#clojure – 2009-09-02

2009-09-02

00:05	technomancy	wow, I just got bit by lisp-1-ness for the first time ever.
00:08	lisppaste8	technomancy pasted "crazy proxy macro" at http://paste.lisp.org/display/86410
00:08	technomancy	so my methods map isn't working with init-proxy; not sure why
00:08	technomancy	guess I should try it with a map that doesn't need tweaking first
00:18	lisppaste8	technomancy annotated #86410 "almost works" at http://paste.lisp.org/display/86410#1
00:19	technomancy	so this works if I remove IMeta from the proxy definition
00:21	technomancy	the meta method just needs to return a map and take zero args, right?
00:21	technomancy	why would it be raising an UnsupportedOperationException?
00:22	hiredman	someing is trying to mutate something
00:22	hiredman	the clojure collections all throw that exception if you call a mutable part of the collections interface
00:23	hiredman	possibly something else of course
00:25	technomancy	http://p.hagelb.org/meta-exception.html <= it looks like the repl is actually raising the exception by trying to look up the type in the meta map
00:26	technomancy	oh... init-proxy _replaces_ a proxy's methods.
00:26	technomancy	I want update-proxy
00:26	hiredman	oh
00:27	hiredman	do you have a fn implementing that method in the map?
00:27	hiredman	proxy throws that for unimplemented interface methods too
00:27	technomancy	nope, just in the original proxy
00:27	clojurebot	proxy is <Chouser> proxy teases with its ease of use, then suddenly betrays.
00:27	technomancy	so true!
00:27	technomancy	,botsnack
00:27	clojurebot	java.lang.Exception: Unable to resolve symbol: botsnack in this context
00:27	hiredman	pashaw
00:27	technomancy	~botsnack
00:27	clojurebot	thanks; that was delicious. (nom nom nom)
00:33	technomancy	well that's enough Macro Madness for tonight.
00:33	technomancy	thanks.
02:07	LauJensen	Doesnt 'if' abort evaluation if the first item or can 'or' body evalutes to false?
02:08	hiredman	…
02:09	brool	LauJensen: not sure I understand the question
02:09	brool	"or" short circuits if any item is true
02:09	hiredman	not the most readable thing
02:09	LauJensen	,(if (or (= 2 1) (= 3 (+ 1 2))) (prn "foo") (prn "bar"))
02:10	clojurebot	"foo"
02:10	LauJensen	Is (+1 2) ever evaluated?
02:10	LauJensen	Answer then is no, thanks :)
02:10	hiredman	uh
02:10	brool	( + 1 2 ) is evaluated there
02:10	brool	because first condition is false
02:10	hiredman	↑
02:11	hiredman	(= 2 1) is false so (= 3 (+ 1 2)) is evaluated
02:11	hiredman	and is true, so the if expression evaluates to the result of evaluating (prn "foo")
02:12	hiredman	which prints out "foo" as a side effect
02:12	brool	,(or (do (println "did first clause... ") (= 2 1)) (do (println "did second clause... ") (= 3 (+ 1 2))))
02:12	clojurebot	true
02:12	clojurebot	did first clause... did second clause...
02:14	LauJensen	Yes yes I got it, thanks :) I wrongfully put in a false first clause
04:06	alinp	hi
04:06	alinp	I have a function like this:
04:06	alinp	(defn my-f [fn] (fn))
04:06	alinp	(def x (fn [a b] (+ a b)))
04:07	alinp	(my-f (x 2 3))
04:07	alinp	how can I make this work ?
04:07	alinp	I mean .. calling fn inside my-f
04:07	hiredman	…
04:07	hiredman	you need to pass my-f a function
04:08	hiredman	(x 2 3) is not a function
04:08	rathore	hiredman: just curious... what timezone are u in?
04:08	hiredman	west coast of america
04:09	rathore	hiredman: ok... just wondering, cause you're always here helping folks... very grateful to you :)
04:09	hiredman	:P
04:12	alinp	hiredman: but I need also to pass the argumens
04:12	alinp	arguments
04:12	alinp	I need to call apply inside my-f ?
04:13	hiredman	no
04:13	alinp	(defn my-f [fn & args] (apply fn args))
04:13	alinp	(my-f x 1 2)
04:13	alinp	this works
04:13	hiredman	sure that'll work too
04:14	alinp	can you please tell me how can I do that beside of apply ?
04:14	hiredman	pass in a functio to my-f
04:14	hiredman	my-f takes one arg fn
04:14	alinp	yeah
04:14	hiredman	the body of my-f is (fn) which is a function call
04:14	alinp	(def x (fn [a b] (+ a b)))
04:15	alinp	x is a function
04:15	hiredman	sure
04:15	alinp	so
04:15	hiredman	but (x 2 3) is not
04:15	alinp	ok ...
04:15	hiredman	(x 2 3) is function application
04:15	alinp	ok ...
04:15	hiredman	so what happens is (x 2 3) evaluates to 5
04:15	alinp	so, 5 is sent instead of fn
04:15	alinp	true
04:16	hiredman	so my-f attempts to call 5 as a function
04:16	hiredman	which, well, doesn't work
04:16	alinp	(my-f (fn [] (x 1 2)))
04:17	alinp	should do the trick
04:17	alinp	right ?
04:17	hiredman	yes
04:17	hiredman	there is also a handy bit of syntactic sugar for this
04:17	hiredman	#(x 1 2)
04:17	alinp	oh, yeah :)
04:17	alinp	I knew that
04:17	alinp	I like to be more verbose when comes to lambda functions ;)
04:18	alinp	but, thanks for reminding me
04:18	hiredman	sure
04:20	nsel	when clicking on stack trace frames in swank-clojure, I get "Synchronous Lisp Evaluation aborted." just want to check before filing a bug report, does anyone here get something different when they click on a stacktrace frame?
