#clojure – 2010-10-03

2010-10-03

01:58	sam1234	j
01:59	amalloy	sam1234: k?
01:59	sam1234	accidently pressed enter
02:08	shachaf	sam1234: Or perhaps ^J?
02:53	amalloy	technomancy: no more TeX? you get that problem figured out, then?
04:02	ashleyw_	Hi, is there a way to expand a vector of values into the arguments of a method? e.g. something like this http://pastie.org/1196308
04:08	LauJensen	ashleyw: Im not 100% sure what you're trying to do, but generally I'll recommend you to use reduce instead of apply if you want t communicate that you're doing some kind of accumulation (instead of spreading of args). If you simply want to add all of the numbers given in 2 vectors, you could
04:09	LauJensen	(defn addnums [[a b]]
04:09	LauJensen	(+ (reduce + a)
04:09	LauJensen	(reduce + b)))
04:09	LauJensen	Where the argumentlist itself is splitting the vector into 2 subvectors. Otherwise, make some kind of flattening operation
05:53	_ulises	morning all
05:54	raek	mornin'
06:33	zmyrgel`	how can I specify that I want symbols value?
06:33	zmyrgel`	I have defined some symbols as (def NORTH 1), (def SOUTH -1) etc.
06:34	_ulises	and you want the 1 for NORTH?
06:34	zmyrgel`	then I get error "can't cast Symbol to Number" in code like (map #(some-function x %) directions)
06:34	zmyrgel`	directions is a list of (NORTH SOUTH)
06:35	zmyrgel`	I assumed it would use the values of the symbols and not the symbols themselves
06:35	mrBliss	zmyrgel`: try a vector instead of a list
06:36	mrBliss	zmyrgel`: or (list NORTH SOUTH)
06:36	mrBliss	zmyrgel`: you're probably quoting your list, so the symbols don't get evaluated
06:36	zmyrgel`	mrBliss: yep, just noticed the quoting thing
06:37	raek	(def north 1) (def south -1) (def directions [north south]) (map #(some-function x %) directions)
06:37	_ulises	this seems to work for me: http://pastebin.com/1n8vxhjC
06:37	_ulises	but maybe I'm not understanding the question :/
06:37	_ulises	(yes, I used a vector instead of a list)
06:39	raek	zmyrgel`: be sure that you understand the difference between (a b c), '(a b c) and (list a b c)
06:39	zmyrgel`	raek: yep, I know that. I just had made the error some time ago
06:42	zmyrgel`	is there any good method of profiling code in clojure?
06:43	mrBliss	zmyrgel`: visualvm is quite easy
07:25	LauJensen	How is the Clojure-Android story ?
07:29	kjeldahl	Is there a new story? The most up-to-date info I've found (with references to github etc) is here: http://stackoverflow.com/questions/973386/android-adverse-to-dynamic-languages
07:30	LauJensen	kjeldahl: I don't know what the official story is on GC, but I know its fixed in several proprietery alternatives to Dalvik
07:49	leafw	is there any lazy reverse fn?
07:51	LauJensen	not possible
07:51	leafw	I believe you, but is there any ratinale wy
07:51	leafw	why
07:52	leafw	also: why :^static fn can only accept ^long and ^double, and not ^int for example, in its args?
07:55	cemerick	leafw: because ^int doesn't give you anything beyond what ^long does.
07:55	leafw	cemerick: you mean performance-wise, they are identical?
07:56	leafw	I can believe that in 64-bit CPUs. Fair enough.
07:56	cemerick	Most JVMs are optimized for doubles and longs these days, yeah.
07:56	cemerick	But that's not the point.
07:56	leafw	math with float is faster than with double. Test it yourself.
07:56	leafw	(in java world, at least)
07:58	cemerick	the Sun jvm is biased towards doubles. Heard it from someone at Sun with my own ears.
07:58	cemerick	Regardless, that's not all that went into it.
07:58	LauJensen	leafw: If you have a sequence X which you want to reverse, you have to start with the last element in that list. To realize that last element, you also need to realize all the elements leading up to it, hence we're no longer lazy
07:59	leafw	cemerick: float is faster (and less accurate). That's a fact.
