	or you could just @#'foo
16:10	sritchie	hey all -- I have two clojure functions that are very similar, and wanted to ask advice on how to abstract out the common pattern. here are the functions -- https://gist.github.com/3148f33f2702b029fe78
16:10	sritchie	I could have all-files call files-with-names and simply not return the ?filename, but the function is exactly the same length
16:11	brehaut	where is <- defined?
16:11	edw	It's cascalog.
16:11	brehaut	oh, cascalog?
16:11	sritchie	brehaut: it's a cascalog macro
16:11	sritchie	brehaut: uip
16:11	sritchie	yup
16:12	sritchie	I suppose I could just give two vectors, input args and output args,to a function
16:12	sritchie	and have it return the common function
16:13	kumarshantanu	hi, can anybody let me know how can i easily extend IFn of invoke (arity 2) to a type?
16:14	kumarshantanu	there are like 21 arity signatures for invoke in IFn, and i don't need all of them
16:22	tomoj	do you own the type?
16:22	tomoj	deftype/defrecord? or what?
16:23	raek	kumarshantanu: from java or using proxy: extend AFn instead of implementing IFn
16:24	sritchie	brehaut: they're similar, but I'm not sure what the pattern is --
16:24	sritchie	in one case, I'm dropping one of the fields
16:25	brehaut	im not sure im much help, i dont know cascalog at all
16:25	brehaut	do you only have the two functions?
16:26	amalloy	sritchie: i suspect that <- is a macro in cascalog, so you can't generalize it without adding a macro of your own - a function can't usefully call it since the options will need to be available at compile time
16:27	sritchie	amalloy: in this case, is it worth it to keep the duplication?
16:27	sritchie	amalloy: the other option is to have the tap return filename for everything, and just ignore it further down the road
16:27	sritchie	amalloy: as that's essentially what "all-files" does anyway
16:28	amalloy	eg (defmacro casca-fn [name doc v1 v2] `(defn ~name ~doc [arg#] (let [source# (hfs-whole-file arg#)] (<- ~(vec v1) (source# ~@v2)))))
16:29	amalloy	i don't know anything about cascalog, but that's a macro you should be able to call as (casca-fn all-files "doc" [?file] [?filename ?file]) to reproduce your second defn
16:31	sritchie	amalloy: interesting, let me try that out
16:31	sritchie	amalloy: this'll be my first macro
16:31	amalloy	sritchie: whew. i just tried it out, and it does
16:31	sritchie	haha, someday I'll be able to churn out working macros in IRC
16:35	amalloy	incidentally, if you wanted to you could eliminate the "source" binding by moving it into the <- in-place
16:42	sritchie	amalloy_afk: done, and playing with macroexpand
16:43	mefesto	,(= (into-array Byte/TYPE [1 2 3]) (into-array Byte/TYPE [1 2 3]))
16:43	clojurebot	java.lang.IllegalArgumentException: argument type mismatch
16:45	mefesto	,(java.util.Arrays/equals (into-array Byte/TYPE (map byte [1 2 3])) (into-array Byte/TYPE (map byte [1 2 3])))
16:45	clojurebot	true
16:47	mefesto	hmm is there another way to test array equality?
16:47	companion_cube	,(= [1 2 3] [1 2 3])
16:48	clojurebot	true
16:48	mefesto	companion_cube: java arrays tho
16:48	mefesto	more specifically, java byte arrays
16:48	companion_cube	don't know
16:50	amalloy	argument type mismatch? that's a weird exception for that
16:50	amalloy	&(= (byte-array [1 2 3]) (byte-array [1 2 3]))
16:50	sexpbot	java.lang.ClassCastException: java.lang.Integer cannot be cast to java.lang.Byte
16:50	mefesto	amalloy: yup, i forgot to (map byte [1 2 3]) it
16:51	mefesto	&(= (byte-array (map byte [1 2 3])) (byte-array (map byte [1 2 3])))
16:51	sexpbot	⟹ false
16:51	mefesto	oh well i guess i can live with j.u.Arrays/equals
16:53	amalloy	mefesto: you could seq them
16:53	amalloy	&(apply = (map seq (repeatedly 2 #(byte-array (map byte [1 2 3])))))
16:53	sexpbot	⟹ true
16:54	mefesto	amalloy: sweet, i'll take that over j.u.Arrays. thanks :)
16:54	amalloy	&(apply = (map identity (repeatedly 2 #(byte-array (map byte [1 2 3])))))
16:54	sexpbot	⟹ false
16:54	amalloy	yeah, just checking. seq is the magic
16:55	opqdonut	gah, lein fails
16:55	opqdonut	can't find org.clojure:clojure:jar:1.2.0
16:55	raek	opqdonut: newest version?
