#clojure – 2011-10-04

2011-10-04

01:06	michaelr525	hello
03:01	fliebel	How is IDE support at the moment? I need to wade through some Java code, so I thought I'd check the clojure plugin for whatever IDE I'm going to use.
03:03	khaliG	fliebel, they say Counterclockwise for eclipse is pretty good, i'm just trying to use it at the moment
03:13	pyr	i have a bit of chicken and egg program
03:14	pyr	problem, rather
03:14	pyr	say you have a protocol defined in a project
03:14	pyr	and you want to allow external lib to implement it
03:14	pyr	to be able to let users of the software provide an implementation
03:14	pyr	(say, for a transport or serializer)
03:15	pyr	how do you manage the cross dependency ?
03:17	cark	hum i don't see a problem ... the library user requires the namespace where you defined the protocol, and that's it
03:17	cark	your library only refers to the namespace where the protocol was defined
03:17	cark	there's no cross-dependency
03:17	pyr	hmm let me be more clear
03:18	pyr	i have a daemon, not a library that relies on transport
03:18	pyr	and i want to be able to switch transports from the configuration
03:18	pyr	the daemon defines the transport protocol
03:19	pyr	oh, I guess i see a way of having this work
03:19	pyr	ok
03:19	cark	ok well : file1 = define your protocol
03:19	cark	file2 : define your demon
03:20	cark	file3 : define your transport
03:20	cark	files 2 and 3 are using file 1
03:20	pyr	with files i'm alright
03:20	pyr	it's libraries that get me confused
03:21	pyr	but yeah
03:21	cark	that's the same thing =P
03:21	pyr	it works that way
03:24	rimmjob	is there a style guide for clojure?
03:25	rimmjob	ive never really writen lisp that other humans had to read..
03:26	rimmjob	mostly just for my own use
03:27	pepuchoZ	quit
03:28	pyr	rimmjob: sp indent, let emacs or vim show the way
03:30	rimmjob	i mean like, after when to newline after a parens
03:30	pyr	if you look at existing code a bit
03:30	rimmjob	ill just look at
03:30	rimmjob	ok
03:30	pyr	you'll see there's no definite style
03:31	rimmjob	oh
03:31	pyr	not at all like C / Java / Ruby
03:31	pyr	because the lack of syntax in clojure makes it almost pointless
03:32	pyr	I tend to newline after "block" calls (doseq, do, when, dosync, ...)
03:33	pyr	using the (defn foo \n "" \n [args]) syntax forces you to put some text in that docstring
03:34	rimmjob	ok, i guess that makes sense. thanks
03:42	dbushenko	hi
03:42	dbushenko	what's the difference between fn* and fn?
03:42	Blkt	good morning everyone
03:53	raek	dbushenko: fn* is the "real" more primitive special form. I think it's like fn, but doesn't do destructuring or something.
03:53	raek	dbushenko: it's not meant to be used directly
03:55	dbushenko	raek, yep, but I just want to dig in the sources of clojure and don't get what for is fn*
03:56	dbushenko	so, thanks for your explanaition
04:24	pyr	is there a way to pass a class around and let another function instantiate it ?
04:25	pyr	i.e: (deftype Foo [x y]) (let [class user.Foo] (instantiate-with class :some :args)))
04:43	michaelr525	pyr: you can use java stuff: (Class/forName "java.lang.String")
04:44	pyr	michaelr525: yep, but then there's no instantiation possible
04:46	pyr	(eval `(new ~(symbol class-as-string) arg1 arg2))
04:46	pyr	that works
04:46	michaelr525	it feels like cheating ;)
04:47	terom	pyr: classes can be instantiated with reflection, also (but it's certainly more complicated than eval)
04:48	raek	pyr: you can use (import 'clojure.lang.Reflector) (Reflector/invokeConstructor class (object-array parameters))
04:48	raek	the Reflector class makes this much simpler compared to java's reflection api
04:49	raek	eval is uncessary here. it will just generate code that calls the invokeConstructor method in this case.
04:50	raek	oh, and 'class' here is an instance of java.lang.Class (the thing you get from "class literals" in clojure)
04:50	pyr	(Reflector/invokeConstructor (Class/forName "user.Foo") (to-array [:a :b]))
04:50	pyr	works
04:51	pyr	nice
04:53	pyr	raek: i looked for clojure reflect api earlier
04:54	pyr	raek: no clear doc
05:03	srdjan	why doesn't the following thing work in clojure 1.3: (-> 4 #(* 10 %)), but (* 10 4) works?
05:04	srdjan	I get the following error: #<CompilerException java.lang.ClassCastException: java.lang.Long cannot be cast to clojure.lang.ISeq, compiling:(NO_SOURCE_PATH:1)>
05:05	srdjan	and the following thing forks: (#(* 10 %) 4)
05:05	srdjan	and does clojure support macros inside of macros, or macros in clojure are not composable?
05:06	Blafasel	,(-> 4 (* 10))
05:06	clojurebot	40
05:07	Blafasel	In other words: Your #() is unnecessary and potentially (I'm clueless) wrong.
05:07	srdjan	Blafasel: please, can you explain me why it has to be written in that way?
05:07	Blafasel	-> already does the magic for you, inserting the current value into the forms
05:07	Blafasel	,(4 (println "foo" "bar"))
05:07	clojurebot	#<ClassCastException java.lang.ClassCastException: java.lang.Long cannot be cast to clojure.lang.IFn>
05:08	Blafasel	,(-> 4 (println "foo" "bar"))
05:08	clojurebot	4 foo bar
05:08	Blafasel	,(->> 4 (println "foo" "bar"))
05:08	clojurebot	foo bar 4
05:08	Blafasel	Spot what happens.
05:09	Blafasel	-> inserts the current value into all forms as first argument. ->> as a last argument.
05:10	raek	srdjan: #(* 10 %) is a shorthand for (fn [%] (* 10 %)) and this translation is done at readtime (before macroexpansion time)
05:10	Chousuke	,'(-> 4 #(* 10 %)); srdjan, spot the problem
05:10	clojurebot	(-> 4 (fn* [p1__131#] (* 10 p1__131#)))
05:11	Chousuke	you're basically asking for (fn 4 [x] (* 10 x))
05:11	raek	srdjan: (-> 4 #(* 10 %)) ---> (-> 4 (fn [%] (* 10 %))) --> (fn 4 [%] (* 10 %))
05:12	Chousuke	macros inside macros work just fine, too
05:12	raek	srdjan: another approach for this case is (->> 4 (* 10))
05:12	srdjan	thanks for the explanation
05:13	raek	to thread functions of one argument you can use comp. (it takes the function in the reverse order compared to -> though)
05:14	srdjan	raek: how is you exaple different wiht mine? (using ->> instead of using -> )
05:14	Blafasel	See my println examples. The argument order is different.
05:14	srdjan	Blafasel: thanks
05:14	Blafasel	->> inserts as last element. -> as first argument
05:14	srdjan	i see now
05:14	Blafasel	Doesn't matter of course for *
05:15	raek	,((comp #(- % 1) #(- 10 %)) 5)
05:15	clojurebot	4
05:16	raek	same as (- (- 10 5) 1)
05:22	srdjan	I was writing SKI evlauator in clojure, as an excercise. I am alowed to discuss the pieces of the code in the chat? (the pieces of the code are about 5 lines each)
05:23	srdjan	Am I alowed*
05:26	Fossi	it's better to use a pastebin
05:26	raek	srdjan: it's perhaps simpler paste the complete code at gist.github.com and the refer to the lines
05:31	srdjan	I have my small SKI evaluator on the http://pastebin.com/Pp33AuNh
05:31	srdjan	I have 3 questions related with my code.
05:32	srdjan	in line 50. I have function 'ff' that evaluates S, K, and I rules.
05:33	srdjan	can anybody suggest me how can this code be improved?
05:33	srdjan	in line 58. I have function 'ap' that tries to apply rule 'ff' recursively.
05:33	srdjan	is it better to combine functions 'ff' and 'ap' in one function?
05:34	srdjan	I have noticed that if I have 'recur' in the pattern matching, the compiler complains.
05:35	srdjan	in the line 66 i have function 'st' that walks the SKI tree and appies the SKI rules to each subtree
05:35	srdjan	this give in the end the minimal SKI tree
05:35	srdjan	is function 'st' idiomatic clojure, or I can simplify it?
05:36	raek	ff and ap looks fine to me
05:36	raek	but (reduce conj ...) = (into ...)
