#clojure – 2009-04-07

2009-04-07

03:41	dysinger	~java.lang.Thread/currentThread
03:41	clojurebot	?
03:42	dysinger	~(.. java.lang.Thread/currentThread getContextClassLoader)
03:42	clojurebot	?
03:43	dysinger	why is this happening ?
03:43	dysinger	Clojure
03:43	dysinger	user=> java.lang.Thread/currentThread
03:43	dysinger	java.lang.Exception: No such namespace: java.lang.Thread (NO_SOURCE_FILE:0)
03:46	dysinger	strange
03:46	dysinger	user=> Thread/currentThread
03:46	dysinger	java.lang.Exception: No such namespace: Thread (NO_SOURCE_FILE:0)
03:46	dysinger	user=> Thread
03:46	dysinger	java.lang.Thread
03:46	dysinger	user=>
03:47	AWizzArd	,Thread
03:47	clojurebot	java.lang.Thread
03:47	AWizzArd	,java.lang.Thread
03:47	clojurebot	java.lang.Thread
03:48	AWizzArd	,Thread/currentThread
03:48	clojurebot	java.lang.Exception: No such namespace: Thread
03:50	AWizzArd	,(Thread/currentThread)
03:50	clojurebot	#<Thread Thread[Thread-16,5,main]>
03:50	dysinger	ah my goof
03:50	AWizzArd	dysinger: currentThread is not a public field but instead a method
03:50	AWizzArd	So it should be put into parens :)
03:50	dysinger	y I goofed
03:50	dysinger	noob
04:06	alphazero	good morning / hello all
04:06	AWizzArd	Hi alpha.
04:07	alphazero	Hi there AWizzArd.
04:08	alphazero	Quick q: has anyone tried bolting clojure to anything besides couchDB for key value hash dbs?
04:25	dysinger	how do I proxy java.util.logging.Formatter ?
04:25	dysinger	(proxy [Formatter] [] (format [record] ... er wait format is clojure method
04:26	dysinger	even emacs gets it all mangled
04:26	Cark	do you have an error trying to compile this ?
04:26	dysinger	the problem is the method to proxf is "format"
04:27	dysinger	not compiling yet
04:27	dysinger	just curious as to what happens - maybe it's fine - other than emacs mangles the formatting
04:27	Cark	it's working for me
04:27	dysinger	ok
04:27	Cark	(.format blah) and (format blah) are two different functions
04:28	dysinger	y it's just not .format in the proxy form it's plain "format" which clojure-mode then things is the clojure method
04:28	dysinger	s/things/thinks
04:34	dysinger	sorry for the noob questions
04:35	dysinger	is there a way to pull an array out of a bean (set) easy? say I want "x" and "y" properties
04:35	dysinger	{ :x 1 :y 2 :z 3 }
04:36	dysinger	repeating [(:x myset) (:y myset)] isn't right I bet
04:44	dysinger	sorry I meant pull out multiple vals from a map
04:44	dysinger	not set
04:44	dysinger	I think I want select-keys
04:46	dysinger	,(select-keys {:one 1 :two 2 :three 3} '(:one :two))
04:46	clojurebot	{:two 2, :one 1}
04:47	dysinger	hmm closer but still not there
04:48	Chouser	,(map {:one 1 :two 2 :three 3} '(:one :two))
04:48	clojurebot	(1 2)
04:49	dysinger	nice
04:49	dysinger	now to guarantee order
04:50	dysinger	just add sort
04:50	Chouser	those will be in the order given in the list
04:50	Chouser	,(map {:one 1 :two 2 :three 3} [:three :one])
04:50	clojurebot	(3 1)
04:51	dysinger	ah that's right
04:51	dysinger	I was thinking of my version above
04:51	dysinger	map would work off a seq in order
04:51	dysinger	thanks
04:52	Chouser	right
05:03	dysinger	Java 1.4-1.6 has java.util.logging -> is there something like the simple formatter that is clojure friendly? the default java formatter does this "Apr 7, 2009 3:55:57 AM sun.reflect.NativeMethodAccessorImpl invoke0 FINE: testing"
05:03	dysinger	I quickly scanned contrib and didn't see anything stand out.
05:04	dysinger	Just want to make sure I am not repeating ground
05:24	cemerick	dysinger: we've had good results using log4j in conjunction with clojure.
05:24	cemerick	not sure what you mean by 'clojure friendly' tho...
05:24	dysinger	I don't want to use commons or log4j
05:25	dysinger	by clojure friendly I mean the namespace instead of "sun.reflect.NativeMethodAccessorImpl invoke0"
05:26	dysinger	I guess I could use log4j but what's the point? that was 10 years ago and I am only on java 6
05:27	dysinger	It doesn't add anything but extra weight IMO
05:27	cemerick	well, java.util.logging is a pale shadow of a kit compared to log4j IMO, but that's emacs v. vi territory.
05:27	dysinger	thats true
05:28	dysinger	I just don't need JMS appenders etc
05:28	cemerick	the class that is emitted in the log msg should be the one where you invoke the log method, regardless of your kit
05:28	cemerick	perhaps the log message you pasted was generated by a call from an anonymous fn or somesuch?
05:29	cemerick	we don't use any of the fancy appenders, either. We're with log4j primarily because of the formatting flexibility.
05:29	dysinger	yes you are right - I was using the repl to try it
05:30	dysinger	but even from within clojure logging I get "sun.reflect.GeneratedMethodAccessor7 invoke"
05:30	dysinger	instead of my Logger's name
05:33	dysinger	ok - I give up - going off to try commons/log4j
05:35	dysinger	They practically cut & pasted doug lea's threading library but couldn't see adopting log4j
05:35	pjstadig	dysinger: does clojure.contrib.javalog not work for you?
05:35	dysinger	whoa - didn't see that - thanks
05:35	dysinger	I don't know contrib well enough yet.
05:35	pjstadig	javalog.clj -- convenient access to java.util.logging in Clojure
05:35	pjstadig	:)
05:35	dysinger	sweetness
05:39	dysinger	hmm only problem is the big warning "Consider it DEPRECATED."
05:39	pjstadig	,(doc clojure.contrib.javalog/logger)
05:39	clojurebot	java.lang.Exception: Unable to resolve var: clojure.contrib.javalog/logger in this context
05:39	pjstadig	oh
05:39	pjstadig	hehe
05:39	pjstadig	i think something like it might still be useful
05:41	pjstadig	,(require 'clojure.contrib.javalog)
05:41	clojurebot	java.io.FileNotFoundException: Could not locate clojure/contrib/javalog__init.class or clojure/contrib/javalog.clj on classpath:
05:41	pjstadig	hiredman: i guess clojurebot doesn't have access to clojure.contrib?
05:41	Chouser	not all of it anyway
05:42	dysinger	maybe it's "This library will not work on Java 5." in javalog
05:43	pjstadig	,(System/getProperty "java.specification.version")
05:43	clojurebot	java.security.AccessControlException: access denied (java.util.PropertyPermission java.specification.version read)
05:43	pjstadig	hehe
05:43	pjstadig	clojurebot: why are you so cagey?
05:43	clojurebot	why not?
05:45	danlarkin	cooked
05:46	dysinger	commons/log4j api is much cleaner
05:46	dysinger	I never got the asinine Fine, finer & finest
05:47	dysinger	I actually complained to Sun when they released 1.4 about logging being so arcane compared to log4j
05:47	Chouser	and they said "good point, we'll get right on that"?
05:48	dysinger	heh
05:48	dysinger	I just wanted to log a complaint
06:06	pjstadig	clojurebot: what is our mascot?
06:06	clojurebot	the official mascot of clojure is a futuristic ninja robot
06:06	pjstadig	clojurebot: just checking
06:06	clojurebot	that's cool
06:07	pjstadig	clojurebot: futuristic ninja robot?
06:07	clojurebot	sooo much better than any of the historical ninja robots
06:20	Guest31655	clojurebot: historical ninja robots?
06:20	clojurebot	historical ninja robots are so 1999...
06:21	Neronus	clojurebot: 1999?
06:21	clojurebot	Gabh mo leithsc�al?
06:21	Neronus	clojurebot: 1999?
06:21	clojurebot	Excuse me?
06:21	Neronus	ahh
06:21	Neronus	clojure: histerical ninja robots?
06:21	Neronus	clojurebot: histerical ninja robots?
06:21	clojurebot	Excuse me?