07:59	cemerick	Well, if you say so, then I guess so. :-P
07:59	leafw	LauJensen: but if one has a list of objects that are all realized, i.e. a non-lazy list, then reverse could be lazy, returning them as it goes. Or is it that what it does current?
08:00	LauJensen	leafw: Wouldn't change anything
08:00	leafw	LauJensen: ok
08:00	leafw	cemerick: YMMV. I work with float[] images ... math is about 1.5x faster than with double[].
08:02	cemerick	leafw: This is an undecidable point with just the two of us dropping anecdotes.
08:02	cemerick	And entirely unrelated to the question of :^static's impl.
08:02	LauJensen	leafw: As I recall, Richs argument was that doubles on the JVM are now typically as fast, sometimes faster than floats
08:04	leafw	LauJensen: thanks.
08:07	LauJensen	np
08:10	kjeldahl	LauJensen: Regarding Dalvik, your information is probably more updated than mine.
08:19	_ulises	how can I dispatch mult-methods based on some property of their 2nd argument? i.e. if i do (defmulti foo :type) then :type is used on the first arg. passed to foo, not its second
08:26	AWizzArd	_ulises: (defmulti foo (fn [arg1 arg2] [(> arg1 arg2) (class arg2)]))
08:27	_ulises	oh! so I could just do (defmulti foo (fn [_ p2] (:type p2)))
08:27	_ulises	excellent!
08:27	AWizzArd	yes
08:27	_ulises	thanks :)
08:28	AWizzArd	first the defmulti is called, and its return value is compared with all defmethods that are available. The right one is selected and then this specific foo is called again with the original arguments
08:30	_ulises	so first the defmulti is called with all the arguments passed?
08:30	_ulises	and then the appropriate method is called, again, depending on the result of the dispatch-fn, with all the arguments passed?
08:33	tobiasraeder	is it possible to (deftype person [proxy firstname lastname] ...) and then do something like (person. "firstname" "lastname")?
08:33	leafw	tobiasraeder: use defrecord for that
08:34	tobiasraeder	@leafw i need to use deftype tho since defrecord doesnt support :volatile-mutable afaik
08:35	leafw	it doesn't :) that's why defrecord is so useful. With deftype and volatile-mutable, you are back into mutable java world.
08:35	leafw	but you knew that already, nm.
08:35	tobiasraeder	yeah, but since this is a huge java interop thing im working on i don't really have a way to bypass it :/
08:36	LauJensen	tobiasraeder: Looks like its possible with a factory fn ?
08:37	tobiasraeder	@LauJensen yeah i guess so
08:51	LauJensen	Any tomcat buffs in here ?
09:13	raek	tobiasraeder: just out of curiosity: did you try the atom solution? if so, what lead you to your decision?
09:30	tobiasraeder	@raek im using deftype with :volatile-mutable now because it seems to fit what i need really
09:32	jaley	Hi guys, anyone able to help out a leiningen newb quickly?
09:33	tobiasraeder	just go ahead ;)
09:33	Bahman	Hi all!
09:33	jaley	i've created a projected added dependencies for compojure, enlive and ring, plus dev-dependencies for swank-clojure, when I run lein deps I get an exception for an attempt to use nth on a Symbol
09:34	tobiasraeder	@jaley can u put your project.clj into a gist or something of that sort?
09:34	jaley	sure, off to pastebin, brb
09:36	jaley	@tobiasraeder http://pastebin.com/Urh0x7SJ
09:36	raek	jaley: wrap [swank-clojure "1.2.0"] in another vector
09:36	tobiasraeder	you need a second pair of [] around your [swank-clojure "1.2.0"]
09:37	tobiasraeder	since the dev dependencies are a vector of dependencies aswell
09:37	tobiasraeder	so it should be :dev-dependencies [[swank-clojure "1.2.0"]])
09:37	jaley	@raek @tobiasraeder that got it, thanks guys
09:43	tobiasraeder	is it possible to generate a vector of java object instances in a macro?