05:36	srdjan	function 'f' was my first try to implement the SKI evaluator, is is to verbose comapred to 'ff'
05:37	srdjan	raek: I see that the '(reduce conj ...)' stinks a little, but I don't know how to improve it.
05:38	raek	I have barely used zippers at all, so I can't give you any feedback on the st function except for that it look small, which is a good thing, I think
05:39	raek	srdjan: also (dorun (for ...)) = (doseq ...)
05:39	srdjan	raek: i have never used zippers in my life, until now, and I don't find the difficult to use (actually, it is the only way that I know to transform the trees in functional language)
05:39	raek	(dorun (for [x tests] (do ...))) = (doseq [x tests] ...)
05:41	raek	you could perhaps make a recusively defined transformation function
05:42	srdjan	raek: you mean to combine 'ff' and 'ap'?
05:43	srdjan	'st' kind of stinks too. I would say that this look like a common pattern in tree transformation.
05:43	raek	no I was thinking about the zippers
05:44	raek	what does st do?
05:44	srdjan	I mean. The common thing is to apply the rule on the subtrees until we come to the end of the zipper, and then return the zip/root. I would expect that there is some function in clojure.zip that does exaclty this, but I could find it. :(
05:45	srdjan	raek: 'st' stans for 'ski-transform-tree'
05:45	srdjan	it walkd the tree in and applies the SKI rules on the subtrees. This gives you the minimal SKI tree.
05:46	srdjan	Example: (st '[S [I I]]) returns [S I], buy applying recuding [I I] to I
05:48	raek	are you looking for something like (st '[S [I I]]) --> (ap ['S (ap [(ap 'I) (ap 'I)])]) ?
05:48	srdjan	raek: 'st' takes the location in zipper, takes the node, applies the rules, replaces the node on the current location, and goes to the next node
05:49	srdjan	raek: 'st' works for me, but the question is if other people can understand the code. (I know that I didn't put the commetns, and that the function names are cryptic)
05:49	srdjan	raek: 'st' goes to S node, applies 'ap'
05:49	srdjan	goes to [I I] node, applies 'ap'
05:50	srdjan	and in the end goes to [S I] and applies 'ap' (the I in this expression is actually the recudtion of [I I] from the previous step)
05:51	raek	srdjan: srdjan can you describe how the resulting tree should be constructed for the '[S [I I]] case using symbols, vector literals and calls to ap?
05:51	srdjan	what is the common way in clojure to represent product types like in ocaml?
05:52	raek	srdjan: a map or a vector, depending on whether you want to name the slots or not
05:52	srdjan	the expression '[S [I I]] is a tree
05:53	srdjan	you can think of '[S [I I]] as a '(S (I I))
05:53	raek	yeah, that is clear
05:54	srdjan	raek: http://people.cs.uchicago.edu/~odonnell/Teacher/Lectures/Formal_Organization_of_Knowledge/Examples/combinator_calculus/
05:54	srdjan	on this link you can see how the tree is represented
05:54	srdjan	I just use clojure vectors to represent the trees
05:54	srdjan	probably I could use clojure list, and do the pattern patchin on them
05:54	srdjan	the end resut doesn't change
05:55	raek	maybe it's simpler to transform the expression in to full parenthesis form first
05:55	raek	and then apply the rules
05:55	srdjan	raek: please, can you elaborate on that?
05:56	raek	now you have examples like '[S x y z w], which really means '[[[[S x] y] z] w]
05:58	srdjan	raek: yes
05:59	srdjan	raek: the representation that I use is easier to use in pattern matching
05:59	srdjan	the end result is the same
06:51	ZabaQ	commas are whitespace!
06:52	ZabaQ	can't decide if thats genius or madness.
07:01	thorwil	depends on how you feel about using ~ to unquote, instead of the comma
07:02	thorwil	can it be that apply doesn't like macros?
07:24	ordnungswidrig	thorwil: you can't apply macros.
07:24	thorwil	ok, wiser about that now after going through http://osdir.com/ml/clojure/2010-01/msg01242.html
07:24	ordnungswidrig	g
07:44	pyninja	Has anyone else had problems with clj-http? Any time I try to do a POST request that works perfectly with curl, it never works. (dakrone are you here?)
07:50	napping	I haven't tried it yet, but I was planning to use it soon
07:51	pyninja	ok cool, looks like they fixed whatever was wrong. i updated it to 0.2.1 and it works now!
07:53	kij
08:33	napping	why is there a not-any? but not an any?
08:34	clgv	napping: english language? there is a 'some though
08:34	cark	,(some even? [1 2 3])
08:34	clojurebot	true
08:35	clgv	&(some identity [nil 2 3])
08:35	lazybot	⇒ 2
08:36	cark	,(some #{1 2 3 4} [1 6])
08:36	clojurebot	1
08:36	cark	=)
08:39	napping	I see it's the only one of the four which makes sense without a ?, but the name still seems a bit funny
08:57	srdjan	,(clojure-version)
08:57	clojurebot	"1.3.0-master-SNAPSHOT"
09:23	mattmitchell	I'm using Ring... and I'm curious to know if there is an elegant way to benchmark individual middleware layers?
09:31	napping	to set up a benchmark for them, or to instrument your server?
09:53	mattmitchell	napping: To benchmark all of them
10:31	darq	Hello. What's the best way to combine 2 vectors? Actually I want [elem] + elem + [elem] . Is there a nice way to add them together to a vector?
10:32	clgv	darq: how about ##(apply conj [1 2 3] 4 [5 6 7 8])
10:32	lazybot	⇒ [1 2 3 4 5 6 7 8]
10:34	darq	Thx clgv :)
10:42	TimMc	clgv: Whoa!
10:42	TimMc	,(doc conj)
10:42	clojurebot	"([coll x] [coll x & xs]); conj[oin]. Returns a new collection with the xs 'added'. (conj nil item) returns (item). The 'addition' may happen at different 'places' depending on the concrete type."
10:43	clgv	,(doc apply)
10:43	clojurebot	"([f args] [f x args] [f x y args] [f x y z args] [f a b c d ...]); Applies fn f to the argument list formed by prepending intervening arguments to args."
10:43	TimMc	No unary conj. That precludes (apply conj base more) when more might be 0-length.
10:44	ilyak	,(conj [1 2 3])
10:44	clojurebot	#<ArityException clojure.lang.ArityException: Wrong number of args (1) passed to: core$conj>
10:45	wink	What's the easiest way to persist some data? jdbc+h2db? or is there anything simpler? Basically just a cache for some maps identified by a primary key
10:46	TimMc	THere are a lot of core functions missing degenerate cases.
10:46	clgv	wink: I used java serialization in some cases
10:46	TimMc	,(doc pr)
10:46	clojurebot	"([] [x] [x & more]); Prints the object(s) to the output stream that is the current value of out. Prints the object(s), separated by spaces if there is more than one. By default, pr and prn print in a way that objects can be read by the reader"
10:46	duck1123	wink: redis is pretty easy
10:46	ilyak	TimMc: Why not (concat base more)?
10:46	ilyak	I agree with you, tho
10:47	wink	duck1123: I'm not inclined to even try to get redis running on windows, sorry :P
10:47	TimMc	ilyak: concat behaves differently than conj -- it uses cons
10:47	clgv	ilyak: concat does return a lazy seq and no vector
10:47	ilyak	Clojure could have less fns but more orthogonal ones
10:47	TimMc	,(concat [] '(1 2 3 4 5))
10:47	clojurebot	(1 2 3 4 5)
10:47	TimMc	,(concat () '(1 2 3 4 5))
10:47	clojurebot	(1 2 3 4 5)
10:48	TimMc	,(apply conj [] '(1 2 3 4 5))
10:48	clojurebot	[1 2 3 4 5]
10:48	TimMc	,(apply conj () '(1 2 3 4 5))
10:48	clojurebot	(5 4 3 2 1)
10:48	TimMc	If you know that base is a vector, it's fine.
10:50	ilyak	TimMc: (conj base (apply vector more))
10:50	ilyak	Oops, no
10:51	ilyak	I don't really understand the performance profile of those operations, so I can't figure out the best way
10:52	ilyak	I guess I rarely need vectors
10:53	clgv	ilyak: (conj base (apply vector more)) should have 2*N runtime if more has N elements
10:54	clgv	ilyak: btw, there is vec which can be used as (vec more)
10:54	ilyak	clgv: It won't work because conj doesn't expect vector as second param
10:55	clgv	ilyak: hm what do you mean?
10:55	ilyak	I'm yet to find a vector concatenation
10:55	jbwiv	can someone tell me how to change the working directory of the repl from within the repl?