04:20	nsel	I've been seeing this behavior for months, and have latest git version
04:21	nsel	wouldn't be surprised if that's just how it's supposed to be, not yet implemented, as well
04:22	nsel	(test case: trigger any error which brings up a slime backtrace and click on a frame).
04:26	nsel	incase this wasn't clear, normally you'd expect it to jump to a source location.
05:03	AWizzArd	when I included a .txt file into my .jar, how can I then slurp the contents of that .txt file?
05:04	Chousuke	hmm
05:04	Chousuke	I think you can use @ in a path to mean "classpath-relative"
05:05	AWizzArd	I would like to include it in the .jar instead of putting it outside
05:05	AWizzArd	I already tried (slurp "@/com/company/file.txt")
05:05	hiredman	(.getResourceAsStream (clojure.lang.RT/baseLoader) "version") where version is the name of a textfile
05:06	hiredman	that gets you an inputstream
05:08	AWizzArd	maybe I can use duck-streams reader function with that
05:10	hiredman	,(doc reader)
05:10	clojurebot	"clojure.contrib.duck-streams/reader;[[x]]; Attempts to coerce its argument into an open java.io.BufferedReader. Argument may be an instance of Reader, BufferedReader, InputStream, File, URI, URL, or String. If argument is a String, it tries to resolve it first as a URI, then as a local file name. URIs with a 'file' protocol are converted to local file names. Uses default-encoding as the text encoding. Should be used in
05:10	hiredman	says it takes an input stream
08:48	lisppaste8	raphinou pasted "help for translating from OO to clojure" at http://paste.lisp.org/display/86427
08:48	raphinou_	if someone has the time to look at this paste, I can use some help
08:49	raphinou_	trying to translate OO code in clojure, and I don't see how to keep funtionality without the instance variables that are used in the OO example
08:54	noidi	raphinou_, you could create check-credentials inside make-login-widget
08:54	noidi	then it could access all the variables in a closure
08:55	raphinou_	yes, seems obvious now you say it :-)
08:55	Chouser	I keep losing the link to that blog post that walks through a few different ways to do "objects" in Clojure.
08:56	Chouser	but generally, closing over the "instance" data is one way. Another is to put that data in a vector or map and pass it around.
08:56	raphinou_	ok
08:57	raphinou_	I was afraid my function would grow horribly, but in a sense it's fine to have it defined in there, as it only makes sense to use it there
08:58	noidi	i thought about this myself a few days ago, and here's the toy code i worked with: http://pastebin.com/d44188f91
08:58	AWizzArd	raphinou_: you can use defstruct to define the fields of your "class" and define multimethods on it. One example, pretty much what Chouser said.
08:59	Chouser	no point in multimethods unless you expect to override them.
08:59	noidi	this macro even gets rid of the duplication in the clojure version http://groups.google.com/group/clojure/msg/5ef807214247472d
09:01	rhickey	raphinou_: just pass the data in a map
09:01	rhickey	then you get a nice function you can test without a gui
09:01	triyo	Anyone know of any native HTML parser for clojure?
09:02	Chouser	triyo: what do you mean by "native"? tagsoup has worked well for me.
09:02	raphinou_	thx all for your answers!
09:03	triyo	Chouser: by native I mean non-java introp way
09:04	Chouser	triyo: nope, don't know of such a thing. What would be the benefit of avoiding interop?
09:05	noidi	Chouser, is this the blog post you mentioned? http://blog.thinkrelevance.com/2009/8/12/rifle-oriented-programming-with-clojure-2
09:05	Chouser	noidi: no, though that's good too.
09:06	triyo	Chouser: no real benefit I guess except that it will perhaps be more lispy. I'll give tagsoup a try.
09:06	noidi	I love the analogy :) "Its a little like saying that a rifle is not arrow-oriented."
09:20	ole3	test
09:29	ambient	hmm, to me OO is mainly about hiding the complexity that arises from mutability. clojure doesnt really have a lot of that so therefore clojure doesn't need OO that much
09:29	ambient	also, OO is more about messages than it is about objects imo
10:23	LauJensen	Gents, what would be the most elegant way to scramble a sequence randomly? ie (scramble [1 2 3 4 5]) => [4 1 5 3 2], (scramble [1 2 3 4 5]) => [1 3 4 2 5] ?
10:25	Chousuke	LauJensen: probably best to use java for that.
10:25	Chousuke	convert to array, scramble, convert back to vector.
10:28	raphinou___	is it normal that drop-fruit is displayed at http://richhickey.github.com/clojure-contrib/sql-api.html ?
10:29	raphinou___	and in general the functions from the sql.test ns ?
10:35	Chouser	hm.. does seem likely that's a mistake.
11:15	LauJensen	Chousuke: When I said 'the most elegant way', what on earth got you thinking Java?
11:16	Fossi	LauJensen: i think there's a function for that in some util in contrib
11:16	Fossi	seq-util prolly
11:20	liwp	,(let [l (java.util.ArrayList. [1 2 3])] (do (java.util.Collections/shuffle l) l))
11:20	clojurebot	#<ArrayList [1, 3, 2]>
11:20	liwp	so how do you make a vector / list out of the ArrayList again?
11:20	Chousuke	LauJensen: because java probably has a method for scrambling arrays :P
11:21	liwp	and no, that's not particularly elegant, especially with the inplace shuffle, but it's quite short
11:21	Chousuke	LauJensen: and it's going to be faster than anything you can do with immutable vectors :)
11:21	liwp	Chousuke: yep, java.util.Collections/shuffle which takes a List
11:22	stuartsierra	c.c.seq-utils/shuffle just calls ArrayList/shuffle.