06:21	Neronus	:-/
06:21	Neronus	clojurebot: You don't pass the turing test
06:21	clojurebot	I don't understand.
06:21	Neronus	clojurebot: yup, thats the problem
06:21	clojurebot	Titim gan �ir� ort.
06:22	Neronus	clojurebot: Wanna talk to a lady I know? She's called eliza
06:22	clojurebot	I don't understand.
06:28	pjstadig	clojurebot: 1999 is <reply>I'm afraid that was before I was born
06:29	clojurebot	Alles klar
06:29	pjstadig	~1999
06:29	clojurebot	I'm afraid that was before I was born
06:29	pjstadig	clojurebot: turing test is <reply>the answer is 42
06:29	clojurebot	'Sea, mhuise.
07:18	danlarkin	please don't subject us to that :-/
12:37	Cark	def~ c.l.PersistentQueue
12:38	Cark	~def c.l.PersistentQueue
12:39	mihand	hello guys. does recur work with lazy sequences?
12:39	Chousuke	what do you mean?
12:40	mihand	i have something like this (lazy-cons x (recur ...
12:40	Chousuke	hmmh
12:40	mihand	i would expect the recur to happen lazyly
12:40	dakrone_hb	when using an exec macro to execute a program, is there a way to get around the limitation of not being allowed to next #( functions?
12:40	kotarak	mihand: call the function recursively instead of recur
12:40	mihand	but instead it loops forever
12:40	Chousuke	I think you need to use a recur... yeah.
12:40	dakrone_hb	s/next/nest/
12:40	Cark	,(conj (new clojure.lang.PersistentQueue) 1)
12:40	clojurebot	java.lang.IllegalArgumentException: No matching ctor found for class clojure.lang.PersistentQueue
12:41	dakrone_hb	ie: http://gist.github.com/91322
12:41	mihand	calling it recursivly works
12:41	Chousuke	,(clojure.lang.PersistentQueue. [1 2])
12:41	clojurebot	java.lang.IllegalArgumentException: No matching ctor found for class clojure.lang.PersistentQueue
12:41	kotarak	dakrone_hb: #() is not a fn replacement. You have to use fn for nesting.
12:41	mihand	but i get stack overflows
12:41	Chousuke	should you be using lazy-cons anyway
12:41	Chousuke	wasn't it done away with in the new release.
12:42	kotarak	mihand: you probably want a new release, with lazy-seq
12:42	Chousuke	though it seems the documentation hasn't been updated yet
12:42	dakrone_hb	kotarak, how to you use fn to do %1 equivalent, would #(doseq [group %1] (...)) == (fn [group %1] (...))?
12:42	dakrone_hb	in the example on the gist
12:42	kotarak	dakrone_hb: (fn [some-seq] (doseq [group some-seq] ...))
12:43	dakrone_hb	kotarak, okay, I'll give it a try, thanks!
12:43	Chousuke	there's no mention of the lazy stuff on the clojure.org front page
12:43	Cark	(conj (clojure.lang.PersistentQueue/EMPTY) :a)
12:43	Cark	,(conj (clojure.lang.PersistentQueue/EMPTY) :a)
12:43	clojurebot	(:a)
12:43	Chousuke	and the docs still talk about lazy-cons
12:44	Cark	hum is it safe as in futureproof to use persistentqueues ?
12:44	Chousuke	Should be. :/
12:44	dakrone_hb	kotarak, that works great, thanks again :)
12:45	kotarak	dakrone_hb: np :)
12:45	danlarkin	ARGH why does (TimeZone/getDefault) return a sun.util.calendar.ZoneInfo instead of a java.util.TimeZone like it says it will?
12:46	dakrone_hb	so fn defines a function, and #( is like a lambda?
12:46	dakrone_hb	I'm still unsure of the main difference between the two
12:46	kotarak	dakrone_hb: fn is like lambda and #() is for very short lambdas.
12:47	kotarak	dakrone_hb: #(foo %) is equivalent to (fn [x] (foo x))
12:47	Cark	,(doc seque)
12:47	clojurebot	"([s] [n-or-q s]); Creates a queued seq on another (presumably lazy) seq s. The queued seq will produce a concrete seq in the background, and can get up to n items ahead of the consumer. n-or-q can be an integer n buffer size, or an instance of java.util.concurrent BlockingQueue. Note that reading from a seque can block if the reader gets ahead of the producer."
12:47	dysinger	how come the clojure api docs say to use (ns foo.bar (:import (java.util Date Timer Random) (java.sql Connection Statement))) I don't see that that works. I have to use (ns foo.bar (import '(java.util Date Timer Random) '(java.sql Connection Statement)))
12:47	dakrone_hb	kotarak, so, why can you nest fn, but not #90?
12:47	dakrone_hb	*#()
12:48	Cark	~def seque
12:48	kotarak	dakrone_hb: because #() is only for very short lambdas. Something like #(instance? SomeClass %).
12:48	dakrone_hb	kotarak, okay, I'll remember that, thanks for the explaination
12:50	pjstadig	~function literals?
12:50	clojurebot	Function literals are cute, but overused
12:55	dysinger	anyone have a noob explaination for my question above on (:import vs (import in (ns
12:55	dysinger	?
12:56	Cark	humhum rhickey: that's why I haven't exposed PersistentQueue yet - I'm afraid of people building polling systems on top of refs
12:57	Cark	well i'm not polling, i only need a queue !
12:57	Chouser	Cark: yes, I think that's specifically the misuse he's trying to avoid
12:58	Chouser	go ahead and use the class directly get the empty one, and you're set from there on in.
12:58	Cark	ok well i'm afraid about it disapearing in a future version
12:58	kotarak	dysinger: the :import for without quoting is correct.
12:58	Cark	i might have to build my own because of this
12:58	Chouser	Cark: I wouldn't worry about that.
12:59	Chouser	Cark: it's more likely to be exposed properly in a future version than it is to go away, I think.
12:59	Cark	also there's a little problem with it's interface
12:59	Cark	when you pop from a collection, it should return both the poped value and the new collection
13:00	Chouser	Cark: forcing you to destructure?
13:00	Cark	as it is you have to do imperative style ugliness
13:00	Chouser	what??
13:01	Cark	first get the value, the return the new queue ...
13:01	Cark	then
13:02	Cark	and yes, forcing destructuring is what you have to do in the absence of multiple return values
13:03	Chouser	it's no more imperative than working with any other seqable. (first s) ... (next s)
13:03	Cark	ah you're right
13:04	Cark	ok well i'll use it ... if it disapear i'll blame it on you !
13:04	Cark	=P
13:05	Chouser	ok
13:05	Chouser	I can handle such blame.
13:06	hiredman	Cark: if you really want to write non-imperitive code, make your functions pointless
13:06	Cark	chouser : damn, you're supposed to be terrified at this point...
13:07	Cark	hiredman, i just don't like to have these (do froms all over the place
13:07	Chouser	hiredman: a lot of the code I write seems pretty pointless, at least when I look back on it a few days later.
13:07	hiredman	:P
13:07	Chouser	Cark: yeah, breaking other people's code just doesn't scare me much. ;-)
13:09	Cark	~chouser is ruthless about breaking other people's code
13:09	clojurebot	Ik begrijp
13:10	Chouser	I think I can back that up with facts, even.
13:11	dysinger	kotarak: "dysinger: the :import for without quoting is correct." that appears to be true from the clojure api but it doesn't work for me trying to import commons logging for example. If use (import '(org..... it works
13:12	dysinger	It doesn't seem clear and clean to me why not
13:12	kotarak	dysinger: I'm not sure, what happens, but (ns foo.bar (:import (some.package ClassA ClassB))) definitively works for me. Do you have an example?
13:13	dysinger	y let me pastie one up and make sure I am correct
13:13	Cark	there are (were?) somme inconsistencies with the diferrent part of the ns form
13:14	Cark	i found that always using vectors works
13:14	Cark	like (ns some.package (:use [blah] [troot]) (:import [some.package Blah]))
13:15	kotarak	Wasn't there some change? I think, the () form was the original one. But I vaguely remember Stephen writing about that on the group.
13:16	Chouser	afaik, no actual change was made.
13:16	Cark	ah could be ... anyways use the vector syntax ...always works !
13:34	dysinger	kotarak - I was in the wrong ns
13:34	dysinger	sorry
13:34	kotarak	np
13:34	dysinger	how do you guys define something like a logger that you want to stick around ?