10:55	ilyak	something like haskell's ++
10:55	ilyak	clgv: conj is (conj vec x & xs)
10:56	raek	ilyak: into
10:56	raek	,(into [1 2 3] [4 5 6])
10:56	clojurebot	[1 2 3 4 5 ...]
10:56	ilyak	Perhaps
10:56	ilyak	TimMc:
10:57	raek	but this does not (and can not) use the fact that the second argument is a vector
10:58	raek	it simply treats it as a sequence and constructs a new vector with those elements added
10:58	raek	the new vector will share structure with the first element, but not the second
10:59	raek	if you want something that can share structure with both arguments, you need another data structure (finger trees can do this)
11:00	khaliG	raek, what's the story with finger trees, are they part of core?
11:00	raek	they are in the "new contrib" I think
11:01	ilyak	raek: Neither does (apply conj)
11:01	khaliG	i'm interested in why they /cannot/ share structure with both arguments
11:01	raek	darq: this is how I would do it (into (conj left center) right)
11:03	raek	khaliG: have you read about how persistent vectors organize their data internally?
11:03	khaliG	raek, nope not yet
11:03	TimMc	raek: Why is that better than (conj left center right)?
11:03	raek	(conj left center right) does not do the same thing
11:04	raek	,(conj [1 2 3] 4 [5 6 7])
11:04	clojurebot	[1 2 3 4 [5 6 7]]
11:04	TimMc	err, throw an apply in there
11:06	jbwiv	is there a clojure debugger with which you can set breakpoints and inspect variable values?
11:06	raek	TimMc: hm, the fact that you cannot call conj with only one argument is not a problem there. I guess they are equivalent then.
11:06	zerokarmaleft	khaliG: http://blog.higher-order.net/2009/02/01/understanding-clojures-persistentvector-implementation/
11:08	TimMc	raek: Fancy. I guess this means that subvec doesn't need the whole tree, just some relevant subtrees. Allows more GC.
11:13	zerokarmaleft	jbwiv: https://github.com/GeorgeJahad/debug-repl
11:14	clgv	zerokarmaleft: that thing is pretty awesome. I use it about 50times a week
11:16	zerokarmaleft	i just discovered it a few days ago...and just read hugod' post about integrating into swank...now i need to look into why i can't examine locals in stack traces
11:16	jbwiv	zerokarmaleft, cool...thanks
11:17	duck1123	I've never managed to get it working very well
11:28	mattmitchell	I'm attempting to use the "time" function in my app. I'm finding it difficult to know which "time" call I'm seeing in stdout. Is there a way to wrap time so that it displays a label along with the time taken?
11:31	TimMc	You could print to out or whatever just before time is called.
11:32	TimMc	Or do somethign tricky so that time prints to a capturing writer.
11:33	Raynes_	&(println "Labeled:" (with-out-str (time (range 3))))
11:33	lazybot	⇒ Labeled: "Elapsed time: 0.513915 msecs" nil
11:33	Raynes_	Hardly tricky.
11:34	TimMc	Raynes_: It would have to be done to all calls to time.
11:34	TimMc	Also, any other output would be captured.
11:34	mattmitchell	Raynes_: The only problem with that the code that's being timed doesn't return the value to the caller
11:34	dnolen	given the existence of JSCocoa, programming iOS devices with ClojureScript is trivial.
11:34	mattmitchell	err sorry, in other words.. I need still need the return value
11:34	TimMc	mattmitchell: Use a let.
11:35	TimMc	Oh, wait... I see.
11:35	Raynes	Yeah, he can't really get the result while wrapping it in with-out-str.
11:36	mattmitchell	exactly
11:36	Raynes	He could rebind out to a writer himself though.
11:36	llasram	&(source time)
11:36	TimMc	Right.
11:36	llasram	&(source time)
11:36	lazybot	java.lang.Exception: Unable to resolve symbol: source in this context
11:36	llasram	Haha
11:36	llasram	Anyway, it's about 4 lines of code
11:36	llasram	Could just re-write it
11:36	TimMc	http://clojuredocs.org/clojure_core/clojure.core/time has source
11:36	duck1123	you could assign it to an atom or some such, but that would affect timings
11:37	TimMc	You also have to account for recursive calls.
11:37	TimMc	If A calls B and both use time, you only want to capture the time output from each level separately.
11:38	mattmitchell	right
11:38	duck1123	right, this atom would be local to the time2 macro
11:38	TimMc	mattmitchell: (let [ret (time ...)] (println "Foo done.") ret)
11:38	TimMc	Macroize that.
11:39	duck1123	,(time (+ 1 1))
11:39	clojurebot	"Elapsed time: 0.086 msecs"
11:39	clojurebot	2
11:40	TimMc	Printing a label to out before calling time is trouble, since an inferior call may also print to time.
11:40	TimMc	Printing afterwards is fine.
11:40	mattmitchell	TimMc: yes that's right. I'll try your let code out.
11:41	TimMc	mattmitchell: Or, as llasram pointed out... just rewrite time. :-)
11:42	duck1123	just not any of the fixed points
11:42	TimMc	fixed points?
11:43	TimMc	Is this some Haskelly nonsense?
11:43	duck1123	Doctor Who reference
11:43	TimMc	ah
11:44	TimMc	got it
11:44	llasram	Ha, the phrase "rewrite time" made me think the same sort of thing, but not a Dr. Who fan, so missed the ref
11:44	duck1123	sorry, you're talking about re-writing time, I couldn't resist
11:45	TimMc	mattmitchell: Write with-time, taking a function that accepts the time string and does something with it.
11:45	TimMc	(with-time ... #(println "Foo: " %))
11:46	TimMc	or (with-time ... #(str "Foo: " %)) with the expectation that the fn's return is printed.
11:51	duck1123	it would be cool if with-time accepted a callback fn that receives info about the run in a future
11:51	clgv	incanter question: somehow $join with three keys does not work like expected. e.g. [[:a :b :c] [:a :b :c]] - it just ignores :c somehow
11:53	mattmitchell	TimMc: You mean the string would act as a template, and with-time would fill it in with "format" or something?
11:58	TimMc	yeah
11:58	TimMc	err
11:58	TimMc	Not really. Depends how you write it.
11:58	TimMc	You could pass a format string, a prefix, a string-building function, or a function that can even print all by itself
11:59	TimMc	Write the simplest thing for now (accept a label) and expand it later as needed.
11:59	robermann	just resolved Gus' Quinundrum problem (http://www.4clojure.com/problem/125) after 4 days of tries - now, after seeing amalloy's solution, I consider myself totally dumb! :)
12:00	robermann	amcnamara's solution uses a fn*
12:00	robermann	what is it?
12:00	TimMc	Compiler primitive. It is what fn uses.
12:00	TimMc	,(macroexpand-all `(fn [x y] (+ x y))
12:00	clojurebot	#<ExecutionException java.util.concurrent.ExecutionException: java.lang.RuntimeException: EOF while reading>
12:01	TimMc	ugh
12:01	TimMc	&(macroexpand-all `(fn [x y] (+ x y)))
12:01	lazybot	java.lang.Exception: Unable to resolve symbol: macroexpand-all in this context
12:01	TimMc	&(macroexpand-1 `(fn [x y] (+ x y)))
12:01	lazybot	⇒ (fn* ([clojure.core/x clojure.core/y] (clojure.core/+ clojure.core/x clojure.core/y)))
12:03	robermann	thank you
12:18	duck1123	Here's a version of with-time that I came up with https://github.com/duck1123/ciste/blob/master/src/ciste/debug.clj#L13
12:18	duck1123	it takes a function to receive the timing data
12:19	mattmitchell	duck1123: awesome, I will need to study that for a bit :)
12:20	TimMc	mattmitchell: most of it is the same as time
12:20	mattmitchell	ok
12:20	TimMc	duck1123: Why a separate thread?
12:21	duck1123	to not interfere with the rest of the flow. I guess that might not be needed
12:21	TimMc	Could be confusing if the timing statements print out of order.
12:22	duck1123	yeah, I think you might be right
12:22	duck1123	the user can always start a new thread if they want it
12:25	glob157	any good "full" application examples, other than ants.clj out there for learning how to build "real" clojure apps ?
12:26	glob157	im still getting lost in a maze of maps with hidden key names etc... i think my code organization is missing something
12:27	mattmitchell	duck1123: This is what I came up with, basically TimMc 's idea of using let https://gist.github.com/1262089
12:28	mattmitchell	not real elegant, but working for now. I'll try your implementation out next.
12:29	TimMc	I'm not sure why time uses prn instead of println.
12:30	duck1123	mattmitchell: won't that print twice?