11:22	Chousuke	,(vec (ArrayList. [1 2 3]))
11:22	clojurebot	[1 2 3]
11:22	liwp	stuartsierra: heh, I didn't look that far :)
11:23	liwp	,(let [l (java.util.ArrayList. [1 2 3])] (do (java.util.Collections/shuffle l) (vec l)))
11:23	clojurebot	[2 1 3]
11:23	Chousuke	that's going to be as elegant as you can get I suppose.
11:23	stuartsierra	Others have written shuffle fns using, e.g., Knuth's algorithm for arrays of unknown length, but ArrayList/shuffle is much efficient.
11:23	stuartsierra	*more efficient.
11:24	liwp	stuartsierra: I'm not seeing a shuffle in ArrayList
11:24	Chousuke	hm
11:24	stuartsierra	It's part of the Collections interface, I think.
11:24	Chousuke	,(instance? List [])
11:24	clojurebot	java.lang.Exception: Unable to resolve symbol: List in this context
11:24	Chousuke	,(instance? java.util.List [])
11:24	clojurebot	true
11:25	stuartsierra	,(java.util.Collection/shuffle [1 2 3 4 5])
11:25	clojurebot	java.lang.IllegalArgumentException: No matching method: shuffle
11:25	stuartsierra	,(.shuffle [1 2 3 4 5])
11:25	clojurebot	java.lang.IllegalArgumentException: No matching field found: shuffle for class clojure.lang.LazilyPersistentVector
11:25	Chousuke	,(java.util.Collections/shuffle [1 2 3 4 5])
11:25	clojurebot	java.lang.UnsupportedOperationException
11:25	Chousuke	,(java.util.Collections/shuffle (transient [1 2 3 4 5]))
11:25	clojurebot	java.lang.Exception: Unable to resolve symbol: transient in this context
11:25	Chousuke	damn :P
11:26	Chousuke	I guess there is no getting around the mutability.
11:26	ambient	back to netbeans from emacs, enclojure seems to have developed nicely last time i tried it. with emacs it's a bit hard to add external libraries
11:26	Chousuke	but if it's only an implementation detail it doesn't exist :)
11:26	stuartsierra	Nope. seq-utils/shuffle copies to an ArrayList and calls shuffle on that.
11:26	clojurebot	shuffle is http://groups.google.com/group/clojure/msg/a2816f1b51d99b79
11:27	stuartsierra	there you go
11:27	liwp	thank you clojurebot!
11:28	Chousuke	for once the autocomment feature triggered at the right time :)
11:39	LauJensen	Sweet :)
11:40	LauJensen	Thanks for taking up the discussions guys :)
11:46	Chouser	I've got a question about the finger-tree paper. I don't suppose anyone here has read it enough to help me come up with an answer?
11:47	Licenser	42!
11:47	rhickey	not necessarily an impediment :)
11:47	rhickey	shoot
11:47	Licenser	rhickey: you're the clojure guy right?
11:47	Chouser	it looks to me like the measure function is applied at all levels of the tree
11:48	rhickey	Licenser: yes
11:48	Licenser	then good work! Keep it going :D
11:49	Chouser	that is, if the measure function is the domain of all measure functions has to include type a (the basic element type), type Node a, type Node Node a, etc.
11:49	slyrus_	hmm.. wasn't sure if there was going to be a #clojure, much less one with 142 users. :)
11:49	Chouser	s/if the/the/
11:49	Chousuke	heh
11:50	Chousuke	this channel is often better for getting in touch with the right people than the group is :P
11:50	slyrus_	group?
11:50	Chousuke	the google group
11:50	slyrus_	ah, ok
11:51	Chouser	bleh, I messed that up too much. I'll try again.
11:51	Chouser	the domain of all measure functions has to include type a (the basic element type), type Node a, type Node Node a, etc.
11:53	Chouser	but that doesn't seem very nice to me. surely the measure fn ought to only have to understand type a
11:55	rhickey	doesn't \|\| pluck out the v?
11:56	Chouser	\|\| on a node simply returns the cached v, yes.
11:56	rhickey	doesn't that happen always before the measure is applied?
11:57	Chouser	But when constructing the node with the "smart constructors" at the top of section 4.2, they just do \|\|a\|\| (+) \|\|b\|\|, where a and b could be of any nested Node depth
11:57	djpowell	Hmm, whilst the 'recur' examples on the transients page work well, I'm finding that doing something like: (reduce (fn [m [k v]] (assoc! m k v)) (transient m1) mymap) - doesn't perform much differently than the persistent version - is that to be expected?
11:57	Chouser	oh
11:58	Chouser	right. ok, if a is a Node, \|\| on it will return the cached.
11:58	Chouser	only if a is a Digit will it try to apply the user-supplied measure fn
11:59	Chouser	somehow I lost this in the conversion from vectors-with-metadata to reify'd objects, and then got too muddled to put it back together.
12:00	rhickey	Chouser: I know you've been planning on exposing the measure system generically, just want to point out the utility of a version with it hardwired: http://hackage.haskell.org/packages/archive/containers/0.2.0.1/doc/html/Data-Sequence.html
12:01	Chouser	rhickey: yes, I have vague plans to do both.
12:01	rhickey	It all starts with vague plans :)
12:02	Chouser	that is, a generic underlying thing with no built-in measure (perfect for simple queues)
12:02	Chouser	then something built on top that implements Indexed by measuring size
12:03	Chouser	perhaps some nicely packaged priority queue. etc.
12:26	djpowell	I think my problem is probably that even the persistent collections perform so well that the calculations I am doing dominate the time taken to build the collection
12:26	djpowell	which is good
12:39	ole3	hello, is there a way to stop and resume processes in clojure?
12:44	leafw	ole3: process are java Threads. So yes.