09:44	_ulises	is there an alternative to re-starting the swank process whenever one needs to redefine things such as multi-methods?
09:45	Raynes	(ns-unmap ns 'multi)
09:46	_ulises	thanks! :)
09:50	tobiasraeder	is there a way to have a function return (MyJavaClass. "parameter") ?
09:52	raek	tobiasraeder: you want the class to be given as an argument?
10:00	raek	tobiasraeder: (.newInstance (.getConstructor MyJavaClass (into-array Class [String])) "parameter")
10:00	Raynes	http://stackoverflow.com/questions/3748559/clojure-creating-new-instance-from-string-class-name
10:01	tobiasraeder	thanks :) i managed to get it to work a bit different, ill gist it real quick. comments are more than welcome
10:03	tobiasraeder	http://gist.github.com/608591
10:34	tobiasraeder	if i have something along those lines (defn myfn [name] '(MyClass. name)) is there any way to get the value of name as first parameter to MyClass. without making it a macro?
10:37	_ulises	can I get your thoughts on "purely functional data structures" by Chris Okasaki? (if anybody has read it)
10:50	raek	tobiasraeder: with java's reflection API, I think you can
10:51	raek	either by doing something like what I posted before, or by using clojure.lang.Reflector/invokeConstructor
10:52	raek	that will not result in any Clojure code that uses the (MyClass. name) syntax, though
10:52	tobiasraeder_	back
10:53	tobiasraeder_	@raek the problem was the quoted form didnt evaluate the name parameter (obviously)
10:53	raek	tobiasraeder: http://pastebin.com/zWTU6NVf
10:53	tobiasraeder_	@raek the (for me) easiest solution seems to be (defn myfn [name] `(MyClass. ~name))
10:54	raek	yes, a macro solution would be one of the simplest ways, I think
10:54	tobiasraeder_	well its a function since i wanna pass it to a map inside another macro
10:54	raek	`Integer
10:54	raek	,`Integer
10:54	clojurebot	java.lang.Integer
10:54	tobiasraeder_	i just didnt realize you could do `~ etc in fns aswell
11:11	jaley	can defrecord take a doc string?
11:19	raek	jaley: no. the doc strings goes on the protocols.
11:20	raek	just like docstrings goes on the defmulti rather than the defmethods
11:20	jaley	@raek right.. that makes me think perhaps i shouldn't use a record for plain old data?
11:21	raek	records are plain old data, with a type
11:22	raek	but if you don't need dynamic dispatch on type, records might be overkill
11:22	jaley	@reak ok.. as i was hoping - but it doesn't make sense to document them?
11:22	raek	what a method should do is defined by the protocol
11:23	jaley	@raek yeah, but I just wanted to note what a collection of fields represent. guess it's doesn't really matter
11:24	raek	I usually describe "data contracts" in the ns docs
11:24	tobiasraeder_	can you define a macro that in turn defines a function like that: (defn create-<firstparameter-of-macro> [& args] (new <firstparameter-of-macro> args)) ?
11:24	tobiasraeder_	i seem to fail horribly :/
11:24	jaley	@raek that works - thanks
11:25	raek	for example, a x-map is a map that has the :a, :b and :c keys, which has the meaning...
11:26	raek	statically typed languages has the advantage that the type is the intuitive place to put the docs at
11:26	raek	but clojure does very much or most of it's things without types
11:27	jaley	@raek yeah that makes sense, i was just look for a sensible place to document what a bunch of fields meant. the ns docstring will be fine
11:28	raek	jaley: an example: http://github.com/raek/quirclj/blob/master/src/se/raek/quirclj/message.clj
11:30	mrBliss	tobiasraeder_: my try: (defmacro defn-create [x] (list 'defn (symbol (str "create-" x)) (vector '& 'args) (list 'new `~x 'args)))
11:31	raek	tobiasraeder_: the problem with the 'new' special form is that you cannot 'apply' it
11:31	mrBliss	tobiasraeder_: I wouldn't use it though ;-)
11:32	tobiasraeder_	@raek in which sense "apply"?