16:55	raek	done lein self-install?
16:56	opqdonut	i downloaded the script and said "lein"
16:56	opqdonut	let's try self-install
16:56	opqdonut	ok, that seemed to do something
16:58	raek	opqdonut: it seems that in the latest versions, self-install should be automatic when you use lein the first time
16:58	opqdonut	ok
16:58	raek	you used the stable version, rigth? https://github.com/technomancy/leiningen/raw/stable/bin/lein
16:58	opqdonut	yeah
17:00	opqdonut	hmm, "lein repl" still complains about org.clojure:clojure:jar:1.2.0 and org.apache.maven:super-pom:jar:2.0 missing
17:00	opqdonut	I kinda hate all these language-specific frameworks, I'm fighting with leiningen and cabal at the moment...
17:02	amalloy	opqdonut: lein repl wants you to be in a project
17:03	amalloy	cake repl will implicitly use the global project
17:03	technomancy	actually lein repl will run in leiningen's classloader if you're outside a project.
17:04	amalloy	technomancy: sorry, i try to avoid singing only-cake's praises, but i forgot this time :)
17:04	technomancy	missing clojure.jar sounds like a firewall issue
17:04	opqdonut	I'm in a project as far as I know
17:05	opqdonut	technomancy: nah
17:05	opqdonut	oh, that's funny
17:05	opqdonut	Exception in thread "main" java.net.SocketException: Network is unreachable (NO_SOURCE_FILE:0)
17:06	opqdonut	I can ping repo1.maven.org and clojars.org just fine
17:06	opqdonut	and browse them too
17:06	opqdonut	and I'm on a basic adsl line
17:09	technomancy	net neutrality?
17:09	technomancy	for Leiningen access you need the Premium Developer Plus DSL bundle.
17:09	opqdonut	har har
17:10	opqdonut	let's try straceing the call
17:10	opqdonut	32316 --- SIGSEGV (Segmentation fault) @ 0 (0) ---
17:11	opqdonut	32316 getsockname(0, 0x7f7c84a8df90, [28]) = -1 ENOTSOCK (Socket operation on non-socket)
17:11	opqdonut	this sounds bad
17:11	kryft	technomancy: Is this something like expensive speaker cables?
17:11	technomancy	right
17:11	sritchie	can anyone recommend a good tool for counting lines of code in a leiningen project?
17:11	opqdonut	hmm, okay, for some reason it's trying ipv6 addresses
17:12	kryft	sritchie: Are you charging per line?
17:12	sritchie	haha, no, they'd soon realize that clojure could all be one line, technically
17:12	sritchie	I'm revamping a rather large python project using clojure, and want to make a point to the team
17:12	amalloy	sritchie: as could anything but python :P
17:13	sritchie	amalloy: true, as I'm comparing to python, maybe my comparison should be: here's our former huge application, in one line.
17:14	raek	opqdonut: could you try renaming your .m2 folder to something else and try again?
17:15	amalloy	sritchie: $ find -name '.clj' \| xargs perl -n -e 'print unless (/^\s(;.*)?$/);' \| wc
17:15	amalloy	is what i use
17:15	opqdonut	raek: same problem
17:15	opqdonut	I even removed all ipv6 addresses from my interfaces
17:19	opqdonut	ah found the debian bug report for this
17:25	amalloy	i see my suggestion was so awesome it broke sritchie's net connection
17:26	opqdonut	yay, workaround
17:59	scottj	Are there any tricks for making lein behave better when there's no internet connection available? does only lein deps cause it to check for updates to dependencies that are SNAPSHOT?