12:46	ole3	leafw: really? I thought a process lives in its own address space.
12:48	leafw	er, a virtual machine is just that, virtual.
12:49	ole3	leafw: ok, can you tell me how?
12:50	ole3	leafw: something like SIGSTOP?
12:51	leafw	ole3: are you talking about OS process, which are just the JVM ?
12:54	ole3	leafw: I'm talking about an OS process started via (.exec (Runtime/getRuntime) ...)
12:54	leafw	ole3: ah! Then, catch the returned object
12:54	ole3	leafw: ok got it, stored in process :)
12:55	leafw	yes
12:55	ole3	leafw: Here it is: #<UNIXProcess java.lang.UNIXProcess@16917ee>
12:55	leafw	unfortunately, the http://java.sun.com/j2se/1.5.0/docs/api/java/lang/Process.html seems not to have anything else to do other than destroy(), so you have to cast blindly
12:56	leafw	to the subclass.
12:56	leafw	i.e. inspect its methods, it may have what you want :)
12:56	ole3	hm, to what subclass?
12:57	ole3	It has exitValue, destroy, waitFor, but not stop
12:57	leafw	java.lang.UNIXProcess is likely a subclass of java.lang.Process
12:58	dave[csc]	Is there an issue with defining an empty vector? a la (def myvec vec)
12:58	ole3	yes, it is one
12:58	leafw	ole3: then the UNIXProcess maybe has other methods. If not, then ... you are stuck.
12:59	ole3	leafw: oh my god
13:01	ole3	leafw: does this mean i have to use jni or jna for stopping processes
13:01	Chousuke	can't you send a suspend signal to the process?
13:02	ole3	i don't know the pid
13:13	Chouser	dave[csc]: vec is not an empty vector, it's a function to create vectors. [] is an empty vector
13:22	ole3	ok, i will use jna, thanks for your help
13:28	ambient	im wondering if i can just use arrays & stuff for mutable data structures and not really care
13:28	ambient	i really love the lispiness but the immutability is seriously giving me a headache
13:30	Chouser	ambient: sure you can, but it'll be better for you in the long run if you push through on the immutability front.
13:31	ambient	well i find it really hard to vision games done mostly with immutable data
13:33	Chouser	performance issues are one thing. headaches another.
13:41	ericthorsen	I see there is a logging namespace in clojure.contrib but it does not appear that anyone is using it (in the contrib lib I mean)
13:42	ericthorsen	Is there a preferred way to handle this in the contrib code? or is logging frowned upon?
13:42	technomancy	ericthorsen: I just started using it
13:42	technomancy	planning on backporting it to the 1.0-compat branch, actually
13:43	Chouser	oh, I see -- if a contrib lib needs to log something?
13:43	ankou	hi, is it possible (and advised) to split one namespace over several files?
13:43	ericthorsen	Chouser: exactly
13:44	Chouser	ankou: certainly possible. proxy and genclass are in clojure.core namespace but live in their own .clj files.
13:44	Chouser	ericthorsen: dunno, haven't needed it. For me it's seemed best so far to leave error handling up to the app -- it can log or abort or whatever as needed.
13:45	ericthorsen	technomancy: it is omitted from the build and does not appear to load unless it can see all the libs (log4j, apache.commons.logging, etc.) which I assume defeats the purpose. Perhaps I do not know how to use it?
13:45	technomancy	ericthorsen: it works with java.util.logging if the others are not found
13:45	ankou	Chouser: is it recommended to split it in several files and load these?
13:46	ankou	Chouser: better question: is this errorprone or something like that?
13:46	Chouser	ankou: Not sure when it's recommended but it's definitely supported.
13:47	ericthorsen	Chouser: You do not see utility in library code providing any logging?
13:47	Chouser	ankou: note how the top of core_proxy.clj differs from most .clj lib files. It's a "secondary" file, so it doesn't define the namespace with an 'ns' macro call. instead it uses in-ns and import.
13:48	Chouser	ericthorsen: I haven't written a libarary where it felt like the right choice. I'm making no claims about the general case. :-)
13:49	ericthorsen	Chouser: ok
13:49	technomancy	ericthorsen: the logging library is quite new.
13:49	ericthorsen	technomancy: I see.
13:51	ericthorsen	Chouser: I find having the ability to enable logging in APIs I am not the author of useful in troubleshooting problems. It has been a life saver working in the behemoth Netbeans framework for example.
13:52	ericthorsen	Chouser: I can do what I need to in contrib without it for sure...just curious as to what the current thinking was there
13:52	ankou	Chouser: so it is convention to use a namespace for each file even in normal non-library programs?
13:52	ankou	Chouser: or do they normally use the User namespace ?
13:53	Chouser	ankou: Most namespaces are defined by a single file.
13:54	Chouser	ankou: I'd not recommend using the 'user' namespace except for repl, work-in-progress, toy code, etc. It won't AOT compile, might clash with others code, etc.
13:55	Chouser	any lib or releasable app ought to use its own namespace
13:55	ankou	Chouser: with several minor namespaces? because I think it would somewhat confusing to have several thousand lines in one file
13:56	opqdonut	sure
13:56	opqdonut	i use subnamespaces all the time
13:57	Chouser	sure, if your code gets to be more than you want for a file, you can add more files, either in the same namespace or in new namespaces.
13:57	Chouser	I guess I would generally add a new namespace when I add a new file, unless I had a specific reason to extend an existing namespace.