12:30	mattmitchell	duck1123: no, it's only printing once
12:31	mattmitchell	duck1123: oh, well yes sorry.
12:32	TimMc	glob157: Here's an application I wrote: https://github.com/timmc/CS4300-HW6 -- can't speak for "good", but it is a decent application
12:33	TimMc	glob157: Sample output: https://github.com/timmc/CS4300-hw6/blob/master/dev/sierpinski%20trace%20pastels%20ambient.png
12:33	duck1123	this is my largest app. https://github.com/duck1123/jiksnu I have a few other libraries on my page
12:33	dnolen	glob157: also Ring source is pretty clean well organized.
12:34	TimMc	dnolen: It can be hard to understand library sources, because you also have to understand the hypothetical clients that use them.
12:34	duck1123	that (should be) what tests are for
12:34	TimMc	mmm, true
12:34	glob157	the ray trace is great , but a little off topic compared with most apps. im thinking of a typical client side app with some application logic, maybe some MVC, etc...
12:35	dnolen	glob157: you might want to check out clooj, people actually use that, it's a lightweight Clojure IDE
12:36	glob157	oh ok
12:36	TimMc	glob157: Like something with widgets in a UI, or a web app?
12:37	TimMc	Widgets and state: https://github.com/timmc/CS4300-HW3/blob/master/doc/screencap-main.png -- another project of mine, involves dragging and stuff.
12:37	glob157	not worried about the particulars - but I wanted to see an app that had good separation of concerns.... hopefully one that explained the design philosophy as well. most of the examples I find are project euler.
12:37	TimMc	Ah, like a proper sample app.
12:38	TimMc	Almost like a tutorial.
12:38	glob157	Yeah, like JPetStore. Anyone interested in collaborating on one ? I have a github project called RudolF that was built to exemplify functional programming for new people like me.
12:39	glob157	ive built a lot of example apps using heroku, wrapping molecular vialization tools, etc, but i never really got a good feel for how complex data and state is handled in clj
12:41	TimMc	I think I did a pretty good job in the HW3 project. State is managed in refs, and the user data state is fully undoable/redoable.
12:41	duck1123	https://github.com/search?language=Clojure&q=tutorial
12:43	TimMc	I should actually split out cascade.clj to a separate utility.
12:44	glob157	which is hw3
12:44	TimMc	https://github.com/timmc/CS4300-HW3/tree/master/src/timmcHW3
12:45	TimMc	core.clj declares some refs at the top (I use earmuffs to mark them, even though that is usually used for rebindable vars) and the data structures are defined in state.clj
12:46	TimMc	cascade.clj is a make-like utility for managing state recomputations
12:46	glob157	oh ok. tim thanks. tim if interested in collab on some open source clojure bioinformatics stuff let me know !
12:46	TimMc	That might be cool.
12:46	TimMc	Do you have a github project (or similar) for it yet?
12:47	glob157	check out https://github.com/jayunit100/RudolF/bioclojure/
12:47	glob157	its all prototype code, but works. we want to bring it into a non prototype stage
12:48	TimMc	bad URL?
12:49	TimMc	-RudolF
12:51	duck1123	,(= 1/2 0.5)
12:51	clojurebot	false
12:51	duck1123	that's more unfortunate than the 2 vs 2.0 IMO
12:51	glob157	https://github.com/jayunit100/RudolF/tree/master/bioclojure
12:57	semperos	I need to call a static method on a Java class; I'm given the name of that static method as a string/keyword; what's the best way to dynamically call that method? any Clojure-specific way?
12:57	jkkramer	,(== 1/2 0.5)
12:57	clojurebot	true
12:59	mattmitchell	What's the preferred json output library for clojure?
12:59	technomancy	~cheshire
12:59	clojurebot	Huh?
12:59	technomancy	...
12:59	dakrone	~json
12:59	clojurebot	http://dakrone.github.com/cheshire/
12:59	joegallo	mattmitchell: https://github.com/dakrone/cheshire
12:59	mattmitchell	cool thanks!
13:00	joegallo	semperos: you'll probably just want to use reflection for that
13:00	technomancy	how do you do redirects/aliases with clojurebot?
13:00	dakrone	clojurebot: cheshire is <reply>https://github.com/dakrone/cheshire
13:00	clojurebot	Ik begrijp
13:00	dakrone	~cheshire
13:00	clojurebot	https://github.com/dakrone/cheshire
13:00	dakrone	like that?
13:01	dakrone	not sure what you meant
13:01	joegallo	dakrone: now define ~alias to output the example :)
13:01	technomancy	dakrone: I was thinking so it would refer to the same entry in memory, but I guess that's not as important in clojurebot since terms only show one entry at a time
13:02	dakrone	ahh
13:02	semperos	joegallo: yeah, I just got .invoke working, wasn't sure if there was a preferred/better abstraction at the Clojure level
13:02	TimMc	semperos: Maybe RT could help.
13:03	joegallo	http://clojure.github.com/clojure-contrib/#reflect
13:03	joegallo	there's that
13:03	zerokarmaleft	dakrone: fyi, the stream encoding example in the readme has args out of order
13:04	semperos	TimMc, joegallo thanks
13:04	semperos	joegallo: that looks about like what I'm doing, so good to know I was on the right track
13:04	dakrone	zerokarmaleft: you are right! lemme update that
13:04	semperos	using 1.3, so not going to depend on a 1.2 contrib fn, but good to see the implementation
13:04	dakrone	zerokarmaleft: thanks!
13:11	devth	hi. trying to upgrade to 1.3 but i'm getting "Warning: classpath not declared dynamic .." and "ClassNotFoundException clojure.lang.ILookupHost java.net.URLClassLoader" when i try to start my repl. what would be a good way to debug? i tried removing deps in my project.clj but makes no difference.
13:12	technomancy	devth: try lein upgrade
13:13	devth	technomancy: upgrading...
13:15	devth	technomancy: upgraded to 1.6.1 but seeing the same issue. deps: https://gist.github.com/1262209
13:15	technomancy	devth: clojure-json is deprecated. but it's more likely lein-nailgun.
13:18	devth	technomancy: hm, still see it after removing both of those
13:18	technomancy	devth: can't repro here; are you sure you're using 1.6?
13:18	devth	technomancy: i'll reinstall. originally installed with homebrew
13:22	tremolo	in Leiningen, what's the preferred way to set up a dependency to another local lein project? is this possible?
13:22	tremolo	i couldn't find any documentation on this
13:23	technomancy	tremolo: add it to :dependencies and then set up the checkouts/ directory. look for "checkout dependencies" in "lein help readme"
13:23	tremolo	technomancy: awesome, thank you
13:24	technomancy	no problem
13:26	devth	technomancy: classpath error is gone -- now just seeing "ClassNotFoundException clojure.lang.ILookupHost java.net.URLClassLoader$1.run (URLClassLoader.java:202)"
13:26	technomancy	huh; never seen that before
13:26	devth	hrm. strange. i will keep messing with it. thanks!
13:48	faust451	hi guys
13:48	devn	How does Haskell's iteratee relate to clojure's chunked sequences?
13:48	devn	Is there any relationship there?
13:48	faust451	i try using Lucene with Clojure, but fail. please help me
13:50	joegallo	faust451: that's a pretty open-ended question. maybe a little more detail?
13:50	faust451	this code http://friendpaste.com/56X61Tkfi6u1iwDEOvVD7A fail with java.lang.ClassNotFoundException: RAMDirectory.
13:50	dakrone	faust451: recommend this: https://github.com/weavejester/clucy
13:51	faust451	dakrone: thanks
13:51	joegallo	faust451: are you using leiningen? is lucene on your classpath? it seems like you are missing the jars you need
13:51	faust451	joegallo: yes leiningen
13:52	joegallo	and if you `ls lib`, do you see the lucene jars there?
13:52	joegallo	because it seems like you might be missing some of them
13:52	TimMc	lein classpath
13:55	faust451	joegallo: ls lib/ lucene-core-3.0.3.jar, lucene-queries-3.0.3.jar
13:56	faust451	joegallo: http://friendpaste.com/2pKJXrnwhWmVGcDieu2ogP
13:57	TimMc	faust451: That's all that's in lib? Try running lein deps
13:59	faust451	TimMc: it's for ls -l lib/ \| grep lucene
13:59	TimMc	k
13:59	technomancy	faust451: I recommend using clucy
13:59	faust451	TimMc: in project file i require only lucene-core and lucene-queries
14:02	joegallo	def index (RAMDirectory.))