18:01	arohner	scottj: lein uses maven under the hood. Have you looked at the offline settings in http://maven.apache.org/settings.html?
18:01	technomancy	scottj: the Maven API exposes a "don't check for snapshots" setting, but it seems to have no effect in my brief experience with it.
18:02	technomancy	scottj: but if there are no SNAPSHOT deps and dependencies are in ~/.m2, then an Internet connection isn't needed
18:02	technomancy	see also :checksum-deps in lein 1.5
18:25	sritchie	hey all -- when defining a multimethod, what's a clean way to direct, say, both byte-arrays and files to the same method?
18:29	wolverian	is there a nicer way than (every? identity coll)? (apply and coll) doesn't work, of course.
18:30	technomancy	wolverian: there's an open ticket for that
18:30	technomancy	http://dev.clojure.org/jira/browse/CLJ-450
18:35	wolverian	that'd be nice.
18:36	technomancy	I concur.
18:36	amalloy	sritchie: ##(doc derive)
18:36	sexpbot	⟹ "([tag parent] [h tag parent]); Establishes a parent/child relationship between parent and tag. Parent must be a namespace-qualified symbol or keyword and child can be either a namespace-qualified symbol or keyword or a class. h must be a hierarchy obtained from make... http://gist.github.com/808684
18:37	sritchie	amalloy: great, just ended up deriving both from ::fileable
18:37	sritchie	thanks
18:37	amalloy	yeah, that's what i was going to suggest
18:38	sritchie	amalloy: all of this syntax ends up being so clean! I'm feeling much clojure love
18:39	amalloy	heh. and now you can count lines, too, right?
18:55	pppaul	anyone know how to use fixtures?
18:55	mefesto	pppaul: clojure.test fixtures?
18:55	pppaul	yup
18:56	pppaul	when i read about them it tells me to add them to the namespace
18:56	pppaul	i tried in the (ns ) macro, but that gave me errors
18:56	mefesto	pppaul: yeah there is a function called clojure.test/use-fixtures i think
18:57	pppaul	so, i just put that in my file and that's it? (that's what have now... haven't been able to run my tests cus i'm fighting with classpaths again)
18:58	mefesto	pppaul: http://clojure.github.com/clojure/clojure.test-api.html
18:58	pppaul	i've read that
18:58	pppaul	a few times
18:59	mefesto	pppaul: in your test ns you define your fixture functions (ex. my-fixture) then call (use-fixtures :each my-fixture)
18:59	pppaul	ok, maybe i'm doing it right
18:59	pppaul	ug, just crashed lein because my classpath for testing was wrong :(
19:01	mefesto	https://gist.github.com/808773
19:54	sritchie	hey all -- I'm looking for an idiomatic way to take every odd element in a collection
19:55	technomancy	,(doc keep-indexed)
19:55	clojurebot	"([f coll]); Returns a lazy sequence of the non-nil results of (f index item). Note, this means false return values will be included. f must be free of side-effects."
19:55	amalloy	sritchie: take-nth
19:56	amalloy	or wait, do you mean every second element, or every element whose value is odd?
19:56	sritchie	every second element
19:56	amalloy	&(take-nth 2 (range 20))
19:56	sexpbot	⟹ (0 2 4 6 8 10 12 14 16 18)
19:56	sritchie	,(take-nth 2 (range 24))
19:56	clojurebot	(0 2 4 6 8 10 12 14 16 18 ...)
19:56	sritchie	cool, that's great
19:57	sritchie	thanks a lot
19:57	sritchie	I'm slowly absorbing the sequence API
19:57	brehaut	,(take-nth 2 (rest (range 24))
19:57	clojurebot	EOF while reading
19:57	amalloy	sritchie: nobody can absorb it all, except chouser
19:57	brehaut	,(take-nth 2 (rest (range 24)))
19:57	clojurebot	(1 3 5 7 9 11 13 15 17 19 ...)