13:35	dysinger	I don't want to make-log for every debug call
13:35	dysinger	I was doing this (def log (LogFactory/getLog "x.y.z"))
13:37	dysinger	seems to work but I want to make sure I am following idiom
13:37	dysinger	idioms
13:38	dysinger	(defn log ..... ) would evaluate it every call right ?
13:39	Chouser	dysinger: yeah, what you've got seems fine.
13:39	dysinger	k
13:40	kotarak	Is there some ant magic, that it stops when c.l.Compile blows up?
13:45	cemerick	~max
13:45	clojurebot	max people is 164
13:45	cemerick	I daresay we've hit a plateau.
13:46	cemerick	who's got day-by-day population data for #clojure to make a pretty graph? ;-)
13:46	kotarak	Hmm.. We dettach from the google group. Should be 170 by now.
13:48	kotarak	When I have a src and a class file. The src in the filesystem, the class created from the source, but in a jar... Which one is loaded by Clojure?
13:49	cemerick	I think it always prefers the classfile.
13:49	cemerick	Wouldn't make sense to do otherwise.
13:49	kotarak	It loads the src if it's newer. But is this information available via jar?
13:50	cemerick	ah -- which reminds me -- why does the clojure-slim.jar file produced by the build retain the .clj files?
13:52	kotarak	The reason, why I'm asking is: I work on a ivy setup for contrib, which generates a jar per module on which I can depend (I don't want a 2.5MB jar, when I just wanted to use defvar...). But it seems, that the .class files provided in a jar are ignored, when the source is available via the filesystem. (I didn't want to rework the whole layout. So every module sees also the source of the other modules..)
13:55	cemerick	I'm not clear on why the source is still available, if you've split contrib into one jar per lib...
13:57	kotarak	cemerick: In contrib there are also dependencies. So, eg. repl-ln depends on def. So it gets provided the freshly built c-c-def.jar, which contains the .class files for c.c.def. But the source is still in the usual src tree. So c/c/def.clj is still in the classpath for c.c.repl-ln. Hence the .class files from c-c-def.jar are ignored and c.c.def is compiled again. Injecting its .class files into the c-c-repl-ln.jar....
13:58	cemerick	hrm.
13:59	cemerick	that's dicey. libs that are dependencies of many other libs (def is a great example) really should just get pushed into core. Who doesn't use defvar-?
13:59	cemerick	that's an orthogonal problem, though.
14:00	cemerick	kotarak: why not compile everything in one shot to one dest dir, and then produce separate artifacts for each lib?
14:02	kotarak	cemerick: when I solve this problem, I also solved the initial one: I have to decide, which file goes to which jar. (Actually the separate build idea is from ivy. If there is another way... I'm happy to use it)
14:03	cemerick	There's absolutely nothing stopping you from compiling all of clojure-contrib in one shot, and then splitting it up into multiple artifacts. You can then define dependency relationships between them in ivy.
14:06	kotarak	cemerick: the problem is: how do I decide which file belongs where? I have to rely on some naming convention for the source files (and hence class files). Clojure itself doesn't care for such a convention. (eg. genclass.clj)
14:07	cemerick	there's N libs in clojure-contrib, each one in clojure.contrib.XXX. Why not one artifact per lib package?
14:09	kotarak	cemerick: yes. That was roughly my plan (maybe grouping the scattered math stuff). Hmmm... The names in contrib are relatively consistent. I will have a look in the file name heuristic for splitting the jars.
14:11	dysinger	pulling my hair out - figured out the problem
14:11	dysinger	my mac's java vm is 6 (set in the prefs panel)
14:11	dysinger	but Aquamacs / Slime is loading 5
14:11	dysinger	:/
14:12	dysinger	user=> (System/getProperty "java.vm.version")
14:12	dysinger	"1.5.0_16-133"
14:12	dysinger	anyone else have this problem ?
14:14	cemerick	nope
14:15	Cark	did you manage to make eclipse work on an x64 computer ?
14:17	dysinger	Eclipse loads with JDK 1.5 on mac but recognizes 1.6 is on the system
14:17	dysinger	(for me)
14:40	durka42	kotarak: bah, my repl isn't working anymore
14:40	kotarak	durka42: huh?
14:40	durka42	\sr isn't bound to anything
14:41	kotarak	hmmm.. which rev are you using?
14:42	durka42	um, i just did hg pull/hg update bleeding-edge
14:42	durka42	vimclojure#NailgunClient is correct
14:43	kotarak	what does b:vimclojure_namespace say?
14:43	durka42	invalid expression
14:43	durka42	:echo b:vimclojure_namespace
14:43	kotarak	Have you started the ng-server? (<- stupid questions first)
14:43	durka42	aah! no i haven't
14:44	durka42	stupid people first
14:44	kotarak	Just start the server and do a :a
14:44	kotarak	I mean :e
14:44	kotarak	This should correctly setup the buffer.
14:45	durka42	you fixed the error conditions then, when there is no server to be found
14:45	durka42	uh oh
14:45	kotarak	It basically just goes on without interactive features.
14:45	kotarak	Starting the server and doing :e will activate them.
14:46	lisppaste8	durka42 pasted "untitled" at http://paste.lisp.org/display/78190
14:47	kotarak	What did you do?
14:47	durka42	i opened the file
14:47	durka42	that also happened when i did :e
14:47	durka42	the server is now running
14:47	kotarak	hmmm... What is gismu?
14:47	durka42	it is the namespace of the file
14:48	durka42	and the filename
14:48	kotarak	You did no \ef or something=
14:48	kotarak	?
14:48	durka42	no
14:48	durka42	i mean, it's true that it isn't on the classpath
14:49	kotarak	Arg. That gives always problems. You have to work with a proper setup.
14:49	kotarak	VC looks into the namespace to determine the syntax highlighting of included libs.
14:50	kotarak	Therefor the namespace must be available.
14:50	durka42	i can open other files
14:51	durka42	wait, it should be on the classpath
14:51	durka42	hmm
14:55	durka42	strange
14:55	durka42	the file starting with "(ns org.durka.nytimes.common" is loaded fine
14:55	durka42	"(ns org.durka.gismu" does not work
14:55	kotarak	what is the complete ns-form of the failing one?
14:56	kotarak	Ah no. Shouldn't of importance.
14:56	kotarak	VC loads via normal require...
14:58	durka42	ah, now it works
14:59	durka42	i restarted a few things after putting it on the classpath
15:00	durka42	so files with compilation errors do not load
15:00	durka42	this may or may not be desired behavior
15:01	kotarak	hmm... It is certainly suboptimal. But as long as VC uses the normal Clojure reader, one has to live with the need for proper source files...
15:01	kotarak	VC heavily relies on the introspection capas of Clojure.....
15:02	durka42	right
15:03	kotarak	Helping it a bit can get you very far. :)
15:08	Neronus	wahh... namespace dependency hell
15:39	eyeris	Is (for [a as b bs] [a b]) supposed to generate the same combinations as (for [a as] (for [b bs] [a b]))?
15:39	Chousuke	probably.
15:40	Chousuke	,(for [a [1 2] b [3 4]] [a b])
15:40	clojurebot	([1 3] [1 4] [2 3] [2 4])
15:40	Chouser	no, 'for' flattens things out, returns a single seq
15:40	eyeris	It will be wrapped in an extra list.
15:40	Chouser	if you nest the 'for' yourself, you'll get nested seqs
15:40	eyeris	But the combinations seem the same when I use numbers
15:41	Chousuke	hm, right.
15:41	eyeris	However I have other code that seems to behave differently.
15:41	eyeris	brb
16:01	eyeris	I'm trying to construct a crosstab from aggregate database results, but I'm having trouble with (for)
16:01	eyeris	http://pastebin.ca/1385227
16:01	eyeris	That paste shows my code and the problem I am having.
16:03	eyeris	I am essentially trying to take two lists and generate a list with the same number of elements as the first list, pairing items from the first and second list where a predicate matches and pairing the item from the first list with nil if there is no match.
16:03	eyeris	What I am getting is a list of all combinations of the items in each list.
16:04	eyeris	Do I just need to replace the second (for) with a searching function?
16:04	kotarak	(map #(if (pred %1 %2) [%1 %2] [%1 nil]) list-1 list-2) ?
16:05	eyeris	kotarak (map) requires equal-length lists.