11:32	raek	mrBliss: when calling (create-foo 1 2 3), wouldn't that try to call (new foo [1 2 3]) instead of (new foo 1 2 3)?
11:33	mrBliss	raek: that's what tobiasraeder_ asked
11:34	mrBliss	a better solution: http://gist.github.com/608662 (not mine)
11:34	raek	I think using java's reflection capabilities is the only general approach
11:35	raek	unless you have fixed arities for your constructors
11:35	tobiasraeder_	well i know the arguments that will be passed to the ctor in the macro, yes
11:36	raek	if you have a seq s of arguments to a function f, you invoke the function with (apply f s)
11:36	Raynes	http://stackoverflow.com/questions/3748559/clojure-creating-new-instance-from-string-class-name I did the same thing in my answer. It doesn't work quite as well as I'd hoped though (read the comments).
11:37	raek	but with a special form like new, there is no such apply function
11:37	tobiasraeder_	alright, thank you
11:58	_ulises	what's the best way of making a single parameter entirely optional? is it using ?
11:58	_ulises	or &?
12:00	raek	one way is (defn foo ([x] (foo x 0)) ([x y] (+ x y)))
12:01	raek	another way could be (defn foo [x & [y]] (+ x (or x 0)))
12:03	mrBliss	,(binding [inc dec] (map inc [1 2 3]))
12:03	clojurebot	(0 1 2)
12:03	mrBliss	,(binding [inc dec] (map #(inc %) [1 2 3]))
12:03	clojurebot	(2 3 4)
12:03	raek	or (defn foo [x & more] (+ x (get more 0 0)))
12:04	jaley	is there a quick way of sending an (in-ns <current namespace>) to the repl with swank?
12:04	raek	(some math operations are static functions, and are optimized by the compiler)
12:04	mrBliss	How can I make my second form do the same as the first? (difference: rebound var inside a anonymous function)
12:04	mrBliss	jaley: C-c M-p
12:05	jaley	mr8liss: merci beaucoup
12:06	mrBliss	solving my problem ^^ would make testing some functions a whole lot easier
12:07	raek	maybe something like (ns ... (:refer-clojure :exclude (inc))) (def inc clojure.core/inc)
12:08	_ulises	raek: thanks
12:09	mrBliss	raek: seems to work in my example, I'll try to incorporate it in my tests. Thanks
12:09	raek	_ulises: I think the first one is the most common
12:10	_ulises	excellent, thanks again.
12:11	rrc7cz`	how can you apply on a Java method? something like (apply (.fillRect gfx) [10 10 20 20])?
12:11	raek	mrBliss: some of the functions in clojure.core is allowed to be inlined. by making the ns of the var being something else than clojure.core, the compiler will not do any of those optimizations
12:12	mrBliss	raek: it was just an example, I only have to rebind my own functions
12:12	raek	then it should always work.
12:13	raek	,(binding [first second] (map first [[1 2] [3 4]]))
12:13	clojurebot	(2 4)
12:14	raek	,(binding [first second] (map #(first %) [[1 2] [3 4]]))
12:14	clojurebot	(1 3)
12:14	raek	hrm
12:14	raek	maybe this behaviour happens for all vars in clojure.core
12:14	raek	,(binding [first second] (doall (map #(first %) [[1 2] [3 4]])))
12:14	clojurebot	java.lang.StackOverflowError
12:15	mrBliss	raek: it's actually quite complicated: the functions I need to rebind are private, in another namespace and I need to use them in the functions that replace them. I might have some problems later when I'll have to rebind functions in futures and pmaps...
12:16	raek	private functions can be used from other namespaces with @#'the-fn
12:16	raek	very useful in testing...