19:58	sritchie	doc in the repl is extraordinarily helpful
19:58	brehaut	hah
19:58	brehaut	&(doc apropos)
19:58	sexpbot	java.lang.Exception: Unable to resolve var: apropos in this context
19:59	amalloy	$findfn 2 (range 10) [0 2 4 6 8]
19:59	sexpbot	[clojure.core/take-nth]
19:59	amalloy	sritchie: sexpbot also has a handy little thing for you
19:59	brehaut	amalloy: what are the limits on the syntax for $findfn ?
19:59	amalloy	it's not very flexible, but it can help if you know what you're looking for but not what to call it
19:59	sritchie	amalloy: wow, that's really handy
19:59	brehaut	sritchie: also look at at clojure.repl/apropos
20:00	amalloy	brehaut: no particular limits. it takes N forms, interprets the last as a return value and the rest as args
20:00	brehaut	it lets you search all the currently loaded ns's for a regexp or string
20:00	amalloy	it evals each form in the same sandbox it uses for ##(count "other stuff")
20:00	sexpbot	⟹ 11
20:01	amalloy	i suppose i could try permuting the arguments so that $findfn (range 10) 2 [0 2 4 6 8] also worked
20:01	technomancy	$find-fn false (loop [] (recur))
20:01	technomancy	ooops; missing a # there
20:01	brehaut	$findfn (fn [a b] (cons b a)) [] [1 2 3] (3 2 1)
20:01	sexpbot	java.security.PrivilegedActionException: java.lang.ClassCastException: java.lang.Integer cannot be cast to clojure.lang.IFn (NO_SOURCE_FILE:0)
20:02	technomancy	was hoping to have it solve the halting problem
20:02	sritchie	haha
20:02	sritchie	and then the forum would implode
20:02	brehaut	$findfn (fn [a b] (cons b a)) [] [1 2 3] (list 3 2 1)
20:02	sexpbot	java.lang.Exception: Unreadable form
20:03	brehaut	whats unreadable about that?
20:03	amalloy	brehaut: dunno
20:03	amalloy	technomancy: it should be eval'ing this in a timeout context. i don't quite see why you never got an exception
20:04	technomancy	ruh roh
20:04	brehaut	somewhere a core is getting hot
20:07	amalloy	lol, no, it probably just doesn't send a message in certain error cases. tbh it's been a while since i wrote this
20:07	amalloy	also sritchie, you can /msg sexpbot if you want to try things like findfn without cluttering #clojure
20:07	brehaut	while that is clearly better, its a ilittle disappointing :P
20:07	amalloy	yeah
20:08	sritchie	amalloy: good advice
20:08	sritchie	amalloy: is there anything like sexpbot out there?
20:09	sritchie	amalloy: what's the competition look like?
20:09	amalloy	sritchie: clojurebot
20:09	amalloy	,'meeeeee
20:09	clojurebot	meeeeee
20:09	amalloy	and there are loads of irc bots for various languages
20:10	brehaut	hash haskell has a pretty cool one
20:10	amalloy	$help heval
20:10	sexpbot	amalloy: Evaluates Haskell code with mueval.
20:10	bhenry1	the sexpbot irc room needs a name change to avoid the occasional creeper in there.
20:10	amalloy	bhenry1: yeah, that's funny stuff
20:12	brehaut	heval take 10 $ let fib = 1 : 1 : zipWith (+) fib (tail fib) in fib
20:13	brehaut	uh oh
20:13	brehaut	do i need to message it directly to sexpbot?
20:16	amalloy	brehaut: $heval
20:16	brehaut	$heval take 10 $ let fib = 1 : 1 : zipWith (+) fib (tail fib) in fib
20:16	sexpbot	⟹ [1,1,2,3,5,8,13,21,34,55]
20:16	brehaut	aha
20:17	amalloy	sexpbot's commands need to be prefixed with $, or anything /msged will be tried as a command
20:17	brehaut	ah right of course
20:37	defn	anyone have the "State of the Art" setup for vim and clojure? Don't worry guys, it's for a friend... :D
20:37	brehaut	a friend eh :P
20:39	defn	brehaut: i have my hand over my heart as we speak
21:37	bartj	is it possible to partition a sequence using a regex ?