14:03	joegallo	you are missing parens around the RAMDirectory. constructor call
14:03	faust451	technomancy: but what i am doing wrong?
14:03	joegallo	you are doing your syntax wrong
14:03	joegallo	your dependencies are fine, but because of the way you have your code typed in, clojure is looking for the non-existent class "RAMDirectory." rather than "RAMDirectory", which you have.
14:05	faust451	joegallo: thanks, let me try to fix
14:33	zodiak	is it jst me that is sitting watching the whole node.js community become rabid dogs and thinking, I don't actually want to convert any of them to clojure :)
14:33	Raynes	It's just you.
14:33	zodiak	awesome :D
14:34	Raynes	;)
14:34	zodiak	all the peeps at work here are trying to jump onto the node.js bandwagon, I keep trying to push clojure+ring
14:35	zodiak	I don't know how to 'soft sell' clojure+ring more though.
14:36	zodiak	there is definitely a visceral reaction when it comes up that clojure is a child (somewhat ;) of lisp
14:38	cemerick	zodiak: just bring up that node.js is a child of javascript; on the merits, that should carry the day.
14:38	amalloy	hahahaha
14:39	amalloy	cemerick: js is a disowned mutant child of scheme though, right?
14:39	technomancy	I've never understood that argument
14:39	technomancy
14:39	technomancy	is it just shorthand for "it has closures"?
14:39	cemerick	technomancy: which one?
14:39	technomancy	cemerick: "JS is secretly just scheme with bad syntax"
14:40	technomancy	feels like a non-sequitur
14:40	amalloy	technomancy: allegedly the guy who wrote js was hired with the promise that he could write "scheme for the browser"
14:40	amalloy	and then was told "oh btw it has to look like java, marketing guys love that shit"
14:40	cemerick	I think it came from Eich having referred to scheme while working on the implementation.
14:40	zodiak	cemerick, sadly, I think that because it's a "language we know" that node.js is "the way to go" .. which is laughable when I bring up js has prototypal inheritance (which is not good imo) and everyone else goes 'huh' in the room
14:41	TimMc	>_<
14:45	PPPaul	does anyone know a good way to flatten a tree into a 1d vector (preserving the paths) :eg [[root child child][root child child]]
14:46	cemerick	zodiak: doesn't sound like people know the language all that much.
14:47	PPPaul	js is a child of lisp too
14:47	PPPaul	my js looks a lot like my clojure code
14:47	Raynes	Probably more parentheses in the JS code.
14:47	PPPaul	maybe
14:52	ibdknox	,(flatten [[:hey :how [:are :you]] [:doing :today]])
14:52	clojurebot	(:hey :how :are :you :doing ...)
14:52	ibdknox	PPPaul: ^
14:53	PPPaul	but, no
14:53	amalloy	PPPaul: your question is pretty vague, it seems to me
14:53	PPPaul	[hey how are... [hey how you...
14:53	PPPaul	i want to be able to re make the tree
14:54	TimMc	PPPaul: Provide sample input and output.
14:54	bsod1	is there a function in clojure like (some), but returns value of first true item instead of true or nil?
14:54	PPPaul	[:a [:b :c]] [[:a :b][:a :c]]
14:54	TimMc	bsod1: That's what some does.
14:55	bsod1	TimMc: no, some returns true or nil, I want the value of item instead of true
14:55	ibdknox	,(some (partial > 3) [1 2 1 1 5 4 3 2])
14:55	clojurebot	true
14:55	TimMc	,(some identity [nil false 5 'bsod1])
14:55	clojurebot	5
14:55	ibdknox	it depends on the function TimMc
14:55	bsod1	,(some #(= % 12) [1 3 12])
14:55	clojurebot	true
14:56	bsod1	oh, wait
14:56	TimMc	Ah, you want to run a predicate too.
14:56	bsod1	TimMc: yes
14:56	symbole	Anybody here at JavaOne?
14:56	bsod1	,(some #(when (= % 12) %) [1 3 12])
14:56	clojurebot	12
14:57	bsod1	any better ways to do this?
14:57	jkkramer	,(some #{12} [1 3 12])
14:57	clojurebot	12
14:57	TimMc	bsod1: That won't work for a predicate that accepts nil or false.
14:58	bsod1	TimMc: I know, I can't find any better solutions
14:58	bsod1	jkkramer: great
14:58	bsod1	thanks
14:59	TimMc	,(some #{nil} [1 nil 5]) :-P
14:59	clojurebot	nil
14:59	amalloy	bsod1: (comp first filter)
15:00	TimMc	nice
15:01	amalloy	personally i wish that's how some behaved, but it's not a huge deal
15:12	apgwoz	has anyone ever use Gson/fromJson with clojure?
15:14	mister_roboto	Is clojure-clr at clojure 1.2 now? 1.3? Can't seem to find any specifics on the github page
15:14	Bronsa	1.3
15:14	mister_roboto	Cool thx
15:19	TimMc	Oh hey, is that still in development?
15:33	hugod	I wonder if it makes sense for clojure.match to allow matching on arbitrary function literals [#(some-predicate? %)]
15:36	amalloy	hugod: isn't that the "predicate dispatch" dnolen has been saying he plans to add eventually?
15:37	amalloy	disclaimer: i know little about either match or predicate-dispatch
15:41	hugod	amalloy: I don't think so - I think that is a dispatch mechanism for multimethods, etc
15:43	hiredman	dakrone: have you ever seen errors about a missing class SyncBasicHttpParams when trying to use clj-http?
15:45	hiredman	ah, ring and clj-http depend on different versions of httpcomponents
15:47	dnolen_	hugod: it also supports that via guards
15:47	dnolen_	s/also/already
15:47	lazybot	<dnolen_> hugod: it already supports that via guards
15:49	srid	what is the primitive for returning a count of true's in a list of true's and false's?
15:49	srid	or should I just use (count (filter ...?
15:49	amalloy	why would there be a primitive for that?
15:50	srid	just wondering if there was (to score in code golf :P )
15:50	amalloy	chuckle
15:50	amalloy	srid: problem 83?
15:50	srid	nope, 83
15:50	srid	nope, 120
15:52	TimMc	$findfn [true false true true false] 3
15:52	lazybot	[]
15:52	TimMc	:-)
15:52	amalloy	TimMc: o/
15:54	srid	is there a function to map f1 f2 .. fn on a coll? got it, (map (comp f1 f2) coll)!
15:56	gtrak	map comp looks to be the same as (map f1 (map f2 coll))
15:56	amalloy	it is the same
15:58	duck1123	isn't it closer to (map #(f1 (f2 %)) coll) ? (same effect though)
15:58	gtrak	but the comp is 1 more character
15:58	gtrak	yea, that's smaller
15:58	hugod	dnolen_: ah, thanks. should have realised that…
15:59	gtrak	actually, mine is fewer chars :-)
16:03	duck1123	wasn't there a contrib library that aliased comp and partial for point-free nuts?
16:04	amalloy	clojure.contrib.haskell?
16:04	clojurebot	clojure-stub is http://github.com/nakkaya/clojure-stub/tree/master
16:04	amalloy	~rimshot
16:04	clojurebot	Badum, ching
16:04	TimMc	ching, really?
16:04	TimMc	more of a tsh
16:04	srid	(filter #(true? %) ... <- can I obviate the true? here?
16:05	TimMc	identity, sort of
16:05	srid	right
16:05	TimMc	,(doc true?)
16:05	amalloy	TimMc: take it up with fsbot; i copied his rimshot to clojurebot
16:05	clojurebot	"([x]); Returns true if x is the value true, false otherwise."
16:05	srid	#(* % %) <- can't this be shortened
16:05	TimMc	srid: identity is going to be sloppier
16:05	srid	?
16:05	srid	it would only accept boolean, so identity is fine.
16:06	gtrak	,#t
16:06	clojurebot	#<ExecutionException java.util.concurrent.ExecutionException: java.lang.RuntimeException: java.lang.ClassNotFoundException: t>
16:07	kzar	Any ideas how I can re-write this blog-snippet enlive snippet? https://github.com/kzar/daveinadub/blob/master/src/daveinadub/views/blog.clj#L10 It looks awful ATM
16:42	technomancy	if you haven't been using M-x clojure-jack-in on account of how much elisp it spews out upon every connect, there's a new swank-clojure that fixes that: lein plugin install swank-clojure 1.3.3
16:46	hugod	is there a predicate to test if a number is boxed or not?
16:47	amalloy	hugod: i don't think that's possible
16:49	duck1123	isn't that a Heisenberg situation? Always true
16:49	hugod	not even in java?
16:49	brehaut	amalloy: really? surely a protocol would make it fairly trivial?