12:16	mrBliss	I'm using with-private-fns for that
12:17	raek	anyway, lookout for binding things in clojure.core and using binding with lazy sequences
12:18	raek	if the lazy sequences escape out of the binding form, the any usage of the rebound var in the lazy seq will see the "outer" value
12:19	raek	mrBliss: also, have you seen this? http://github.com/marick/Midje
12:22	mrBliss	I only need it in my tests, so I don't think I'll run into that problem. Yes I've seen Midje and put it on my todo list. But I'd rather switch to lazy-test when that's final. Thanks for the help
12:25	solussd_	hm. when did clojure start supporting metadata on functions?
12:26	raek	in 1.2
12:26	solussd_	neat
12:34	ldhanson2	i've got a map on which I call (keys my-map), which gives me a clojure.lang.APersistentMap$KeySeq. I'm trying to use do/recur to loop through each key in the sequence, but calling 'first' on the key sequence gives "java.lang.ClassCastException: clojure.lang.PersistentHashMap cannot be cast to java.util.Map$Entry at clojure.lang.APersistentMap$KeySeq.first (APersistentMap.java:130)". Am I missing something?
12:39	mrBliss	ldhanson2: can you post the code?
12:56	ldhanson2	mrBliss: i'm currently trying to isolate the code to reproduce it, it's not occurring in simple test scenarios. doing so may well help me discover the problem, but i was hoping it rang a bell for somebody
12:59	mrBliss	ldhanson2: rather strange error message
13:02	ldhanson2	mrBliss: indeed. before I call first on it i even (assert (seq? keyseq)) so one would think that's not a problem
13:09	carkh	a question about protocols and namespaces : i have a protocol Matcher in namespace A, it has a single method "match". I want to use it from namespace B but i have reflection warnings. Is there a way to "import" the protocol into namespace B so that i don't have to type hint with the fully qualified and very long name of namespace A ?
13:28	carkh	so here is the answer to my question : do not use the namespace A as the Matcher name will be to the var created by the defprotocol form
13:29	carkh	inted i need to require the namespace, then import A.Matcher
13:29	carkh	instead*
13:29	carkh	which is quite a lot of work =/
13:31	bhenry	carkh i'm pretty sure you can import just the matcher without the require, unless you need other stuff in namespace A.
13:32	carkh	i get a class not found
15:16	tobiasraeder_	Anyone got an idea why i might get "Class clojure.lang.Reflector can not access a member of class MyObjectProxy with modifiers "public"" when i try to create an instance of MyObjectProxy?
15:17	tobiasraeder_	it got a public constructor (whcih is pretty obvious i guess) which im trying to call
15:19	tobiasraeder_	solved it
16:01	leafw	is zipmap lazy?
16:02	leafw	it's doing a loop/recur ... likely not
16:02	leafw	or is loop lazy?
16:02	AWizzArd	not lazy
16:02	AWizzArd	loop + recur = goto
16:02	leafw	thanks
16:03	leafw	I just blowed up 12 Gb of RAM, and I was wondering. zipmap looks like the problem
16:03	leafw	also: is (apply merge-with + ... not lazy either? I.e. will all maps be generated, and then apply applies?
16:03	AWizzArd	leafw: how much RAM do you have in your box?
16:04	leafw	this one has 16 Gb
16:04	AWizzArd	nice
16:04	leafw	useful :)
16:04	AWizzArd	ahead the average
16:04	leafw	oh no, this machine is actually small here. Standard is more like 24 or higher
16:05	hiredman	the only lazy thing in clojure is lazy-seqs
16:05	leafw	hiredman: map is lazy, reduce is lazy ... or arent' they?
16:05	hiredman	well my laptop has 8 gigs, so …
16:05	AWizzArd	also very nice
16:05	leafw	a laptop with 8 Gb is a treat
16:05	hiredman	map produces a lazy seq so it is lazy
16:05	hiredman	reduce is not lazy
16:05	leafw	hiredman: what about reduce?