16:06	kotarak	,(map #(vector %1 %2) [1 2 3] (iterate inc 5))
16:06	clojurebot	([1 5] [2 6] [3 7])
16:07	cads	,(zip [1 2 3] [1 2 10])
16:07	clojurebot	java.lang.Exception: Unable to resolve symbol: zip in this context
16:08	kotarak	,(interleave [1 2 3] [1 2 10])
16:08	clojurebot	(1 1 2 2 3 10)
16:08	cads	,(interleave [1 2 3] [1 2 10 6 6 6])
16:08	clojurebot	(1 1 2 2 3 10)
16:09	cads	nice!
16:09	cads	can we easily map a function over pairs in an interleaved list?
16:09	eyeris	,(into (1 2 3) (10 11 12))
16:09	clojurebot	java.lang.ClassCastException: java.lang.Integer cannot be cast to clojure.lang.IFn
16:09	eyeris	,(into [1 2 3] [10 11 12])
16:09	clojurebot	[1 2 3 10 11 12]
16:10	kotarak	cads: you mean like partition?
16:10	slashus2	,(zipmap [1 2 3 4] [4 5 6 7])
16:10	clojurebot	{4 7, 3 6, 2 5, 1 4}
16:10	kotarak	(map identity (partition 2 (interleave [1 2 3] [:a :b :c])))
16:10	kotarak	,(map identity (partition 2 (interleave [1 2 3] [:a :b :c])))
16:10	clojurebot	((1 :a) (2 :b) (3 :c))
16:11	Chouser	,(map vector [1 2 3] [:a :b :c])
16:11	clojurebot	([1 :a] [2 :b] [3 :c])
16:11	Chouser	,(map list [1 2 3] [:a :b :c])
16:11	clojurebot	((1 :a) (2 :b) (3 :c))
16:11	eyeris	,(map vector [1 2 3] [:a :b :c :d])
16:11	clojurebot	([1 :a] [2 :b] [3 :c])
16:11	cads	:)
16:11	eyeris	,(map identity (partition 2 (interleave [1 2 3] [:a :b :c :d])))
16:11	clojurebot	((1 :a) (2 :b) (3 :c))
16:11	kotarak	eyeris: seems that map doesn't require equal-length lists ;)
16:12	slashus2	,(vec (zipmap [1 2 3] [:a :b :c]))
16:12	clojurebot	[[3 :c] [2 :b] [1 :a]]
16:12	eyeris	kotarak It doesn't require them to execute, but it only consumes the smaller of the number of entries in each list
16:14	cads	,(let (zipmap2 #(map %1 (vector %2 %3))) (zipmap2 identity [1 2 3 4] [:dog :cat :bat]))
16:14	clojurebot	java.lang.IllegalArgumentException: let requires a vector for its binding
16:15	cads	,(let [zipmap2 #(map %1 (vector %2 %3))] (zipmap2 identity [1 2 3 4] [:dog :cat :bat])) ; like that?
16:15	clojurebot	([1 2 3 4] [:dog :cat :bat])
16:16	cads	hrm, I missed what Chousuke did
16:17	eyeris	cads: Here's a fleshed out example: http://pastebin.ca/1385227
16:18	eyeris	I don't even need the items paired, if the mismatches are eliminated.
16:22	eevar_	is this still true? -- When you ask a sequence on its last element for the rest it returns nil, saying, "there is no more".
16:23	slashus2	,(rest [3])
16:23	clojurebot	()
16:23	eevar_	http://clojure.org/lisps
16:24	Chouser	,(next [3])
16:24	clojurebot	nil
16:24	slashus2	,(next [3])
16:24	clojurebot	nil
16:25	eevar_	that whole paragraph seems pre-lazy-patch to me
16:26	Chouser	eevar_: yes it does
16:27	eyeris	Is there a function in the api like filter, but returning the first match instead of all matches?
16:28	eyeris	,(first [])
16:28	clojurebot	nil
16:28	slashus2	some
16:28	slashus2	,(some even? [1 3 4])
16:28	clojurebot	true
16:29	slashus2	await.
16:29	eyeris	Thanks
16:30	slashus2	No, some doesn't do that.
16:30	slashus2	You could just use first on filter.
16:30	eyeris	(first (filter even? [1 3 4]))
16:30	eyeris	,(first (filter even? [1 3 4]))
16:30	clojurebot	4
16:31	eyeris	,(first (filter even? [1 3 5]))
16:31	clojurebot	nil
16:31	eyeris	Perfect
16:31	kotarak	Ins't there a find-first or seek or something in contrib.seq-utils?
16:31	AWizzArd	there is
16:31	AWizzArd	(find-first even? [1 3 5])
16:31	Chouser	eyeris: (defn left-join [as bs pred] (for [a as] [a (first (filter #(pred a %) bs))]))
16:32	eyeris	Chouser: That is what I was just trying! :)
16:32	eyeris	You beat me to it :)
16:32	Chouser	ah, sorry.
16:33	eyeris	Earlier I sd "Do I just need to replace the second (for) with a searching function?", but I didn't know what that "searching" function was :)
16:33	eyeris	I knew it had to be simple
16:34	Chouser	by the time I understood what you were trying to produce, I had the solution
16:34	Chouser	since you're working with collections of maps and have names like "left-join", you might be interested in clojure.set
16:35	eyeris	Docs? I don't see it on the libraries page.
16:35	slashus2	http://clojure.org/api#toc649
16:37	Chouser	things named like 'xrel' should be sets of maps
16:37	bradford	anyone using clojure-test-mode?
16:39	cemerick	is there a map fn across associatives? Or is (apply hash-map (mapcat ...)) the best there is?
16:40	kotarak	(into (empty thing) (map foo thing))?
16:42	cemerick	yeah, somewhat better.
16:42	bradford	re: clojure-test-mode...I am able to run tests, but unsure what buffer the output is being dumped to - any ideas?
16:42	cemerick	heh, I always forget about into
16:42	kotarak	cemerick: me too. :)
16:43	cemerick	I think part of is that I intuitively think of setting all of the values in a map being cheaper when "updating" a map that already has all of the keys I want.
16:44	cemerick	...which is surely wrong in the new clojure world order :-)
16:49	Chouser	no, I think you're right.
16:50	Chouser	vectors and (array-)maps are created more efficiently when given the full set of values at creation
16:52	cemerick	well, in that way, into and (apply hash-map ...) are in the same boat, aren't they (e.g. they're just reducing over the seq of kvs being provided)
16:53	slashus2	It would be neat if the map constructors could also take in a map.
16:53	slashus2	(sorted-map map-of-another-kind)
16:54	antifuchs	how would the resulting sorted map be ordered? (:
16:54	slashus2	by the keys
16:55	Chouser	cemerick: for hash-map, yes. But (apply array-map ...) can be O(1) instead of O(n)
16:55	cemerick	slashus2: (into (sorted-map) map-of-another-kind)?
16:56	antifuchs	makes as much sense as any other (:
16:56	slashus2	cemerick: Yes that does work.
16:56	slashus2	When I first saw (sorted-map) I expected it to take a map as an argument to coerce it.
16:57	cemerick	Chouser: ah. That'd get coerced up to a hash-map pretty quickly, though.
16:57	Chouser	hm, and in practice you'd have to work pretty hard to make it O(1)
16:57	cemerick	yeah
16:58	cemerick	it's possible if your datasets are small, but in that case, you probably don't care about the costs of reducing through your new kvs anyway
16:58	slashus2	,(. clojure.lang.PersistentTreeMap (create {:b 5 :a 4}))
16:58	clojurebot	{:a 4, :b 5}
16:58	Chouser	even a LazilyPersistentVector will clone it's array before handing it to a PersistentArrayMap
16:58	slashus2	That will do it, but that is not how the constructor is set up.
16:59	Chouser	slashus2: that's some old-school syntax you've got there. ;-)
16:59	Chouser	,(clojure.lang.PersistentTreeMap/create {:b 5 :a 4})
16:59	clojurebot	{:a 4, :b 5}
16:59	slashus2	Chouser: I copied that out of the core.
16:59	Chouser	ah
16:59	slashus2	The core has some old school syntax :-P
17:00	cemerick	heh -- that's probably worth tweaking, so people can pick up "modern" idioms
17:00	slashus2	Chouser: Is that more efficient than doing into?
17:01	slashus2	into seems faster.
17:01	Chouser	both essentially do 'assoc' on all the values
17:01	slashus2	hmm
17:02	Chouser	on all the entries
17:02	slashus2	No, I was just doing a mini benchmark that doesn't show the variability of the times.
17:02	slashus2	I guess they are very similar.