13:58	hiredman	calling them subnamespaces is some what confusing
13:59	ankou	but you know what I mean
13:59	ankou	or opqdonut xD
13:59	ankou	okay thanks
13:59	hiredman	as far as I can tell there is no hierarchy of namespaces just like there is no hierarchy of java packages
13:59	opqdonut	yeh
13:59	Chouser	right, just a hierarchy of names. :-/
14:17	mibu	I have a peculiar bug with for. The first binding form is always binding to the first element of a sequence. Can someone shed some light on this behavior? http://codepad.org/8BY0KlQi
14:17	mibu	I'm using 1.1
14:20	stuartsierra	for iterates over all bindings, last-fastest
14:20	stuartsierra	if the last binding is to an infinite sequence, the first binding will never advance
14:21	mibu	what do you mean?
14:21	Chouser	it's like a nested loop.
14:21	stuartsierra	,(for [i (range 5) j (range 5)] [i j])
14:21	clojurebot	([0 0] [0 1] [0 2] [0 3] [0 4] [1 0] [1 1] [1 2] [1 3] [1 4] [2 0] [2 1] [2 2] [2 3] [2 4] [3 0] [3 1] [3 2] [3 3] [3 4] [4 0] [4 1] [4 2] [4 3] [4 4])
14:21	mibu	oh, I see.
14:21	mibu	so, you can't bind to infinite sequences in for?
14:22	stuartsierra	Not if they're nested.
14:22	Chousuke	you can, but it makes no sense :)
14:22	mibu	what's a better elegant way to create pairs from a single sequence?
14:22	Chousuke	at least, not without :while
14:22	Chouser	sure you can, but if any level is infinite and not terminated by a :while, then you'll just never get to the next iteration of the outer level
14:22	Chousuke	partition?
14:23	mibu	chousuke: yes! thank you.
14:23	mibu	I hate it when I forget simple useful functions.
14:24	Chousuke	note that partition may require padding if you have an uneven sequence
14:24	mibu	I have an infinite sequence.
14:24	Chousuke	oh, okay :)
14:26	mibu	There ought to be a place with all those little gotchas for every function. I haven't coded in clojure for two months and I already forgot plenty of those gotchas.
14:27	Chousuke	start a blog series.
14:27	Chousuke	:)
14:28	ambient	reverse of partition seems to be for
14:28	mibu	I was thinking maybe an add-on to the internal documentation (doc). you know, with examples, gotchas, guidelines, and other cookbook stuff.
14:28	Chousuke	heh
14:28	ambient	[[2 4] [3 5]] (for [x coll] (first x)) seems to be shortest way ot generating seq from the first items
14:29	Chouser	(map first coll)
14:29	ambient	:( dohh
14:29	Chousuke	and the reverse of partition is concat
14:29	Chousuke	,(apply concat (partition 2 [1 2 3 4])
14:29	clojurebot	EOF while reading
14:29	Chousuke	whoops
14:29	Chousuke	,(apply concat (partition 2 [1 2 3 4]))
14:29	clojurebot	(1 2 3 4)
14:34	ambient	(map #(play-note chan %1 %2) (map first phr) (map second phr)) im sure there's more elegant way to put this
14:37	Chousuke	hm
14:38	Chousuke	play-note looks like it should have a side-effect :P
14:38	ambient	it has
14:38	Chousuke	(doseq [[x y] phr] (play-note chan x y)) then
14:40	Chousuke	you should avoid mapping functions with side-effects
14:41	Chousuke	,(do (map println ["foo" "bar"]) "foiled")
14:41	clojurebot	"foiled"
14:41	ambient	because map doesn't guarantee sequential access?
14:41	Chousuke	,(do (doall (map println ["foo" "bar"])) "not foiled")
14:41	clojurebot	"not foiled"
14:41	clojurebot	foo bar
14:41	Chousuke	because map is lazy :)
14:41	Chousuke	doall forces eagerness
14:42	Chousuke	or dorun, if you don't care about the results.
14:43	ambient	this is going to be absolutely horrible..
14:43	ambient	im generating a thread for each channel and each pattern
14:43	Chousuke	hm.
14:43	Chousuke	you might have problems keeping them in sync :P
14:46	ambient	or alternatively i could convert these: (def phr-melod-1 [[60 250] [64 250] [60 250] [65 250]]) into event model
14:47	ambient	so i need to convert from note duration to its position in time
14:57	ambient	damn.. its surprisingly complex to build accumulating value over seq and make a new seq out of it
14:58	Chouser	:-)
14:58	ambient	or perhaps i could use reduce
14:58	Chouser	,(reduce (fn [[acc time] [dur freq]] [(conj acc [time freq]) (+ time dur)]) [[] 0] [[1 :a] [4 :b] [3 :c] [1 :d]])
14:58	clojurebot	[[[0 :a] [1 :b] [5 :c] [8 :d]] 9]
14:58	rhickey	,(reductions + (range 5))
14:58	clojurebot	java.lang.Exception: Unable to resolve symbol: reductions in this context
14:58	rhickey	(clojure.controb.seq-utils/reductions + (range 5))
14:58	rhickey	,(clojure.controb.seq-utils/reductions + (range 5))
14:58	clojurebot	java.lang.ClassNotFoundException: clojure.controb.seq-utils
14:59	rhickey	,(clojure.contrib.seq-utils/reductions + (range 5))
14:59	clojurebot	java.lang.ClassNotFoundException: clojure.contrib.seq-utils
14:59	Chouser	,(use 'clojure.contrib.seq-utils)
14:59	clojurebot	nil
14:59	rhickey	,(reductions + (range 5))
14:59	clojurebot	(0 1 3 6 10)
14:59	ambient	nifty :)
15:00	Chouser	the problem in this case is you're off by one
15:00	Chouser	,(reductions (fn [[p v] [d v]] [(+ p d) v]) [[1 :a] [4 :b] [3 :c] [1 :d]])
15:00	clojurebot	([1 :a] [5 :b] [8 :c] [9 :d])
15:01	Chouser	:b should start at time 1, not 5. 5 is when :b should end and :c should start.