21:37	bartj	grr, re-seq
21:38	brehaut	bartj: can you rephrase the question? its a bit ambigious
21:39	brehaut	perhaps and example of your hypothetical inputs and outputs?
21:40	bartj	brehaut, basically a re-split which works on sequences ?
21:41	bartj	eg: (func #"^$" ["a" "b" "" "c" "d"])
21:41	bartj	would give (("a" "b") ("c" "d"))
21:45	brehaut	bartj: hmm. partition-by will get you half way there
21:46	brehaut	,(partition-by empty? ["a" "b" "" "c" "d"])
21:46	clojurebot	(("a" "b") ("") ("c" "d"))
21:47	bartj	brehaut, that's exactly what I need, thanks!
21:48	brehaut	bartj: its not _exactly_ it
21:48	brehaut	a: the pivot is still present
21:48	brehaut	b: it'll pivot again if another empty string is found
21:49	brehaut	,(partition-by empty? ["a" "" "b" [] "c"])
21:49	clojurebot	(("a") ("") ("b") ([]) ("c"))
21:49	bartj	though, I am stumped by:
21:49	bartj	,(empty? '())
21:49	clojurebot	true
21:49	brehaut	the empty list is empty?
21:50	brehaut	partition-by partitions the seq whenever the result of the fun its provided changes
22:01	brehaut	bartj: heres an exceedingly scruffy pivot-on fun that does what you wanted https://gist.github.com/808961
22:02	brehaut	and by exceedingly, it looks like a cat wandered into a text editor and vomited up square brackets
22:03	brehaut	the clearly ignores the obvious optimisation of just using the remainder as 'r' as soon as the pivot predicate is matched
22:08	pdk`	could always be fancy and write (vec this that...) everywhere
22:08	pdk`	if those []s bug ya
22:09	pdk`	i assume this is for a functional version of quicksort or something?
22:09	brehaut	pdk oh man i hope not :P
22:10	bartj	brehaut, thanks, I think I'll stumble through
22:10	bartj	I've used:
22:10	brehaut	bartj: i have a loop recur version that is cleaner :P
22:11	bartj	brehaut, thank you, I am done with what I want :)
22:11	brehaut	haha
22:11	brehaut	no problem
22:12	brehaut	i just couldnt leave the scruffy version as the only version :P
22:14	gfrlog	(doc vec)
22:15	gfrlog	,(doc vec)
22:15	clojurebot	"([coll]); Creates a new vector containing the contents of coll."
22:15	clojurebot	Execution Timed Out
22:15	gfrlog	I knew (vec this that) didn't make sense
22:16	brehaut	gfrlog: it'll fit right in with my terrible function implemention :P
22:20	pdk`	well
22:20	pdk`	vector :p
22:20	bhenry	,(vector :this :that)
22:20	clojurebot	[:this :that]
22:20	pdk`	hmm
22:21	pdk`	(fn [[[l r] pivoted] v]
22:21	pdk`	what exactly is going on with the destructuring for the fn args there
22:21	brehaut	yikes man, no need to go spread that mess about :P
22:21	pdk`	it's a learning exercise :p
22:21	brehaut	ok
22:22	brehaut	(pivot-on empty ["a" "b" "" "c" "d"]) ;=> [["a" "b"] ["c" "d"]]
22:22	brehaut	in the reduce, im using a vector as a pair for my accumulator
22:23	brehaut	the pair consists of the final result (another pair of vectors) and a flag to tell the reducer if it has pivoted
22:25	brehaut	pdk`: does that make it clearer?
22:26	bartj	,(doc pivot-on)
22:26	clojurebot	Excuse me?
22:26	bartj	huh?
22:27	zippy314	Hi. How do I determine if a cloujure object is callable as a function?