17:02	pjstadig	,(class (into [] '(a b c)))
17:02	clojurebot	clojure.lang.PersistentVector
17:03	pjstadig	i thought into was lazy
17:03	Chouser	no
17:03	slashus2	Maybe it should have three explanation points on it.
17:04	slashus2	(into!!! (sorted-map) {:b 5 :a 4})
17:04	pjstadig	~def into
17:04	dakrone_hb	from: (exec "shellscript.sh" (fn [groups] groups)) I'm getting java.io.IOException: Stream closed, is this because of my syntax, or can I not just get a list of outputed lines like this?
17:04	Chouser	it produces non-lazy collections. no way for it to be lazy
17:05	pjstadig	yeah i wasn't sure why i thought that
17:05	pjstadig	i hadn't thought it all the way through
17:05	Chouser	dakrone_hb: what's 'exec'?
17:05	pjstadig	i just thought it was supposed to be efficient
17:05	dakrone_hb	Chouser, http://gist.github.com/91322 very first defmacro
17:07	Chouser	that macro looks bad -- it needs to force the lazy line-seq before returning out of with-open
17:07	dakrone_hb	Chouser, unfortunately, I don't completely understand how it works, found it elsewhere and needed something to run other programs
17:07	Chouser	or I suppose you have to be careful to force as much as you need in your 'pred'
17:07	dakrone_hb	clojure needs a popen method
17:07	Chouser	dakrone_hb: you might look at clojure.contrib.shell-out
17:08	dakrone_hb	Chouser, I will take a look at it
17:08	kotarak	It doesn't even need to be a macro...
17:09	Chouser	shell-out returns non-lazy things specifically to avoid that kind of problem
17:09	dakrone_hb	is there a unified place for clojure-contrib documentation, or do I need to go through all the source to understand it?
17:09	dakrone_hb	the google-code wiki only has 1 page in it
17:10	dakrone_hb	or, run (doc <blah>) on everything I guess
17:10	dakrone_hb	if there is a web-form
17:10	Chouser	there's not much -- most of the source files have a line or two in the header comment that describes the purpose of that lib
17:11	Chouser	most of them also have expamples of usage at the end, in a comment.
17:11	dakrone_hb	I see, seems like a javadoc equivalent would be nice to have for it
17:12	dakrone_hb	Chouser, thanks for the pointer, I'll check it out
17:22	duncanm	is there an equivalent to begin0 or prog1 in Clojure?
17:24	kotarak	(defmacro do1 [form & body] `(let [result# ~form] ~@body result#))
17:26	kotarak	or: (defn [form & _] form), but the macro is clearer in its intention, doesn't hide the main point (side-effects of _) and doesn't live from assumptions about the evaluation order of function arguments.
17:27	kotarak	So although not strictly necessary, I would prefer the macro in this case.
17:28	Chouser	interesting
17:28	slashus2	(defn returning [returnval & body] returnval)
17:29	slashus2	Does that work?
17:29	kotarak	yes, but for the above reasons I would not recommend this function.
17:31	slashus2	yeah
17:33	kotarak	But what are my recommendations... YMMV as always.
17:34	replaca	dakrone_hb: I've also been thinking about standarizing the doc on clojure.contrib so that folks kind understand it's capabilities, but I haven't started the discussion yet
17:34	dakrone_hb	replaca, that would definitely be helpful
17:34	dakrone_hb	should write a javadoc tool to go through and run (doc) on everything and convert it to html for publishing
17:35	replaca	I think so. We need to discuss and get some consensus on the best location, format, etc.
17:35	replaca	dakrone_hb: yeah, that's one component. Rich already does that for the core API
17:35	replaca	We should probably follow him on that for consistency
17:36	kotarak	There is already clj-doc, which is supposed to do that, IIRC.
17:36	dakrone_hb	one thing I'd love to see (as an extreme newbie) is non-trivial examples for all the core API functions in the docs
17:36	kotarak	Maybe it could serve as a starting point.
17:37	replaca	kotarak: yeah, we just need to automate it and host it
17:37	replaca	dakrone_hb: someone already did some of the in the wikibook. Did you check that out?
17:37	kotarak	dakrone_hb: http://en.wikibooks.org/wiki/Clojure_Programming/Examples/API_Examples
17:37	kotarak	clojurebot: examples is also http://en.wikibooks.org/wiki/Clojure_Programming/Examples/API_Examples
17:37	clojurebot	Ik begrijp
17:38	replaca	kotarak: thanks, saved me the trip!
17:38	dakrone_hb	kotarak, replaca, awesome! was reading through it and hadn't gotten there yet
17:38	replaca	as for contrib doc, I'd like to see it get a little richer than just API doc
17:39	kotarak	Namespaces may also carry docstrings. Some more verbose docs with examples could go there.
17:39	replaca	which implies the ability for authors to add their own html, markdown whatever
17:39	replaca	kotarak: so that did happen? I remembered it being discussed, but didn't know if it happened
17:39	replaca	that's also a great component
17:40	slashus2	So a "JavaDoc" generator for clojure?
17:40	kotarak	But it is currently broken in so far, that the metadata on namespaces is lost when loading from a class file.
17:40	kotarak	At least it was the last time I checked.
17:41	replaca	I'd imagine a home page with a list of contrib namespaces, the namespace doc and a link to the API doc (autogenerated) and also a link to some contributer/user generated doc if it exists
17:41	replaca	maybe hosted on the main clojure site off the contrib link.
17:42	replaca	kotarak: :-( I'll take a look and see if there's an issue on it
17:42	replaca	we're realy going to want that for tools like this
17:42	kotarak	replaca: I asked rhickey several times, whether I should open an issue, but never got an answer.
17:44	replaca	hmm, maybe we should just go ahead and do it :-) we could try to see if there's an easy patch
17:44	replaca	the desire for someone to do and maintain this system (which I'm willing to do) could serve as a motivator
17:45	replaca	I'll write up a post to the group (it's been on my mind anyway)
17:45	replaca	and see what people think
17:46	cp2	si there a built in function that returns the version of clojure?
17:46	Chouser	cp2: no
17:46	cp2	aye
17:47	Raynes	Write one.
17:47	replaca	Raynes: Can you determine it from the jar? I don't think so
17:47	slashus2	I don't think clojure keeps up with its current version.
17:48	cp2	Raynes: i wanted to see which version this clojure plugin downloaded
17:48	cp2	(la clojure for IDEA)
17:48	cp2	but im just going to build one from svn
17:48	replaca	It would have to be part of the compilation
17:48	kotarak	Up to now I had the impression, that the current version number doesn't keep up with the dev speed of clojure... ;)
17:48	replaca	I'd like to see a version and build number in there too, though
17:48	Raynes	cp2: I always use an external Clojure with La Clojure and Enclojure.
17:49	cp2	yeah Raynes, thats what i will be doing
17:49	replaca	most production systems that I've been involved with stash their svn version or something like it in the output somewhere
17:50	Chouser	http://clojure-log.n01se.net/date/2009-03-20.html#11:05b
17:50	cp2	heh
17:51	cp2	darn, revision 1338
17:51	cp2	i missed rev 1337
17:51	replaca	we discussed 1337 in here when it went by
17:52	replaca	that was a moment of high geekiness
18:03	dakrone_hb	Chouser, do you have a moment I could ask you some questions about your shell-out library?
18:07	cp2	does la clojure have the ability to spawn a repl?
18:08	cp2	Raynes: do you know ?
18:09	cp2	ah i see
18:09	cp2	nevermind
18:15	Raynes	cp2: I'm sorry for not replying sooner. La Clojure doesn't have a way to spawn a standalone REPL, or load files into an REPL, it will be included in the future. If you want to strike an REPL, the easiest way to do so is to create an empty file and then run it as a script and check run in REPL. Little tedious but it works.
18:16	cp2	yes i noticed :)
18:16	Raynes	The way it indents let and if forms, is what keeps me from using it any more than just testing it. I've talked with llya or whatever his name is about it and I guess he's going to include a settings page to fix it, but I'm not sure if we got the messages through to each other very well because of the language differences.
18:17	cp2	hm
18:17	cp2	i only just installed it, so i dont know about the way it indents
18:17	cp2	let me check that out
18:18	cp2	i see
18:18	cp2	yeah, i dont prefer my indentation to be like that either
18:18	Raynes	cp2: I think he thinks that is the convention :\
18:18	cp2	lol
18:19	cp2	i can tolerate that, but it would be nice to have it configurable
18:19	Raynes	I'm too picky to tolerate that.