15:05	Chouser	,(map (fn [[[t] [_ v]]] [t v]) (partition-all 2 1 (concat [[0]] (reductions (fn [[p] [d v]] [(+ p d) v]) [[1 :a] [4 :b] [3 :c] [1 :d]]))))
15:05	clojurebot	([0 :a] [1 :b] [5 :c] [8 :d] [9 nil])
15:05	Chouser	destructuring ftw!
15:07	LauJensen	ftw?
15:08	Chouser	For The Win!
15:08	LauJensen	ah okay
15:09	ambient	(map #(- % (second (first phr))) (reductions + (map second phr))) seems to work ok :)
15:10	Chouser	ambient: heh. well, that's better.
15:10	ambient	but thats not the best way to do it i think
15:14	ambient	(all-but-the-last (reductions + (map second (cons [0 0] phr-melod-1))
15:15	ambient	no.. (all-but-the-last (cons 0 (reductions + (map second phr))))
15:16	Chousuke	all-but-the-last is called butlast :P
15:16	ambient	ok, good to know
15:16	Chouser	or drop-last
15:17	Chouser	drop-last is lazy, butlast is not.
15:23	ankou	hi again, is there a possibility to use string manipulation with symbols or how are the .javaShortFunctionCalls created?
15:24	hiredman	huh?
15:25	Chouser	,(str 'foo) ; like this?
15:25	clojurebot	"foo"
15:25	hiredman	,(name 'foo)
15:25	clojurebot	"foo"
15:25	Chouser	,(symbol (str "." "foo" "Bar")) ; or like that?
15:25	clojurebot	.fooBar
15:26	ankou	oh great :D
15:28	Chousuke	ankou: if you need it in a macro you can just use . directly
15:34	ambient	(partition 2 (interleave (map first phr)
15:34	ambient	(drop-last (cons 0 (reductions + (map second phr)))))))
15:34	ambient	isnt very pretty :/
15:34	ambient	but gets the job done
15:35	LauJensen	ambient: thats kinda funny because I've heard Chouser described that exact way by his employer
15:35	ambient	hehe
15:38	ambient	in an imaginary language the problem would be about: new second element = accumulate old from 0
15:39	ambient	how that could be expressed in sexps is beyond me
15:40	hiredman	(-> phr ((partial map second)) ((partial reductions +)) ((partial cons 0)) drop-last)
15:41	Chousuke	:P
15:42	Chousuke	that's just ugly.
15:42	hiredman	(comp drop-last (partial cons 0) (partial reductions +) (partial map second))
15:42	hiredman	Chousuke: hardly
15:42	Chousuke	the latter is much better.
15:43	Chousuke	the first one is just abuse of ->
15:43	Chousuke	it happens :)
15:44	ambient	comp/partial, seems useful
15:44	Chousuke	mostly the weird double parentheses offend me.
15:44	hiredman	whenever I play around try to work out a function in a repl it always turns into a big long -> expression
15:44	hiredman	you get used to the double parens
15:45	Chousuke	I don't use partial that much either
15:45	LauJensen	Chousuke: Did you catch this anonymous contribution?
15:45	LauJensen	,(reduce #(let [i (rand-int %2) j (dec %2)] (-> %1 (assoc i (%1 j)) (assoc j (%1 i)))) [1 2 3 4 5] (range 5 1 -1))
15:45	clojurebot	[2 1 3 5 4]
15:46	hiredman	gnarly
15:46	Chousuke	it's better in haskell because of autocurrying
15:46	Chousuke	but I like being explicit.
16:16	angerman	how would I nuke all lines from a string that start with %.... ?
16:17	angerman	e.g. I have a file that uses % as comment indicator
16:18	tomoj	angerman: you read the entire file into a string? or a seq of lines?
16:18	ankou	is there a function to expand inner lists? something which makes ((1 2) (3 4)) to (1 2 3 4)?
16:18	angerman	tomoj: I read the whole file it's not that large
16:19	angerman	tomoj: but I guess it's like (split on \n resp \r\n)
16:19	angerman	filter the lines and glue it back together.
16:19	hiredman	,(doc line-seq)
16:19	clojurebot	"([rdr]); Returns the lines of text from rdr as a lazy sequence of strings. rdr must implement java.io.BufferedReader."
16:20	hiredman	,(remove #(re-find #"^%.*$" %) ["foo bar baz" "%foo" "bar"])
16:20	clojurebot	("foo bar baz" "bar")
16:21	angerman	hiredman: thanks. that looks not too bad :)
16:21	tomoj	eh
16:22	angerman	well I guess I screwed up :...
16:22	tomoj	why use a regular expression to do .startsWith?
16:23	angerman	hm line-seq really wants a BufferedReader... slurp's so nice
16:23	tomoj	I guess if whitespace is allowed before the % then a regular expression becomes more useful
16:23	hiredman	tomoj: shurg
16:23	tomoj	#"^\s*%" ?
16:26	angerman	hmm.
16:26	angerman	maybe I should tackle the problem from a differnt side
16:26	angerman	my file consists of blocks separated by two new lines.
16:26	angerman	and there may be lines in there that start with %. so all I need at first is just the blocks as a sequence.
16:27	angerman	my first thought was: break file into lines, remove all those that startwith %, glue back together and split on \n\n
16:31	hiredman	,(doc partition)
16:31	clojurebot	"([n coll] [n step coll] [n step pad coll]); Returns a lazy sequence of lists of n items each, at offsets step apart. If step is not supplied, defaults to n, i.e. the partitions do not overlap. If a pad collection is supplied, use its elements as necessary to complete last partition upto n items. In case there are not enough padding elements, return a partition with less than n items."