16:05	leafw	ok
16:05	leafw	thanks
16:05	hiredman	the only lazy thing in clojure is lazy-seqs
16:05	AWizzArd	in 1990 a real good computer had maybe 8 mb of RAM, so we got a factor of 1000 in two decades :)
16:06	AWizzArd	every two years a doubling, which will give us average RAM of around 100 gb by the end of this decade
16:06	hiredman	to be fare I had to buy more ram because right now even with 8gigs I only have 90mb free
16:06	hiredman	fair
16:06	AWizzArd	uhm, what are you doin? :)
16:07	leafw	hiredman: "for" is also lazy
16:07	AWizzArd	Btw, a very good time to buy ram now, very nice price
16:07	hiredman	leafw: for produces a lazy-seq
16:07	leafw	I see
16:07	hiredman	the only lazy thing in clojure is lazy-seqs
16:07	leafw	hiredman: I get it :)
16:07	AWizzArd	for is built with lazy-seq's under the hood
16:08	leafw	do all fn that return a lazy-seq are labeled as lazy in their docs?
16:08	leafw	is there a list of lazy fn to look at?
16:08	AWizzArd	leafw: you can do (source your-fn) and have a quick look
16:08	leafw	AWizzArd: (find-doc "lazy") lists a few
16:08	AWizzArd	There is no regularity in the docs I think
16:11	jaley	hey guys - i was wondering.. what's the best way to organize test data files in a leiningen project? can i somehow avoid making assumptions about the working directory?
16:13	jk_	AWizzArd: where does "source" come from? if i type (source _something_) in the repl, source is not found
16:14	freakazoid	has anyone built a transactional map yet? one where concurrent updates, including inserts and deletes, to different keys don't conflict?
16:14	freakazoid	I've thought of a way to do it but I wanted to make sure it hadn't already been done first
16:18	technomancy	jaley: usually you just have the same structure but with _test.clj tacked on to the end
16:18	technomancy	each implementation namespace corresponds to one test namespace
16:22	jaley	@technomancy yeah actually I wanted to have some extra text files with test data in them
16:22	technomancy	jaley: test-resources is for those
16:22	jaley	@technomancy and was wondering if there's a standard place to keep test data
16:22	leafw	ok, so it's clear: apply is evil ... realizes all sequences first, then applies.
16:23	jaley	@technomancy ah ok, in the root of the test folder? is there some documentation about this somewhere that i've missed?
16:23	technomancy	jaley: in the root of the project
16:23	jaley	@technomancy great - thanks!
16:23	technomancy	it's documented in sample.project.clj; I guess it should be in the tutorial too
16:25	jaley	@technomancy it's fine - i even had that open in another tab, just totally missed it (d'oh!). thanks again
16:25	technomancy	sure
16:30	chouser	leafw: just got here, but ... what? apply is lazy.
16:30	leafw	chouser: then I don't get it. Now it's churning away at 1 GB of RAM ... no more blowing up
16:30	chouser	that is, apply doesn't realize the seq you give it any more than is required by the function you're passing it to.
16:31	leafw	let me paste it somewhere
16:32	chouser	,(let [f (fn [a b c & more] (+ a b c (first more)))] (apply f (range 10)))
16:32	clojurebot	6
16:32	chouser	er
16:32	chouser	,(let [f (fn [a b c & more] (+ a b c (first more)))] (apply f (range)))
16:32	clojurebot	6
16:32	leafw	wasn't there a paste bin where #clojure was listed? not paste.lisp.org?
16:32	chouser	leafw: paste.lisp.org or gist.github.com
16:33	leafw	thanks
16:35	chouser	ah, apply on merge-with will consume the whole seq
16:35	leafw	http://gist.github.com/608905
16:36	chouser	the problem there is merge-with, which returns a map, and maps aren't lazy.
16:36	technomancy	it's not that apply is lazy, it's that lazy functions can be applied
16:36	leafw	any nicer ways to write the second version?
16:37	chouser	if any link in the chain is eager, the input seq will be fully realized. apply won't do that, but something else (like merge-with) certainly can.
16:37	leafw	chouser: I was expecting merge-with to only care about two maps at a time
16:37	chouser	intersting. reduce will certainly realize the whole input seq to
16:37	leafw	and to forget all maps that it has already seen
16:37	chouser	too
16:37	chouser	hm
16:38	leafw	well, at least I have a solution, but it's not pretty to say the least
16:39	leafw	I'm computing node centrality in a tree (graph without loops) that represents a neuron ... looks way cool in 3D
16:41	chouser	:-) I'm sure
16:47	AWizzArd	leafw: how deep is your tree?