18:44	Chouser	dakrone_hb: sure
18:44	dakrone_hb	so I was able to split command output by line with this: (doseq [i (.split (:out (sh "ls" "-l" :return-map true)) "\n")] (println "line:" i)) Is this the right way to do that?
18:44	Chouser	sure
18:46	dakrone_hb	and now I'm trying to write a wrapper or a macro around that so I can do something like (doseq [i (sh-lines "ls" "-l")] (println "line:" i))
18:46	dakrone_hb	because I'm going to be calling a lot of external programs and iterate over the lines of output
18:47	Chouser	sure, a function like that should be pretty easy to write.
18:47	dakrone_hb	I'm trying to wrap my head around writing it
18:48	dakrone_hb	and I have (defn sh-lines [& body] (.split (:out (sh body :return-map true)) "\n"))
18:48	dakrone_hb	but that doesn't work, something with the arguments I believe
18:48	dakrone_hb	am I on the right track? wrong track?
18:49	Chouser	looks pretty close
18:49	Chouser	try (apply sh :return-map true body)
18:50	dakrone_hb	Chouser, excellent!
18:50	dakrone_hb	thanks works :)
18:51	Chouser	great. you see why you need apply?
18:51	dakrone_hb	Chouser, I think so, is it because body is a list of arguments?
18:51	Chouser	right, body is a list but sh is expecting several args instead of a single list
18:51	dakrone_hb	and apply formats it correctly for the (sh) function?
18:52	dakrone_hb	how does apply know how to apply ':return-map true' to 'sh'?
18:52	dakrone_hb	*know not to
18:52	slashus2	,(doc apply)
18:52	clojurebot	"([f args* argseq]); Applies fn f to the argument list formed by prepending args to argseq."
18:52	Chouser	apply take then fn, then several args (or none), and finally a seq -- the seq is "unrolled" into the args to the fn
18:53	dakrone_hb	okay, so it applies to the last arg, no matter how many other arguments you have
18:53	Chouser	right
18:53	dakrone_hb	I understand now :) thanks
18:53	Chouser	which is why the body arg has to go last
18:53	dakrone_hb	okay
18:53	Chouser	fortunately sh doesn't demand that the options come at the end
18:54	dakrone_hb	if it did, would this not work?
18:54	Chouser	right, you'd have to do something more complicated
18:54	Chouser	(apply sh (concat body [:return-map true])) or something
18:55	dakrone_hb	ahh, to add the option only to the end first
18:55	dakrone_hb	thanks for the shell-out library, much better than the macro I was using :)
18:56	Chouser	you're welcome -- I only started it, several others have added on now.
18:58	Chouser	you might want to be careful spliting on "\n" if you intend to work on other OS's
18:58	dakrone_hb	Okay, I'll switch to a separator that'll work with others
18:58	dakrone_hb	thanks for the heads up
18:59	Chouser	could be another option, I suppose
18:59	dakrone_hb	for 'sh'?
18:59	Chouser	:out :line-vec
19:00	dakrone_hb	ahh, to split by lines automatically?
19:00	Chouser	right
19:00	dakrone_hb	that would be useful
19:07	Chouser	would there be any value to you if it returned a lazy sequence?
19:32	hiredman	lisppaste8: url?
19:32	lisppaste8	To use the lisppaste bot, visit http://paste.lisp.org/new/clojure and enter your paste.
19:32	lisppaste8	hiredman pasted "fast-proxy" at http://paste.lisp.org/display/78214
19:34	Cark	the regular proxy is slow ?
19:34	lisppaste8	hiredman annotated #78214 "usage example" at http://paste.lisp.org/display/78214#1
19:34	hiredman	no
19:34	hiredman	fast in the writing sense
19:35	Cark	kk
19:35	Chouser	the first arg to the given fn is the name of the method being called?
19:35	hiredman	as a keyword
19:36	durka42	so the only function you give fast-proxy is a dispatch?
19:36	hiredman	so you can use it with a multimethod
19:36	hiredman	no, it is a function
19:36	hiredman	if it is a multimethod it will need it's own dispatch function
19:37	hiredman	something like (comp first list)
19:38	hiredman	(defn fn->al [func] (fast-proxy ActionListener (fn [_ event] (func event))))
19:50	durka42	why did you use _SOMEFUNCTION instead of a gensym?
19:50	clojurebot	why not?
19:50	durka42	good answer
19:51	hiredman	different symbols got generated
19:51	hiredman	(let* [fn__688__auto__ (constantly 1)] (clojure.core/proxy [clojure.lang.IDeref] [] (deref [& args] (clojure.core/apply fn__687__auto__ :deref args))))
19:52	durka42	interesting
19:55	Chouser	that's because auto-gensym only works right within a single syntax-quote
19:55	Chouser	you'd have to use (gensym) manually
19:56	hiredman	in the outer let?
19:57	Chouser	sure, next to 'class'
19:59	lisppaste8	hiredman annotated #78214 "fixed to use gensym" at http://paste.lisp.org/display/78214#2
20:13	lisppaste8	hiredman annotated #78214 "filter out methods from ancestors" at http://paste.lisp.org/display/78214#3
20:15	durka42	is that what getDeclaredMethods does?
20:15	hiredman	is it?
20:16	hiredman	~javadoc java.lang.Class
20:17	cp2	getDeclaredMethods returns all methods regardless of access flags
20:17	cp2	whereas getMethods only returns public methods
20:18	durka42	i mean, do you want this method to filter out superclass methods
20:18	durka42	this macro, i mean
20:18	hiredman	yes
20:19	durka42	what if you're proxying something that descends from jframe, say
20:19	hiredman	I don't want to proxy wait() etc
20:19	hiredman	fast-proxy is for simple things
20:19	hiredman	at least that is the idea
20:20	durka42	ok
20:21	hiredman	I guess I could just filter out methods from Object
20:22	lisppaste8	hiredman annotated #78214 "filter out methods from object" at http://paste.lisp.org/display/78214#4
20:29	hiredman	~ticker IDSOX
20:29	clojurebot	IDSOX; -0.30
20:29	hiredman	my first day of my first index fund
20:33	cp2	lol
20:34	Cark	~google IDSOX
20:34	clojurebot	First, out of 167 results is:
20:34	clojurebot	IDSOX: Profile for ING DIRECT INDEX PLUS SMALL CAP - Yahoo! Finance
20:34	clojurebot	http://finance.yahoo.com/q/pr?s=idsox
20:58	jonathanturner	newbie questions... I'm trying to get numbers from 1 to 1000 using this: (filter (partial < 1000) (iterate inc 1)) but that doesn't seem to stop at 1000
20:59	hiredman	you want take-while
20:59	durka42	(partial < 1000) is #(< 1000 %)
20:59	hiredman	durka42: shhh
20:59	hiredman	clojurebot: function literals
20:59	clojurebot	Function literals are cute, but overused
21:00	durka42	no, i was giving a hint as to why it didn't work
21:00	hiredman	filter has no way of knowing that sometime after 1000 there isn't something that is less then 1000
21:00	durka42	that too
21:00	hiredman	clojurebot: filter?
21:00	clojurebot	filter is not map
21:01	durka42	clojurebot: map?
21:01	clojurebot	map is LAZY
21:01	hiredman	clojurebot: literal [?]
21:01	clojurebot	1
21:01	durka42	clojurebot: lazy?
21:01	clojurebot	lazy is hard
21:01	jonathanturner	hiredman: I thought that too, but then why would it return the values if they didn't meet the requirement?
21:01	hiredman	clojurebot: literal [?] filter
21:01	clojurebot	1
21:01	durka42	,(take 3 (filter (partial < 1000) (iterate inc 1)))
21:01	clojurebot	(1001 1002 1003)
21:01	hiredman	clojurebot: filter is also <reply>filter doesn't stop
21:01	clojurebot	c'est bon!
21:01	hiredman	jonathanturner: you are using the wrong function
21:02	hiredman	,(doc <)
21:02	clojurebot	"([x] [x y] [x y & more]); Returns non-nil if nums are in monotonically increasing order, otherwise false."
21:02	hiredman	,(< 1000 1)
21:02	clojurebot	false
21:02	jonathanturner	huh, why is that 1001 1002 1003 instead of 1 2 3?