16:31	hiredman	(comp (partial partition 2) line-seq)
16:31	hiredman	or something
16:49	angerman	what would be a lisp way to scan a list and when one finds a symbol takes all that has been scanned a turns it into a list, from there on starts with a second list and so on
16:50	angerman	e.g. (... "a" ... ... "a" ... ... "a" ... ...) -> (( ... ) (... ...) (... ...))
16:51	ambient	split?
16:51	angerman	would still have to scan the last (... ...)
16:51	ambient	that oddly reminds me of cartman
16:51	hiredman	reduce
16:51	angerman	ambient: well the symbol is actually two in a row. so I though about using peek
16:52	hiredman	if you are parsing stuff, you could just, you know, using a parsing library like fnparse
16:52	hiredman	of course fnparse apears to be under "heavy developement" at the moment
16:52	hiredman	:/
16:53	hiredman	~fnparse
16:53	clojurebot	Pardon?
16:53	hiredman	~parse
16:53	clojurebot	parser is http://github.com/joshua-choi/fnparse/tree/master
16:59	angerman	so basically I can use fnparse to construct me a tree like structure?
17:01	ambient	anyone know how i can easily manage additional java libraries in emacs?
17:01	ambient	or must i always change .emacs by hand
17:01	hiredman	angerman: you can use it to parse things
17:03	angerman	hiredman: ok. I guess I'll look a lot deeper into that.
17:06	powr-toc	ambient: if you use a script to start clojure, you can have that script consult a .clojurerc file which contains a classpath decleration
17:08	nsinghal	(let [{:keys [a b & ss]} {:a 1 :b 2 :c 3 :d 4}] ss)
17:08	nsinghal	I am trying to capture rest of map in "ss" (other than a & b). How can i do that?
17:09	Chousuke	hm, you can't do that, but you can have access to the whole map.
17:09	hiredman	,(dissoc {:a 1 :b 2 :c 3 :d 4} :a :b)
17:09	clojurebot	{:c 3, :d 4}
17:09	ambient	i got (0 250 500 750 1000), how do i get (250 250 250 250)?
17:10	Chousuke	,(let [{:keys [a b c] :as m} {:a 1 :b 2 :c 3 :d 4}] [a b c m])
17:10	clojurebot	[1 2 3 {:a 1, :b 2, :c 3, :d 4}]
17:10	hiredman	,(repeat 4 (second '(0 250 500 750 1000)))
17:10	clojurebot	(250 250 250 250)
17:11	ambient	:p yeah.. im actually looking for the differences
17:11	ambient	but i think it will come to me after a while, i shall save my "points" for asking actually difficult questions
17:12	nsinghal	i thought it wont work - but still was worth a try to ask
17:13	tomoj	,(map (fn [[a b]] (- b a)) (partition 2 1 '(0 250 500 750 1000)))
17:13	clojurebot	(250 250 250 250)
17:13	nsinghal	it will be nice to have that - validating map input might become easy. (defn [{:keys [a b & ss]}] (when-not ss ...)
17:14	hiredman	,(map #(- %1 (* %2 250)) '(0 250 500 750 1000) (iterate inc -1))
17:14	clojurebot	(250 250 250 250 250)
17:17	bostwick	Hi
17:18	bostwick	Does anyone have experience with the clojure plugin for eclipse?
17:19	bostwick	I'm writing a program to do some xml processing, and when I attempt to print a hash representation of my xml from the repl, eclipse freezes
17:24	tomoj	seems like a "mapply" might be useful sometimes
17:25	tomoj	like, map, but instead of calling the function on each element, apply the function to each element (where each element is a seq itself)
17:26	replaca	you mean like (map (partial apply f) ...)?
17:27	tomoj	guess so
17:27	replaca	I agree with whoever said earlier (Chosuke?) that they missed "auto-currying from Haskell
17:27	tomoj	,(map (partial apply #(- b a)) (partition 2 1 '(0 250 500 750 1000)))
17:27	clojurebot	java.lang.Exception: Unable to resolve symbol: b in this context
17:27	replaca	it would be nice to have comp and partial be a little more natural
17:27	tomoj	err
17:28	tomoj	,(map (partial apply #(- %2 %1)) (partition 2 1 '(0 250 500 750 1000)))
17:28	clojurebot	(250 250 250 250)
17:28	tomoj	yup
17:28	Chousuke	partial apply?
17:28	brool	replaca: yah, i would like that as well
17:29	replaca	(defn mapply [f & rest] (map (partial apply f) rest))
17:29	replaca	I think
17:30	tomoj	but then all the seqs are in a seq
17:32	Chousuke	some reader syntax for comp or partial might be nice... but is there anything left? :P
17:35	ambient	i need to separate items in a list like [4 4 4 3 3 1 1 1 1] -> [ [4 4 4] [3 3] [1 1 1 1] ] any easy way to do this?
17:35	ambient	ive been browsing the API but my eyes start to water
17:37	Chousuke	(doc group-by)
17:37	clojurebot	"([f coll]); Returns a sorted map of the elements of coll keyed by the result of f on each element. The value at each key will be a vector of the corresponding elements, in the order they appeared in coll."
17:38	Chousuke	hm not that.
17:43	tomoj	ambient: my attempt: http://paste.lisp.org/display/86466
17:44	ambient	tomoj, well that was my backup plan if there weren't any short combination of existing functions :)
17:46	tomoj	seems like a weird thing to do
17:46	ambient	not so weird when its part of a larger thingie
17:46	ambient	((9 35 0) (4 60 0) (4 0 250) (4 64 250) (9 0 500) (9 38 500) (4 0 500) (4 60 500) (4 0 750) (4 65 750) (9 0 1000) (4 0 1000)) i need to group this by the last items
17:47	tomoj	oh
17:47	tomoj	that's different
17:47	tomoj	:)
17:47	tomoj	I was thinking it was weird because having [1 1 1 1] seems silly if you know that they're all the same anyway
17:48	ambient	but i can just (map #(weird-thingie (last %)) myseq)
17:48	ambient	hmm wait
17:49	tomoj	you can just use my split-groups except pass a function in to use instead of just doing plain =
17:49	ambient	im still trying to come up with a higher level solution
17:50	tomoj	group by works but it gives you something different
17:50	ambient	but i got your version saved for future reference
17:50	tomoj	depends on if order matters
17:50	ambient	order matters
17:50	tomoj	within a group only? or you need the groups in the right order too?