16:52	leafw	AWizzArd: about 3000 nodes
16:52	leafw	with 88 branch points
16:54	AWizzArd	so, do you mean the deepest point is 88 nestings?
16:56	AWizzArd	leafw: (reduce (fn [l _] (cons l ())) (range 88)) or range 3000?
16:58	leafw	no, deepest degree is about 500 or so
16:59	AWizzArd	ok
17:06	chouser	hm, well I don't see where merge-with would be holding the head of the seq.
17:08	alistairk	Can anyone point me (a Clojure newbie) in the right direction to install Clojure 1.2 with Emacs on Ubuntu 10.04
17:08	alistairk	?
17:09	leafw	chouser: no idea either. The whole script is here: http://github.com/acardona/Fiji-TrakEM2-scripts/blob/master/TrakEM2/backbone_.clj
17:09	AWizzArd	alistairk: the Emacs Starter Kit could be a good point to start
17:09	AWizzArd	alistairk: basically you download a folder .emacs.d and put it into your home, and make sure that there is no .emacs file in your home.
17:10	technomancy	alistairk: cljr is a nice place to start experimenting, then once you want to create a project look at leiningen
17:10	technomancy	also relevant: http://www.assembla.com/spaces/clojure/wiki?id=clojure&wiki_id=Getting_Started_with_Emacs
17:12	alistairk	Thanks guys - I'll check your suggestions out. The problem is specifically with Clojure 1.2 - I have Emacs and Clojure 1.1 working
17:12	alistairk	but obviously want to start playing with Clojure 1.2
17:12	alistairk	;-)
17:12	jaley	what would cause a NoSuchMethodError <init> when trying to instantiate a record? :s
17:13	technomancy	it should be easy to swap in 1.2 in your project.clj
17:16	leafw	good night all
17:16	leafw	thanks for the help and the company.
17:26	jaley	hm weird. i did a lein clean and recompiled and the problem went away
17:26	jaley	it's nice when that happens
17:32	florianjunker	Is there some way to use leiningen with a SOCKS proxy?
17:40	alistairk	(progn (load "/home/alistairk/.emacs.d/elpa/slime-20100404/swank-loader.lisp" :verbose t) (funcall (read-from-string "swank-loader:init")) (funcall (read-from-string "swank:start-server") "/tmp/slime.7839" :coding-system "iso-latin-1-unix"))
17:40	alistairk	Clojure 1.2.0
17:40	alistairk	user=> java.lang.Exception: Unable to resolve symbol: progn in this context (NO_SOURCE_FILE:1)
17:41	technomancy	alistairk: you're launching slime in some way that assumes you're working with common lisp
17:41	technomancy	have you read the swank-clojure readme?
17:41	alistairk	Sorry, will do. Is Emacs Starter Kit 'compatible' with Clojure 1.2.0?
17:42	technomancy	sure; they're orthogonal
17:42	alistairk	Brilliant! Thanks for your patience...
17:42	alistairk	I'll get there ;-)
17:45	alistairk	technomany: by the way - I like how the Emacs Starter Kit configures Emacs - very nice
17:46	alistairk	Installing Leiningen...
17:55	lpetit	Hmmm
17:57	lpetit	I need help understanding Open source license issue
17:58	lpetit	Say I have written a program under the EPL. This program includes an antlr grammar for clojure.
17:59	lpetit	Now I see the antlr grammar has been included into another program. This program is not licensed under EPL at all. Sort of a BSD-like "own copyright without any guarantee, etc."
17:59	lpetit	A
18:00	lpetit	At least there's credit for my work in a ATTRIBUTION.txt file.
18:00	alistairk	Installed Leiningen. Added the swank-clojure "1.2.1" dependency to the project.clj file and ran 'lein swank [PORT=4005] [HOST=localhost]'
18:00	lpetit	Isn't this a violation of the EPL ?