21:02	hiredman	,(> 1000 1)
21:02	clojurebot	true
21:02	hiredman	filter is jumping over 1-1000
21:02	jonathanturner	aahhh, oops
21:03	jonathanturner	I had it in my head that filter was drawing out what you wanted, not removing it
21:03	hiredman	,(doc take-filter)
21:03	clojurebot	java.lang.Exception: Unable to resolve var: take-filter in this context
21:03	hiredman	er
21:03	hiredman	,(doc take-while)
21:03	clojurebot	"([pred coll]); Returns a lazy sequence of successive items from coll while (pred item) returns true. pred must be free of side-effects."
21:03	jonathanturner	(take-while (partial < 1000) (iterate inc 1)) returns nothing
21:04	jonathanturner	or () rather
21:05	hiredman	wrong function again
21:05	hiredman	,(take-while (partial > 1000) (iterate inc 0))
21:05	clojurebot	(0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 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 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 1
21:06	cmvkk	i this case, i actually think #(< % 1000) is more readable
21:07	jonathanturner	hiredman: I'm scratching my head as to why you say greater than instead of less than
21:07	cmvkk	because you're taking anything that 1000 is greater than.
21:07	dnolen	jonathanturner > is the function you setting it's first argument to 1000
21:07	dnolen	the second argument will be the number passed to take-while to test.
21:08	jonathanturner	dnolen: ahh, I think too many years of infix has melted my brain
21:08	hiredman	jonathanturner: ,(doc >)
21:08	hiredman	,(doc >)
21:08	clojurebot	"([x] [x y] [x y & more]); Returns non-nil if nums are in monotonically decreasing order, otherwise false."
21:08	cmvkk	this is why #(< % 1000) makes more sense. partial makes it look wrong...
21:09	jonathanturner	monotonically decreasing order? heh, glad I didn't look it up first
21:09	jonathanturner	cmvkk: yeah, I see your point
21:09	hiredman	,(doc <)
21:09	clojurebot	"([x] [x y] [x y & more]); Returns non-nil if nums are in monotonically increasing order, otherwise false."
21:10	hiredman	~def <
21:16	jonathanturner	I get it. So I had filter right in my head, it does draw out when things are true. Not that I was using it right, but had the idea
21:16	cmvkk	yes.
21:17	jonathanturner	though I guess you guys are prob tired of answering questions like that... I bet a decent number of people try the project euler stuff in clojure
21:17	slashus2	If an agent blocks, do the agents queued after it run?
21:17	slashus2	while it is blocked.
21:18	hiredman	depends
21:18	Chouser	if you used send-off, yes
21:18	hiredman	^-
21:18	cmvkk	if you send-off a function to an agent, nothing gets queued after it though, right?
21:19	slashus2	I am playing with a socket server, and I have the agent blocking and waiting for imput from the socket client, but the behavior seems to be that both clients have to enter input before the agents return anything.
21:19	cmvkk	unless it's another function to the same agent, in which case it would not run.
21:19	slashus2	cmvkk: That is the case that I am playing with.
21:19	slashus2	I think I just need to use threads.
21:19	cmvkk	if you send two functions to the same agent, then they won't run at the same time, i don't think...
21:19	slashus2	cmvkk: I think you are right.
21:19	cmvkk	(i don't actually know though)
21:19	hiredman	that is correct
21:20	slashus2	It is an agent queue.
21:20	hiredman	,(map #(.getName %) (.getMethods (class (agent nil))))
21:20	clojurebot	("releasePendingSends" "getErrors" "clearErrors" "getQueueCount" "shutdown" "dispatch" "deref" "setValidator" "getValidator" "notifyWatches" "getWatches" "addWatch" "removeWatch" "alterMeta" "resetMeta" "meta" "wait" "wait" "wait" "hashCode" "getClass" "equals" "toString" "notify" "notifyAll")
21:21	cmvkk	is there a way to use a Ref from java without using a class that implements Callable?
21:22	hiredman	~def c.l.Ref
21:22	durka42	fail
21:24	cmvkk	dosync runs Ref.runInTransaction, with the body as a callable fn, i think
21:24	Chouser	oh, by "use" you mean "mutate"?
21:24	cmvkk	yes.
21:25	cmvkk	i think foo.deref() is good enough to read
21:25	Chouser	yes
21:25	cmvkk	but ideally there would be start and stop methods for transactions or something
21:26	hiredman	runInTransaction doesn't seem to be a method on Ref
21:26	Chouser	well, it has to be able to restart the transaction
21:26	cmvkk	oh right, it's in LockingTransaction
21:26	cmvkk	Chouser, oh you're right. that's a good point.
21:26	cmvkk	so i really do have to encapsulate transaction code in its own class.
21:27	Chouser	why are you writing Java code?
21:27	Chouser	gun to your head?
21:27	hiredman	annon inner class
21:27	cmvkk	to learn java :)
21:28	cmvkk	i don't mind java, as long as I can do stuff that needs to be dynamic in clojure. which i can.
21:28	cmvkk	but i'm writing a 2D game engine, and i want it to work without me having to worry about boxing of numbers between function calls, etc
21:29	jonathanturner	what's lazy-cons in the most recent clojure?
21:29	cmvkk	lazy-seq ?
21:29	cp2	jonathanturner: cons
21:29	jonathanturner	I notice it's still on the website under Sequences but not in the api docs
21:30	cp2	cons is now lazy, so no more lazy-cons
21:30	hiredman	cp2: nope
21:30	hiredman	cons is not lazy
21:30	Chouser	lazy-cons is gone. use lazy-seq and cons in similar circumstances
21:30	cp2	oh what
21:30	cp2	oh right
21:30	hiredman	lazy-seq is lazy
21:30	cp2	Chouser got it right
21:30	cp2	i was mixed up
21:30	cmvkk	ha ha, someone didn't update the website properly
21:30	hiredman	you wrap cons in lazy-seq
21:31	cp2	right, listen to them jonathanturner
21:33	Chouser	see http://clojure.org/lazy
21:36	jonathanturner	Chouser: thx
21:42	cmvkk	ugh, i can't commute without passing an IFn either
21:43	Chouser	there's a reason rhickey created a new language
21:43	cmvkk	this stuff should be usable from java though,
21:43	cmvkk	i mean, having something like a Ref is useful in any language.
21:43	Chouser	it is usable, it's just not convenient
21:44	cmvkk	it just reminds me of some place where he was talking up the fact that the persistent data structures are all easily drop-in usable in java
21:45	cmvkk	if that's such a boon, then why not the STM?
21:45	cmvkk	oh well. i can do it this way anyway.
21:45	Chouser	all you'd need is macros in Java, and STM could be convenient
21:46	cmvkk	in this case, it's not something that's a restriction of the language, though.
21:46	msingh	did anyone find it weird getting used to square parens coming from lisp?
21:46	cmvkk	see, refs have a set(), which takes a new value and that's all.
21:46	cmvkk	but commuting has particular characteristics that are favorable (it's 'more concurrent')
21:47	Chouser	cmvkk: the fn passed to commute gets run twice
21:47	cmvkk	that is true...
21:48	dnolen	msingh: only a little, have built in reader support for defining vectors [], maps {}, and set #{} is worth the tradeoff, also it makes destructuring bind quite fun!
21:48	cmvkk	which brings up an interesting point. 90% of the time, all i need to do is commute the ref's value. if it were 100%, i could just use an agent or atom.
21:49	cmvkk	but 10% of the time, i need to change two refs in reference to each other
21:49	cmvkk	there should be a way to commute outside of a transaction. since i don't intend to do anything else inside the transaction but change that one value.
21:50	msingh	dnolen: humm.. may'be they'll grow on me :)
21:50	arohner	if commute were defined as (dosync (_commute ...)), wouldn't that work?
21:51	cmvkk	yeah. i guess the point was, in that situation, there's no need to run commute twice.
21:51	arohner	because if you wanted to do stuff in a "real" transaction, you would write (dosync (commute foo) (commute bar))
21:51	cmvkk	or rather, there's no need for the fn passed to be called twice.
21:52	Chouser	that's a very interesting observation
21:53	cmvkk	it would be like calling swap! on a ref.
21:53	Chouser	right
21:53	cmvkk	well, kind of. maybe more like 'send' but without the new thread.
21:54	cmvkk	anyway
21:54	Chouser	hm.
21:56	Chouser	no, I think it'd be synchronous, like an atom
21:58	cmvkk	oh yeah, i guess i was thinking that if it was commutable you wouldn't have to retry ever, but really you still would.
22:00	Chouser	I see in the code that retries can happen when committing commutes, but I think that's different from what I understood previously.
22:00	cmvkk	well if two changes are being made to a ref at once, it doesn't matter if they're commutable. only one can go through.