16:49	amalloy	since functions can only take objects, which are necessarily boxed. in 1.3 you can have a function that takes unboxed primitives, but afaik it can't say "i take a single arg which is either Object or long or double"
16:50	amalloy	brehaut: no way
16:50	napping	maybe something like warn-on-reflection could exist
16:50	brehaut	amalloy: but you can make a protocol know about arrays of bytes
16:50	hiredman	ugh
16:50	brehaut	i guess theres a backing class though isn't there
16:51	hiredman	anyone know if clj-http's :form-params actually work?
16:51	amalloy	hugod: you can't do it in java either, i think, but i'm less certain
16:52	amalloy	it's a weirder question in java, because java has static type information: "Did I declare this variable as an int or an Integer?" is not a useful question to ask
16:52	dakrone	hiredman: have you tried the latest clj-http version? someone sent a pull recently that was merged
16:52	hugod	amalloy: I thought it might be possible with overloading
16:53	amalloy	hugod: maybe. that's certainly the only approach that might work
16:53	brehaut	hugod: in java i think it would; but only because the type system already knows what that the local vars are primitives already
16:53	amalloy	but i don't know the details of how clojure passes numbers to java methods; it might notice that the thing takes a primitive long and unbox it
16:54	brehaut	hugod: i recall method overload selection being a compile time rather than runtime decision
16:54	hiredman	dakrone: I'm on the latest release
16:55	dakrone	cool, how's it not working for you?
16:55	amalloy	brehaut: correct. but since he's calling it from clojure, he'll be using reflection with runtime data
16:56	brehaut	amalloy: at which point, you are using the reflection data anyway, so why go to java?
16:56	hiredman	dakrone: from what I can tell (running it against ring, ring just printing out request map) it looks like it is ignoring :form-params
16:56	amalloy	good point
16:56	dakrone	hiredman: GET or POST?
16:56	hiredman	client/post
16:57	dakrone	can you send me a snippet?
16:58	terom	,(.isPrimitive (class 3))
16:58	clojurebot	false
16:58	terom	,(.isPrimitive (Integer/TYPE))
16:58	clojurebot	true
17:04	kzar	ibdknox: You there?
17:06	ibdknox	kzar: hey
17:07	kzar	ibdknox: Is there a way have multiple routes in a defpage? I wanted /blog to set :page to 0 and /blog/archive/:page to set :page to what's given
17:08	ibdknox	take a look at (render)
17:08	ibdknox	with a named route for blog-page, it would look like this:
17:09	ibdknox	(defpage "/blog" (render blog-page {:page 0}))
17:09	kzar	ibdknox: Ah cool I see, cheers
17:10	kzar	ibdknox: Oh wait, how do name blog-page? defpage doesn't take a name
17:11	ibdknox	it does
17:11	ibdknox	as of 1.2
17:11	kzar	aha gotya
17:11	ibdknox	(defpage blog-page "/blog/archive/:page" {:keys [page]} ...)
17:11	ibdknox	or rather
17:11	ibdknox	it can :)
17:14	ibdknox	kzar: you can do the same thing without naming it
17:14	ibdknox	(render "/blog/archive/:page" {:page 0})
17:17	kzar	ibdknox: cool OK, yea I got both ways working fine :) cheers
17:19	carllerche	I have a leiningen project that has a ragel -> java step, what would be the easiest way to have leiningen run a couple of ragel commands before javac?
17:34	semperos	I'm not strong on my Java jarring skills, getting following error while trying to uberjar
17:34	semperos	java.lang.SecurityException: no manifiest section for signature file entry org/bouncycastle/asn1/ocsp/ResponderID.class
17:34	semperos	I see that I can add a :uberjar-exclusions ["signature file here"] in my project.clj as one possible solution
17:35	semperos	but I don't entirely understand
17:38	semperos	nm, just needed to learn more about jar structure...
17:44	mwillhite	hey all - I'm confused. working with the jets3t lib and I can get a list of all the objects in my bucket, however when I try to apply a prefix (filter) I get an IllegalArgumentException
17:44	mwillhite	it may be a clojure syntax issue, just not sure
17:44	mwillhite	https://gist.github.com/1262913
17:44	mwillhite	any help will be greatly appreciated!
17:46	hiredman	mwillhite: check the javadocs, the method you are trying to call doesn't exist
17:47	mwillhite	I'm looking at the method in the docs…
17:47	mwillhite	and why would it work without the prefix arg?
17:47	wink	listObjects only takes one argument
17:47	wink	what is stammer? your bucket?
17:47	mwillhite	check the next method definition
17:47	mwillhite	yeah its the bucket
17:47	amalloy	mwillhite: listObjects doesn't seem to have a two-arg version
17:47	amalloy	one, or three
17:47	mwillhite	oh (I know nothing about java)
17:48	mwillhite	so I need to fill in the other arguments?
17:48	semperos	yep
17:48	mwillhite	thanks :)
17:48	semperos	that API offers several different method signatures
17:48	semperos	you have to support at least one, giving all arg's for that one
17:48	mwillhite	okay cool
17:48	mwillhite	good to know
17:49	wink	so (.listObjects s3-foo bucket prefix delimiter) basiclaly
17:49	mwillhite	what is that last thing "long maxListingLength"
17:49	mwillhite	what does that mean? its not an arg?
17:49	wink	that's the 4 arg version
17:49	mwillhite	okay
17:50	wink	oh how I love people who use @deprecated without stating what is the best way now
17:50	mwillhite	hehe thanks all, its working now!!
17:53	wink	hm, when "lein deps" fails to find a dependency and tells me to mvn install:install-file - what do I do if I seem to have no maven because I'm on windows and lein magically works anyway?
17:54	hiredman	wink: install maven
17:57	duck1123	or download the project in question and build it, but install maven anyway
17:58	technomancy	could be your :dependencies has a typo or something
17:58	wink	it worked when just copying the jar inside lib/ - but another lein deps kills it
17:58	wink	it's h2 - not in any repo I've seen
17:59	technomancy	probably need to file a bug with h2 then
18:00	technomancy	in the mean time that's something you still do need maven for
18:00	wink	oh? it sounded like total intended behavior, so I wasn't really bothered
18:00	wink	yeah, on it
18:00	hiredman	h2? the embedded sql? it is in repos
18:00	joegallo	http://mirrors.ibiblio.org/pub/mirrors/maven2/com/h2database/h2/
18:00	technomancy	well if their intention is to not publish an artifact then their intention is wrong
18:01	joegallo	for future reference:
18:01	joegallo	http://mvnrepository.com/search.html?query=h2
18:01	hiredman	~google maven h2
18:01	clojurebot	First, out of 22400 results is:
18:01	clojurebot	Maven Repository: com.h2database » h2
18:01	clojurebot	http://mvnrepository.com/artifact/com.h2database/h2
18:01	technomancy	joegallo: nah: http://search.maven.org/#search%7Cga%7C1%7Ch2
18:01	technomancy	crappier URL, but nicer otherwise
18:02	wink	oh well, if they suggest org.h2 instead of com.h2database - that could be it
18:03	wink	again, you're all really helpful and I'm the idiot. thanks a lot
18:03	joegallo	technomancy: nice, and thanks! (keyword added for search!)
18:04	wink	works :)
18:11	TimMc	OK, I just got a new laptop, and I am picking a hostname. I'm thinking about "thunk". My question: If I go to the Conj, is everyone else going to have the same damn hostname?
18:12	sjl	Anyone know if there's a reason I shouldn't be able to start the Noir server from inside of swank?
18:12	technomancy	TimMc: there's a who occasionally visits this channel who goes by thunk
18:13	wink	TimMc: Did you really thunk hard about that hostname? Was it even a ThunkPadded?
18:13	TimMc	wink: It's a ThinkPad, yes.
18:13	TimMc	Which is also built pretty sturdily, so it was a shoo-in name.
18:13	wink	a colleague of mine named it lolnovo
18:13	TimMc	haha
18:14	technomancy	niiiice
18:21	technomancy	cemerick: looking forward to trying out pomegranate
18:21	technomancy	I've had the idea for something like that for a while
18:21	cemerick	technomancy: I demo'd it live at J1 about an hour ago, so it seems to work :-)
18:22	technomancy	gutsy for a brand-new project =)
18:22	cemerick	I'm trying to go big these days.
18:22	cemerick	or, bigger :-P
18:22	technomancy	my plan was to have the users edit project.clj and have the running instance recalculate it, but of course that only works for additive deltas.