17:50	hiredman	,x
17:50	clojurebot	((9 35 0) (4 60 0) (4 0 250) (4 64 250) (9 0 500) (9 38 500) (4 0 500) (4 60 500) (4 0 750) (4 65 750) (9 0 1000) (4 0 1000))
17:50	hiredman	vals (reduce (fn [map group] (update-in map [(last group)] conj group)) {} x))
17:50	hiredman	er
17:51	hiredman	,(vals (reduce (fn [map group] (update-in map [(last group)] conj group)) {} x))
17:51	clojurebot	(((4 0 1000) (9 0 1000)) ((4 65 750) (4 0 750)) ((4 60 500) (4 0 500) (9 38 500) (9 0 500)) ((4 64 250) (4 0 250)) ((4 60 0) (9 35 0)))
17:51	tomoj	wat
17:51	tomoj	oh, that's just like group by though, right?
17:51	ambient	well those are instrument events (channel note time-location)
17:51	ambient	when note = 0 its note off
17:51	tomoj	I mean, wrong order
17:52	hiredman	by grouping you are changing the order
17:52	hiredman	I don't see how you can expect to keep an order
17:52	ambient	main thing is that note-offs dont come after the note-ons in the same time-slot
17:52	tomoj	uhh
17:53	hiredman	,(vals (reduce (fn [map group] (update-in map [(last group)] conj group)) (sorted-map) x))
17:53	clojurebot	(((4 60 0) (9 35 0)) ((4 64 250) (4 0 250)) ((4 60 500) (4 0 500) (9 38 500) (9 0 500)) ((4 65 750) (4 0 750)) ((4 0 1000) (9 0 1000)))
17:53	ambient	it can be explicitly defined though
17:53	tomoj	[4 4 4 4 3 3 3 1 1 1] -> [[4 4 4 4] [3 3 3] [1 1 1]] ?
17:53	tomoj	ah, that works
17:53	hiredman	tomoj: his sample is not ordered that way though
17:53	hiredman	,x
17:53	clojurebot	((9 35 0) (4 60 0) (4 0 250) (4 64 250) (9 0 500) (9 38 500) (4 0 500) (4 60 500) (4 0 750) (4 65 750) (9 0 1000) (4 0 1000))
17:54	hiredman	oh, it is, anyway
17:54	ambient	note-on note-off only matters in the same channel, which is the first val
17:54	hiredman	you would map a sort function over each group
18:14	lisppaste8	ambient pasted "split-groups" at http://paste.lisp.org/display/86467
18:14	ambient	i did scheme-y version of it that maintains order
18:14	ambient	it's pretty simple, but i dont know what patterns can be extracted from that
18:21	ambient	nvm.. i need to sleep
18:21	slyrus_	I get a bunch of errors like this: [java] java.io.FileNotFoundException: Could not locate clojure/stacktrace__init.class or clojure/stacktrace.clj on classpath: (jmx.clj:10)
18:21	slyrus_	when trying to build clojure-contrib
18:21	slyrus_	should I worry?
18:22	ambient	i usually have to tell clojure.jar path manually to ant: ant -Dclojure.jar=../clojure/clojure.jar
18:37	slyrus_	ambient: yeah, I did that... it fails earlier if I don't do that
19:35	rhickey	put Clojure data directly into RDF triple store: (store-root #^{:woo :hoo} {:a 1 :b 2 :c #{3 4 5} :d {:e 6 :f [7 8 9]} :g [1 2 3] :h "ten" :i 10.42})
19:37	rhickey	reload and recover full structure in a single call
19:37	rhickey	org.clojure.rdfm=> (reload (::id ^*1))
19:37	rhickey	{:i 10.42, :h "ten", :g [1 2 3], :d {:f [7 8 9], :e 6}, :c #{3 4 5}, :b 2, :a 1}
19:38	rhickey	org.clojure.rdfm=> ^*1
19:38	rhickey	{:org.clojure.rdfm/id #<URI urn:uuid:13e52fec-ce9b-4060-bc64-5d02792495fa>, :woo :hoo}
22:10	tomoj	what's the "triple" there?
22:15	hiredman	http://www.thelastcitadel.com/images/prolog.jpg <-- $2 at goodwill
22:25	tomoj	is there a "clojure monads for dummies" out there?
22:26	tomoj	I am tremendously confused by both of the intros I've tried to read
22:26	hiredman	monads are confusing
22:38	tomoj	maybe I should just finally learn haskell
23:33	technomancy	so I'm trying to use update-proxy to provide a proxy with a map of methods.
23:33	technomancy	I'm using zipmap to turn the keyword keys into strings, and that works fine
23:34	technomancy	I also want to wrap the fn values in the map in another map, but I get a strange ClassNotFoundException when I try it.
23:34	technomancy	http://p.hagelb.org/applet.clj.html
23:34	technomancy	^ if I replace the bind-applet with identity in the zipmap, it works fine
23:39	technomancy	it looks like if bind-applet returns a named function it works, but returning an anonymous function breaks
23:41	technomancy	even stranger: returning a named function from clojure.core or clojure.xml works, but returning a named function from my own namespace or from contrib breaks.
23:44	hiredman	sounds like a classloader issue