22:02	replaca	Q: does anyone know how your supposed to add metadata to a namespace (like a doc string)
22:02	Chouser	but I thought they queued up and happened without retrying
22:03	replaca	we were talking about this earlier, but I don't see how to do it
22:03	replaca	*your => you're!
22:04	cmvkk	Chouser, commutes queue? that would block the tread the commute occurred in though, right?
22:04	cmvkk	which i thought nothing particularly ever blocked.
22:04	cmvkk	the reason agents queue is because the code isn't being executed in the calling thread anyway.
22:11	replaca	oops, never mind - that was obvious
22:13	Chouser	I thought the commute happened once for the "in-transaction" value, and then queued for the commit value.
22:13	cmvkk	if you mean it waits until the transaction is committed to run again, then yeah.
22:14	cconstantine_	has anyone here used enclojure?
22:14	cmvkk	but what if, for example, the transaction finishes, then when the commute is running, another transaction happens?
22:15	slashus2	Since an agent can block the agent queue, send-off wouldn't help on an agent that blocked indefinitely, it would help on an agent that blocked, but not indefinitely.
22:15	cmvkk	send-off helps for agents that block indefinitely in that they don't hog one of the threads in the pool.
22:16	cmvkk	other than that, it makes no difference.
22:17	cmvkk	also, there's not just 'an agent queue'. every agent has its own queue... at least i think.
22:17	cmvkk	if you send-off a function that runs forever, to an agent, it won't affect the way any other agent runs.
22:17	Chouser	if an action is going to be slowed down by anything other than the CPU, it should be done in a send-off
22:18	Cark	i think there's a single queue for all agents
22:18	Cark	~def c.l.agent
22:18	Chouser	each agent has a queue of actions. also, the "send" pool has a queue of agent/actions
22:18	Cark	right
22:19	Cark	that's what they just changed in erlang
22:19	Cark	they were seeing contention because of this
22:19	replaca	Cark: did you get the pretty printer working OK?
22:19	cmvkk	well anything done with a send-off won't block the pool queue, right?
22:19	Cark	replaca : yes i did, thanks a lot
22:19	Chouser	cmvkk: not at the java level, right.
22:19	replaca	you were having problems with it yesterday, weren't you?
22:20	Cark	was only a matter of flushing out =/
22:20	Chouser	cmvkk: of course both pools are sharing the same hardware, so... :-)
22:20	replaca	cool, glad to hear it. I'll ge tthe upper stream to flush the underlying strem before closing
22:20	cmvkk	well right.
22:20	replaca	which should make that invisible to you
22:20	Cark	replaca : mhh i'm not sure it's the job of your library to do this
22:21	Cark	library which is highely usefull, thanks for it !
22:21	replaca	well, I'm creating a wrapper stream, if you didn't do it yourself, so I should probably flush on close
22:21	Cark	so you'll be flushing each time i use pprint ?
22:21	replaca	if people are doing larger things, they coul dopen a longer term stream
22:22	replaca	yeah, does that seem like a bad idea?
22:22	replaca	I've been torn on the subject, which is why I hadn't done it
22:22	Cark	it depends, can you forsee some inefficiencies coming out of this ?
22:22	replaca	(you're welcome, btw :-) )
22:23	Cark	i think only the user knows when it's best to flush
22:24	replaca	yeah, though there's a workaround: (binding [out (PrettyWriter. out)] ...) (more or less)
22:24	replaca	that's why I'm on the fence. The other ide of the story is that you have confusion.
22:25	replaca	well, maybe we'll see if more people are confused by it
22:25	Cark	well i have a feeling that most programs spend a little more time doing output than my little test, the flushing problem then becomes easy to spot
22:26	Cark	maybe give a hint about it in the documentation
22:26	replaca	good idea
22:27	replaca	Q: anyone ever think about the problem of "find all the namespaces in a jar?"
22:28	Chouser	I've asked around about "find all the classes in a package"
22:29	Chouser	but apparently the closest you can come to that is poking around if the filesystem or .jar files directly.
22:29	albino	jar -tf ?
22:29	mattrepl	unzip -l
22:30	hiredman	maybe clojure should have it's own equiv of the manifest?
22:30	hiredman	or embrace and extend the manifest
22:30	Chouser	what's missing from java's manifest?
22:30	mattrepl	you can grab all class files from a jar, if that's what you're thinking
22:31	Chouser	or all the .cljs
22:31	mattrepl	or that
22:31	Chouser	and look for (ns ...) at the top
22:31	hiredman	Chouser: a list of the namespaces in the jar
22:31	clojurebot	Chouser might make night
22:32	hiredman	clojurebot: bah!
22:32	clojurebot	I don't understand.
22:32	Chouser	clojurebot: Chouser is <reply>Who??
22:32	clojurebot	Chouser might make night
22:32	Chouser	what does that even mean?
22:32	Cark	~chouser?
22:32	clojurebot	chouser is ruthless about breaking other people's code
22:33	Cark	=)
22:33	hiredman	someone was thanking you for something, and it was late
22:33	Chouser	much better.
22:33	Cark	case sensitive
22:33	Chouser	ohhh.
22:36	Chouser	but why won't it let me overwrite?
22:36	Cark	oh is there only one definition per word ?
22:36	Cark	~foo
22:36	clojurebot	namespaces are (more or less, Chouser) java packages. they look like foo.bar; and corresponde to a directory foo/ containg a file bar.clj in your classpath. the namespace declaration in bar.clj would like like (ns foo.bar). Do not try to use single segment namespaces. a single segment namespace is a namespace without a period in it
22:36	hiredman	it thinks lines that end with ? are questions
22:37	hiredman	clojurebot: literal [?] Chouser
22:37	clojurebot	1
22:37	hiredman	clojurebot: literal [1] Chouser
22:37	clojurebot	<reply>Who??
22:37	Chouser	race
22:37	hiredman	looks overwitten to me
22:38	Chouser	~Chouser?
22:38	clojurebot	Who??
22:38	Chouser	yep, you gave me the hint I needed, thanks.
22:56	replaca	sorry, I had to go put my boy to bed
22:56	slashus2	cmvkk: The reason I said that earlier is because I wrote a server thingy it had the following behavior.
22:57	replaca	I'm thinking there's gotta be a Java API just to read through the jar and then I can load up all the clj files and see what (all-ns) is like at the end
22:58	hiredman	oh sure
22:58	hiredman	there is a java api for jars
23:01	lisppaste8	slashus2 pasted "server thingy" at http://paste.lisp.org/display/78229
23:05	cmvkk	slashus2, maybe that's socket behavior?
23:05	slashus2	cmvkk: It works the way I expected when I use the executors.
23:06	cmvkk	i guess i don't see where you're using more than one agent, by the way.
23:06	slashus2	I am using one agent.
23:06	cmvkk	so, an agent can only do one thing at a time.
23:06	slashus2	Right.
23:06	cmvkk	so that would be why.
23:06	slashus2	send and send-off have the same behavior in this case?
23:06	cmvkk	what you should be doing is using a separate agent for each connection.
23:06	cmvkk	yes.
23:07	cmvkk	the thread pool can only do n things at once (where n is the number of threads), but
23:07	dreish_	send-off only promises not to deplete the common agent thread pool.
23:07	cmvkk	an agent will only take up one thread at any time.
23:07	cmvkk	whether that thread is part of the pool or not.
23:08	cmvkk	(.execute pool ...) is like creating a brand-new agent and using it only once.
23:08	slashus2	cmvkk: Yeah
23:08	cmvkk	i.e. (send (agent nil) client-handler-agt-fn client-socket)
23:11	hiredman	the is a threadpool (two of them) backing agents, and each agent has it's own queue of actions
23:12	cmvkk	i think it might be useful to think of an agent as its own thread with an attached value, even though at the java level that's not what's going on
23:13	hiredman	so send or send-off to a single agent multiple times will block until the previous action completes
23:13	slashus2	I was confused because it said that send-off should be used in situations with blocking actions.
23:13	hiredman	send's go to a bounded threadpool so they may also need to wait for other agents to complete
23:14	hiredman	send-off's go to an unbounded threadpool
23:14	dreish	slashus2: I agree the docs should do a better job of explaining _why_ the distinction between send and send-off exists. Otherwise one can't help but form theories.
23:15	cmvkk	oh yeah, the docs for send-off are ambiguous in that way.
23:45	arohner	,(str nil)
23:45	clojurebot	""