18:23	cemerick	well, pomegranate isn't any smarter
18:23	cemerick	It's a very long length of rope
18:23	technomancy	yeah, it's just more specific
18:23	technomancy	but then the responsibility lies with the user to keep the instance in sync with the deps declaration
18:24	cemerick	sure
18:25	cemerick	I think we're talking about the same thing in any case; the 'recalculate' part is in aether, insofar as existing dependencies are in .m2/repo
18:26	cemerick	i.e. I don't think it matters if the same jar URL is registered with a classloader twice
18:26	technomancy	right
18:27	cemerick	As I was writing it, I was thinking "jeez, leiningen already has some of these utilities" :-P
18:27	technomancy	I meant "recalculate from project.clj" rather than taking the additional deps as function args"
18:28	cemerick	Have you started with aether in lein yet, or is that still just dangling over your head?
18:28	technomancy	I have a stash for it, but it's mostly cribbed from a gist I got from hugod
18:28	ibdknox	cemerick: technomancy: it'd be nice to have a lib that just generally lets you interact with the classpath
18:28	ibdknox	for example, there are somethings in tools.namespace that I really wished worked in more contexts
18:29	ibdknox	some things*
18:29	cemerick	ibdknox: It's tricky. Once you're talking about the classpath as a mutable thing, you have to ask, "which classpath?"
18:30	hiredman	yeah, there is no classpath
18:30	hiredman	there are classloaders
18:31	ibdknox	sure, is there no way for us to create an abstraction that would make that an implementation detail?
18:31	cemerick	either no classpath, or 1 + N classpaths, yes.
18:31	ibdknox	these are the things that are currently available to us
18:31	cemerick	ibdknox: pomegranate is 99% of the way there, as long as you only care about clojure
18:32	cemerick	module systems and app servers suck for these sorts of things
18:33	ibdknox	cemerick: in theory does pomegranate work in a war?
18:33	cemerick	sure
18:33	cemerick	actually, it will currently attempt to find the "eldest" classloader, which may be above the war's classloader
18:34	cemerick	That may be a problem. :-P
18:34	ibdknox	haha :)
18:34	cemerick	Maybe I'll look at that in week 2.
18:34	ibdknox	why the caveat "only about Clojure", are things significantly different for loading random java stuff? (I know very little about classloaders and such)
18:36	cemerick	Random Java stuff has static names in it. You need something like JRebel for that.
18:36	cemerick	or osgi :-P
18:36	ibdknox	ah
18:37	cemerick	vars are the real hero of clojure's dynamicism.
19:15	sjl	ibdknox: Hey, you're the author of Noir, right?
19:15	ibdknox	sjl: yep
19:16	sjl	ibdknox: any idea why I'd be getting this when trying to use 'lein run' in a noir project? http://dpaste.com/627489/
19:16	ibdknox	those are netty errors
19:16	ibdknox	are you using aleph?
19:18	ibdknox	sjl: that's from something external to noir. It doesn't use contrib or netty :)
19:19	ibdknox	sjl: is your project on github? I can take a quick look to see if I can spot anything
19:19	shep-home	So, one of my functions goes off into an infinite loop when I test it. How do I jump in to that running thread with a debugger?
19:19	sjl	ibdknox: nevermind, I think I figured it out... for some reason Puppet doesn't want to install and start Redis until I provision it a second time
19:19	shep-home	I'm running it inside of emacs with clojure-test-mode
19:20	technomancy	shep-home: M-x slime-list-threads maybe
19:25	shep-home	technomancy: that seems to go off into it's own little hole
19:27	shep-home	M-x slime-interrupt doesn't
19:28	shep-home	(work that is)
19:30	technomancy	sorry don't know
19:31	shep-home	ANy idea how to get the JVM running so that I can attach with a java debugger?
19:32	shep-home	I'm currently using clojure-jack-in
19:33	technomancy	maybe the cdt documentation would say
19:35	shep-home	I'll dig into that, thnks
19:38	shep-home	Is there a way I can change what `lein jack-in` does?
19:39	TheBusby	only if you purchase a lein gold account
19:39	ibdknox	lol
19:42	technomancy	shep-home: you can add additional elisp files to be piggy-packed, though it's not well-documented
19:42	technomancy	is that what you mean?
19:51	shep-home	technomancy: move like trying to figure if I can add -X... options when it starts Java
19:51	technomancy	sure; that's orthogonal to jack-in
19:51	technomancy	:jvm-opts in project.clj
19:52	shep-home	thx
19:54	whidden	Hello all, is the function 'some efficient?
19:56	arohner	whidden: it's O(n)
19:59	whidden	arohner: hmm that's what I thought but when i profile my use of some it doesn't look like O(n), more like O(e^n)
20:00	arohner	whidden: what's in your predicate? I'm confident some by itself is O(n)
20:01	whidden	arohner: yeah all things point to my predicate, which is a partial that at its core does a range check.
20:01	amalloy	arohner: prime? :)
20:02	arohner	amalloy: factor-rsa :-)
20:02	whidden	Are there issues with using the clojure.contrib.profile collection of tools?
20:53	paul_	anyone here use clutch?
20:54	brehaut	paul_: just go ahead and ask your question; if someone can help they will answer
20:55	paul_	ok
20:55	paul_	i want to change the port for clutch
20:55	paul_	i think i need to use "with-bindings"
20:55	paul_	i don't know how, though
20:57	brehaut	paul_: its the slightly more awkward version of binding; (with-bindings {#'foo :bar} …) rather than (binding [foo :bar] …)
20:57	ibdknox	woah what is that nonsense?
20:58	brehaut	ibdknox: i didn't think i was that far off track
20:58	ibdknox	no no
20:58	ibdknox	why doesn't that use a vector?
21:00	brehaut	ibdknox: so that you can build up a map of bindings programattically
21:00	ibdknox	hm
21:02	ibdknox	brehaut: is this the clutch we're talking about? https://github.com/ashafa/clutch
21:03	brehaut	i think so
21:03	ibdknox	I looked through the source
21:03	ibdknox	with-bindings isn't there?
21:03	ibdknox	lol
21:03	brehaut	i think with-db probably uses bindings rather than with-bindings?
21:04	brehaut	would you look at that it does
21:04	brehaut	https://github.com/ashafa/clutch/blob/master/src/com/ashafa/clutch.clj#L129-135
21:04	ibdknox	hah
21:05	ibdknox	I thought with-bindings was something specific to clutch. I didn't realize it was a core thing
21:05	brehaut	oh right haha
21:05	ibdknox	I don't think he needs with-bindings.
21:06	brehaut	i dont either
21:06	brehaut	i think with-db will do it
21:06	ibdknox	paul_: ^
21:06	brehaut	paul_: the db map is i think {:host "hostname" :port int? :name "dbname"}
21:07	brehaut	paul_: see also set-clutch-defaults! https://github.com/ashafa/clutch/blob/master/src/com/ashafa/clutch.clj#L54-62
21:08	brehaut	paul_: way back when i last used clutch i found the source to be pretty readable in general. theres very little complicated stuff
21:12	cgray	is there a more idiomatic way of doing (comp not =) ?
21:12	brehaut	(complement =)
21:12	cgray	and is comp idiomatic in general?
21:13	ibdknox	,(doc not=)
21:13	clojurebot	"([x] [x y] [x y & more]); Same as (not (= obj1 obj2))"
21:13	ibdknox	cgray: yeah, comp is idiomatic
21:13	cgray	ibdknox: thanks
21:14	khaliG	wonder why they didn't just go with the java !=
21:15	ibdknox	! means side-effects in clojure
21:15	ibdknox	as a matter of convention
21:15	khaliG	as a prefix though?
21:15	brehaut	and not= is the lisp convention isn't it?
21:16	zippy314	Hi, what's the correct way to use regular expressions in clojurescript. i.e. what's the equivalent of (re-find #"^a" some-str)?
21:16	ibdknox	(/=) is lisp I think
21:16	ibdknox	err CL that is
21:16	khaliG	ibdknox, yeah /= in CL
21:17	TimMc	,(doc re-find)
21:17	clojurebot	"([m] [re s]); Returns the next regex match, if any, of string to pattern, using java.util.regex.Matcher.find(). Uses re-groups to return the groups."
21:17	TimMc	zippy314: I'm not a cljs user -- does it have regex literals yet?
21:17	ibdknox	zippy314: it's the same?
21:17	ibdknox	zippy314: https://github.com/clojure/clojurescript/blob/master/src/cljs/cljs/core.cljs#L2576
21:18	zippy314	no, it's not
21:18	ibdknox	ah
21:18	TimMc	zippy314: So what happens if you do ##(re-find #"^a" "abc") in cljs?