18:00	lpetit	I'm referring to this : http://www.eclipse.org/legal/eplfaq.php#USEINANOTHER
18:00	alistairk	Exception in thread "main" java.lang.NumberFormatException: For input string: "[PORT=4005]" (NO_SOURCE_FILE:1)
18:00	carkh	lpetit: i think they can use your work, but need to include the license for that part
18:00	Chousuke	lpetit: well at least the antlr grammar should be licenced under the EPL AFAIK.
18:01	lpetit	it's part of ccw and ccw is licensed under EPL.
18:01	alistairk	Does that mean that I have to download the Swank Clojure source first (it doesn't download it?)???
18:01	chouser	lpetit: I think that is a violation
18:01	alistairk	Guys, is this the right place for me to be asking these questions?
18:01	carkh	yep they cannot change the license of your code
18:02	lpetit	carkh: they did not, or it's very well hidden
18:02	chouser	lpetit: I'm sure they'd appreciate a friendly reminder to pay attention to license details.
18:02	chouser	alistairk: yep, right place. just hang on. :-)
18:02	lpetit	The ATTRIBUTION.txt file is just an acknowledgement of the provenance of the work, there's no mention of license
18:03	alistairk	thanks - no hurry
18:03	chouser	alistairk: ok, unfortunately I don't use emacs or swank and don't really know what to advise
18:03	lpetit	Well, I'll think about how to turn my answer. Don't necessarily want to get a new enemy on this little earth, just to make things clear :)
18:04	ansra	alistairk: the bit in brackets can be left out, those are showing default values for the port and host, running 'lein swank' should work
18:04	chouser	lpetit: sure. you could offer to grant them a custom license, if you own that whole file
18:05	chouser	lpetit: or just suggest they include an EPL license notice for that one file
18:05	lpetit	chouser: yeah, thanks for the advice.
18:07	alistairk	oops, just running 'lein swank' seems to be working :-)
18:08	chouser	lpetit: since they took your work without permission, it should be possible to be polite and undemanding and still generate a sense of guilt in them. :-)
18:09	alistairk	thanks ansra
18:09	alistairk	yep it seems to be working
18:09	alistairk	thanks guys
18:09	ansra	sure thing, have fun :0
18:10	lpetit	chouser: hmm, last question. Would you ask this to him publicly ? (answering the email on #clojure ?).
18:12	carkh	huh i wouldn't (on the first try)
18:18	lpetit	too late :)
18:18	lpetit	but things were not as bad as I first thought
18:20	carkh	ohwell your message isn't too stern anyways =)
19:10	defn	greetings gents
19:12	Raynes	chouser: He was overreading the README. 'lein swank [PORT=4005] [HOST=localhost]' should have been 'lein swank 4005 localhost'
19:12	Raynes	In case you were wondering, and for future reference. I bet he read that command, and those were just vague placeholders.
19:22	AWizzArd	~seen rhickey
19:22	clojurebot	rhickey was last seen quiting IRC, 4113 minutes ago
19:22	AWizzArd	,(/ 4113 60.0)
19:22	clojurebot	68.55
19:22	Raynes	$seen rhickey
19:22	sexpbot	rhickey was last seen quitting 4113 minutes ago.
19:22	Raynes	sexpbot agrees
19:23	AWizzArd	(:
19:23	AWizzArd	good night
19:23	Raynes	Night.
23:34	notostraca	I am trying to find information on a nice deed the clojure community did -- a bunch of people paid for someone's airfare to a clojure conference, but I can't find anything using google
23:35	notostraca	(trying to evangelize a bit)
23:43	maravillas	notostraca: http://cemerick.com/2010/09/10/a-clojure-scholarship-lets-send-raynes-to-the-conj/
23:43	maravillas	and the next post, http://cemerick.com/2010/09/10/anthony-simpson-will-receive-his-clojure-scholarship-thanks-to-you/
23:44	notostraca	thanks a ton
23:44	maravillas	np