22:27	brehaut	bartj: its in that gist i made
22:27	bartj	oh
22:27	zippy314	i.e. would return true for a keyword but false for an integer
22:27	bartj	zippy314, ASFAIK most clojure objects are functions
22:27	bartj	eg:
22:27	brehaut	,(fn? inc)
22:27	clojurebot	true
22:27	brehaut	,(fn? {})
22:27	clojurebot	false
22:28	brehaut	(ifn? {})
22:28	bartj	,(doc keyword?)
22:28	brehaut	,(ifn? {})
22:28	clojurebot	"([x]); Return true if x is a Keyword"
22:28	clojurebot	true
22:28	bartj	zippy314, here is my version:
22:28	bartj	(if (keyword? obj) true false)
22:29	brehaut	zippy314: ifn? for anything callable, fn? for actually a function
22:29	zippy314	thats it!
22:29	zippy314	,(doc ifn?)
22:29	clojurebot	"([x]); Returns true if x implements IFn. Note that many data structures (e.g. sets and maps) implement IFn"
22:29	brehaut	bartj: that can be reduced to (keyword? obj)
22:29	phenom_	out of curiosity, what is the plan for binding in clojure 1.3?
22:31	dnolen	phenom_: plan ?
22:32	phenom_	dnolen: rebinding a function in a binding clause no longer works in 1.3, but does for 1.2
22:33	brehaut	phenom_: by default. if you add the :dynamic meta to the var it does
22:34	phenom_	hmmm, correct me if im wrong (and i mean in that in the newbie sense not in the pretentious sense) but does that mean I need to know ahead of time which functions will potentiall be rebound and add the :dynamic meta info ?
22:34	brehaut	yes
22:34	pdk`	i was asking more about the args list for the fn in the line i posted bartj
22:35	Razwelles	How do you load a .net dll and use its classes in clojureclr?
22:35	pdk`	how you got 2 levels of nesting going on in the args list so i was wondering how the destructuring there was playing out
22:35	brehaut	phenom_: its good practise to ear-muff your rebindable vars anyway
22:36	phenom_	brehaut: how would stubbing or mocking work for testing ?
22:36	brehaut	phenom_: pass?
22:37	phenom_	brehaut: :S im not familiar with that
22:37	brehaut	phenom_: i mean i dont know sorry
22:38	phenom_	brehaut: kk, thnx
22:41	dnolen	phenom_: known issue, http://clojure-log.n01se.net/date/2010-10-29.html
22:42	dnolen	phenom_: not sure if it's been addressed yet.
22:43	dnolen	phenom_: it can be done by changing the var root binding, but I think it's a bit tedious, thus the discussion about a convenience macro for mocking/stubbing.
22:47	phenom_	dnolen: good catch, thnx :)
22:50	pppaul	anyone here work with lein?
22:50	bhenry	pppaul: what;s up?
22:50	pppaul	i'm trying to do testing in it, but i'm not sure how to set up my namespace
22:50	bhenry	you mean like lein test?
22:52	pppaul	(ns test.solve360.name-fixer is my namespace, but when i run lein test it breaks. my folder structre is /test/solve360/name_fixer.clj
22:53	pppaul	test/ is in my classpath too
22:54	pppaul	also, when lein test is run, lein totally breaks... i need to restart it
22:55	bhenry	try ns solve360.name-fixer
22:55	bhenry	https://github.com/technomancy/leiningen/blob/master/test/test_compile.clj
22:55	bhenry	example
22:55	bhenry	^^^ pppaul
22:57	pppaul	thanks
22:58	pppaul	how would that namespace work? because i have that namespace already taken in my src dir
22:58	bhenry	i think that's why all the leiningen tests have test in the name.
22:59	bhenry	just change the name of your test file to test-name-fixer
22:59	pppaul	seems a bit redundant
22:59	pppaul	thanks, though
23:00	bhenry	pppaul: i agree.
23:00	pppaul	oh, i think i did it wrong... should be solve360.test.name-fixer
23:00	pppaul	i hope
23:01	pppaul	yup, works
23:02	bhenry	cool.
23:02	bhenry	i haven't done too much with testing. should probably get in the habit of that.
23:06	pppaul	i really need to test, i can't trust my code
23:31	brehaut	technomancy: does lein backup the 'lein' script when you do upgrade ?