21:18	lazybot	⇒ "a"
21:18	ibdknox	(re-pattern)
21:20	zippy314	It's weird: https://github.com/clojure/clojurescript/wiki/Differences-from-Clojure says you have to use Javascript regexs (see Other Functions)
21:20	ibdknox	zippy314: I just did it in the repljs
21:21	zippy314	ibdknox: what exactly?
21:21	ibdknox	(re-find #"a" "hhha")
21:21	zippy314	hmmm..
21:22	ibdknox	also fine: (let [x "some string"] (re-find #"a" x))
21:22	zippy314	Is there version thing going on here?
21:23	zippy314	In my code that totally fails.
21:23	ibdknox	I have the latest
21:23	zippy314	I have (if (re-find #"/^multipart/" "multipart") (js/alert "test")) and I don't get the alert...
21:24	ibdknox	get rid of the /
21:24	zippy314	duh!
21:24	zippy314	It worked. Thanks!
21:34	paul_	for clutch i tried (db/set-clutch-port! "5985" (db/all-databases)) but it doesn't work
21:34	paul_	going to look at the src
21:36	devth	am i correct to assume it's not possible to (:require [com.foo.bar.*]) to require all namespaces starting with `com.foo.bar`?
21:37	paul_	seems to work with set-clutch-defaults!
21:39	paul_	ok, got it working... i'm now trying to query a view, but i'm having trouble doing so too... going to look at the source :D
21:46	gfredericks	devth: darn tootin you're correct!
21:50	amalloy	ask lazybot, he knows everything. lazybot, is it possible to (:require [com.foo.bar.*])??
21:50	lazybot	amalloy: Uh, no. Why would you even ask?
21:51	gfredericks	lazybot, is it possible to (:require [clojure.set])??
21:51	lazybot	gfredericks: What are you, crazy? Of course not!
21:51	amalloy	gfredericks: you just have to not ask him any dumb questions. it's all your fault for phrasing it wrong
21:52	gfredericks	lazybot: is it possible for me to ask you a dumb question??
21:52	lazybot	gfredericks: Uh, no. Why would you even ask?
21:54	gfredericks	amalloy: given that he just contradicted you, I have to conclude that you're right.
21:59	ibdknox	lazybot: is gfredericks trying to game the system???
21:59	lazybot	ibdknox: Oh, absolutely.
22:00	ibdknox	lazybot: is that acceptable??
22:00	lazybot	ibdknox: Definitely not.
22:00	brehaut	lazybot: is this a question?????
22:00	ibdknox	lol
22:00	amalloy	brehaut: he probably knows better than to respond to people with five question marks. answering such people only leads to sadness
22:00	ibdknox	> 4 ?'s and he should tell you to go to hell ;)
22:00	hsbot	<hint>:1:6: lexical error in string/character literal at character ' '
22:00	ibdknox	oh no's
22:00	ibdknox	lol
22:00	brehaut	we have a haskell bot?!
22:01	amalloy	brehaut: we have two
22:01	brehaut	fantastic :)
22:01	amalloy	i'm not sure who owns hsbot, but lazybot does haskell too
22:01	brehaut	> print "hello, world!"
22:01	hsbot	No instance for (GHC.Show.Show (GHC.Types.IO ())) arising from a use of `M5580730128600360166.show_M5580730128600360166' Possible fix: add an instance declaration for (GHC.Show.Show (GHC.Types.IO ()))
22:01	amalloy	$heval [1,3...15]
22:01	brehaut	baha fail
22:01	lazybot	⟹ Not in scope: `...'
22:01	amalloy	$heval [1,3..15]
22:01	lazybot	⟹ [1,3,5,7,9,11,13,15]
22:02	ibdknox	$heval print "hey"
22:02	lazybot	⟹ No instance for (GHC.Show.Show (GHC.Types.IO ())) arising from a use of `M1183673886.show_M1183673886' at <interactive>:(2,0)-(4,30)Possible fix: add an instance declaration for (GHC.Show.Show (GHC.Types.IO ()))
22:02	brehaut	clearly they are sanely not operating in the IO monad
22:02	brehaut	its a pretty good sandbox
22:02	amalloy	hah
22:02	brehaut	> show 1
22:02	hsbot	"1"
22:02	amalloy	> [1, 2..]
22:02	hsbot	[1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,...
22:03	brehaut	> 1 + read "1"
22:03	hsbot	2
22:04	brehaut	> fib where fib = 1:1: zipWith (+) fib $ tail fib
22:04	hsbot	<hint>:1:5: parse error on input `where'
22:04	brehaut	> let fib = 1 : 1 : zipWith (+) fib $ tail fib in fib
22:04	hsbot	Couldn't match expected type `[a0]' with actual type `[b0] -> [c0]'
22:04	brehaut	and this is why i use clojure, not haskell, for my day to day programming
22:04	ibdknox	what I'm learning here is that haskell sucks
22:05	brehaut	ibdknox: i think if i threw some braces at the first one it would be ok
22:05	ibdknox	haha I was just trying to see who would yell at me ;)
22:05	brehaut	or at least present the same type con flit as the second
22:05	ibdknox	I know basically nothing about haskell
22:06	gfredericks	I look away for two seconds and suddenly everybody knows haskell.
22:06	ibdknox	gfredericks: welcome to #haskell, hope you brought your helmet.
22:06	amalloy	i love that it thinks :1:5: parse error on input `where' is a good hint
22:07	ibdknox	lol
22:09	brehaut	ibdknox: i only know enough haskell to get the really cryptic compiler warnings
22:09	mjonsson	> let fib = 1:1:zipWith (+) fib (tail fib) in fib
22:09	hsbot	[1,1,2,3,5,8,13,21,34,55,89,144,233,377,610,987,1597,2584,4181,6765,10946,17711,28657,46368,75025,121393,196418,317811,514229,832040,1346269,2178309,3524578,5702887,9227465,14930352,24157817,39088169,63245986,102334155,1...
22:09	ibdknox	brehaut: about enough I'd say
22:10	brehaut	mjonsson: was it the $ that tripped it up?
22:10	mjonsson	brehaut: yep
22:17	gfredericks	it's so weird to look at recursive code and not worry about what the poor jvm is going to think
22:18	gfredericks	> let fib = 1:4:zipWith (+) fib (tail fib) in fib
22:18	hsbot	[1,4,5,9,14,23,37,60,97,157,254,411,665,1076,1741,2817,4558,7375,11933,19308,31241,50549,81790,132339,214129,346468,560597,907065,1467662,2374727,3842389,6217116,10059505,16276621,26336126,42612747,68948873,111561620,180...
22:20	sleepynate	lambdabot in #clojure, this just seems so backward
22:22	brehaut	sleepynate: i think you're thinking of algolbot
22:24	sleepynate	har har
22:51	srid	&(or (and (zero? 1) 'zero) 'nonzero)
22:51	lazybot	⇒ nonzero
22:51	srid	&(or (and (zero? 0) 'zero) 'nonzero)
22:51	lazybot	⇒ zero
22:53	brehaut	srid: that looks like a complicated way to write if :P
22:54	amalloy	brehaut: i've written if that way before, in a language that had and/or expressions, and if statements
22:54	brehaut	amalloy: likewise; python frinstance
22:54	srid	its the habbit of doing "expr and 'foo' or 'bar'" in languages like python
22:55	srid	in lisp, an `if` is nothing special.
22:56	brehaut	srid: I'm also curious why you used quoted symbols rather than keywords (not that its really important here)
22:57	srid	oh right. forgot keywords. scheme-habit
22:57	srid	though i use them pretty often in my html code using hiccup
23:19	miwillhite	good eve; I know they are the same thing but which is more standard? (:name bucket) or (bucket :name)
23:19	miwillhite	assuming bucket is a map of course…
23:20	dnolen	miwillhite: (:foo x) preferred.
23:20	dnolen	,(nil :foo)
23:20	clojurebot	#<CompilerException java.lang.IllegalArgumentException: Can't call nil, compiling:(NO_SOURCE_PATH:0)>
23:20	dnolen	,(:foo nil)
23:20	clojurebot	nil
23:21	miwillhite	cool thanks
23:27	amalloy	miwillhite: http://stackoverflow.com/questions/7034803/idiomatic-clojure-map-lookup-by-keyword/7035984#7035984 gives a little detail on why that is, if you're interested
23:37	khaliG	so i've got a bunch of data files in the form of clojure readable sexps - and i'd like to batch modify the lot. is there an existing tool suitable for this purpose or shall i go ahead and roll my own?