#clojure – 2010-04-29

2010-04-29

00:27	replaca	did I hear that clojure-swank is broken with the latest 1.2 clojure?
00:35	replaca	ah, the tumbleweeds are rolling around in here tonight...
01:30	palamas	Good evening all. Would someone have a moment to answer a question about labrepl? When I try to start it it throws "java.lang.ExceptionInInitializerError (control.clj:9)
01:32	leifw	technomancy\|away: I take issue with a commit of yours in clj-processing. If you come back and have a chance to look at it, I'd be elated.
01:33	unlink	How do I achieve the effect of apply when I want to apply to a Java method call? e.g. (.someMethod obj [expand me])
01:33	Chousuke	wrap it in a function
01:34	leifw	unlink: I remember hearing that memfn was bad for some reason, but I'd probably do (apply (memfn .someMethod obj) list)
01:34	Chousuke	leifw: I don't think memfn is bad as much as outdated (#() is easier most of the time)
01:35	Chousuke	and I think memfn doesn't need the . in the method name
01:35	Chousuke	(doc memfn)
01:35	clojurebot	"([name & args]); Expands into code that creates a fn that expects to be passed an object and any args and calls the named instance method on the object passing the args. Use when you want to treat a Java method as a first-class fn."
01:35	leifw	true
01:40	unlink	Chousuke: How would you wrap it in a function?
01:43	Chousuke	just (apply #(.someFoo %1 %2) [1 2])
01:43	jacortinas	test
01:43	jacortinas	woo
01:43	Chousuke	but then you need to know how many parameters there are
01:45	unlink	Chousuke: right, that's the issue I'm facing
01:45	Chousuke	then try memfn
01:45	unlink	It's a static function
01:46	Chousuke	hm
01:46	unlink	er
01:46	unlink	but memfn pokes a hole in the place of the object
01:47	Chousuke	,(apply (memfn Integer/valueOf) ["1"]) testing
01:47	clojurebot	java.lang.IllegalArgumentException: No matching method found: valueOf for class java.lang.String
01:47	Chousuke	right. :(
01:47	Chousuke	I suppose then you're left with manually writing a multi-arity fn
01:47	unlink	meh, I've never heard a better excuse to use a macro
01:48	unlink	um. false.
01:48	unlink	yeah I guess you're right.
01:49	Chousuke	You can write a macro if you're going to need to do this a lot
01:49	hoeck	unlink: or try reflection
01:49	unlink	but the number of args isn't known in advance
01:50	unlink	(defmacro wrap [method args] `(info (apply #(. EWrapperMsgGenerator ~method) ~args)))
01:50	unlink	obviously the apply #( ... ) doesn't work
01:50	Licenser_	morgn
01:50	unlink	but it will be used like this: (proxy [cls] [] (meth1 [& args] (wrap meth1 args)) (meth2 [& args] (wrap meth2 args) (meth3 [& args] (wrap meth3 args)))
01:51	hoeck	Licenser_: moring
01:51	unlink	close parens to taste :-)
01:51	Chousuke	unlink: that doesn't look right.
01:52	unlink	I would have liked to put a macro in there that would let me do (proxy [cls] [] (wrap meth1 [arg1 arg2 arg3]) (wrap meth2 [blah bluh bleh]))
01:52	unlink	Chousuke: Yeah it's a funky API
01:53	Chousuke	unlink: would the vectors always be literal like that?
01:53	unlink	Chousuke: in which case?
01:53	Chousuke	for your wrap invocations
01:53	unlink	Chousuke: those are supposed to be argument vectors
01:54	unlink	Chousuke: actually, I can accomplish this with a macro that wraps proxy.
01:54	unlink	...which I would really not like to do.
01:54	Chousuke	well, I'll take your word for it because I have no clue what you're trying to do
01:55	unlink	Chousuke: I'm implementing one interface which basically delegates its work to another object (which doesn't implement that interface)
01:55	unlink	s/object/class
01:56	Chousuke	once or many times?
01:56	unlink	many times, like 50
01:56	Chousuke	ie. do you have many class/class pairs doing this delegation?
01:56	unlink	oh, no
01:56	Chousuke	or just objects?
01:57	Chousuke	remember that proxy can close over locals, so if you have only one class you need to wrap, I think you could do it by returning a proxy from a function
01:57	unlink	I have an object with about 50 different methods which all get delegated to a class with methods of the same name
01:57	Chousuke	ah.
01:58	Chousuke	so that's why you want a macro .p
01:58	Chousuke	I think writing a macro that generates a proxy form would be the easiest
01:58	unlink	I promise I had nothing to do with the design of those APIs :-)
01:58	unlink	Yeah I can't think of anything else
01:59	Chousuke	(make-wrapping-proxy from-class to-class methodname [arg1 arg2] methodname2 [arg1 arg2] ...)
01:59	Chousuke	though I'm not sure how you'd handle the static/instance method difference in this case.
02:04	Chousuke	In any case, you should end up with something like (defmacro foo [a b & args] (let [partition-here] `(proxy [whatever] ~@(map (fn [[methodname argvec]] `(~methodname ~argvec (. foo ~methodname ~@argvec))))))
02:04	Chousuke	except with fewer metasyntactic variables :P
02:05	unlink	wow
02:05	unlink	that is almost identical to what I wrote
02:06	unlink	I used a real classname instead of whatever and foo, and meth and args instead of methdname and argvec, but otherwise identical
02:06	Chousuke	heh
02:13	unlink	except it's a little more complicated because of multiple arity methods
02:14	Chousuke	hm
02:14	Chousuke	well you can also try using reflection in the macro
02:15	Chousuke	(note: not in the generated code :))
02:15	unlink	haha
02:15	Chousuke	but that'll be a bit more difficult to write
02:16	Chousuke	but it's entirely possible as long as the class is known at compile-time
02:16	Chousuke	which it has to be, I suppose.
02:25	unlink	Wow, really? http://paste.plurk.com/show/242337/
02:30	vegai	uh oh, I've been voted to speak about clojure at work next friday
02:30	vegai	I guess it's high time to learn the thing :P
02:31	zmila	vegai: you speak about Clojure once, then second time, finally you learn it.
02:32	vegai	;)
02:38	Raynes	"Scala & Clojure. I think Scala is more similar to Ruby than Clojure. But if you ask me which I like, I’ll say Clojure." -- Matz
02:38	vegai	bogglesome.
02:39	vegai	I know I shouldn't say this, having designed no languages... but I don't have much respect for Matz
02:39	vegai	oops, I said it.
02:39	Raynes	At least Ruby isn't Java.
02:39	Raynes	Or C.
02:41	Raynes	vegai: Elaborate! :D
02:42	vegai	I've used ruby at work for about 3 years
02:45	leifw	I would probably use ruby if there were no lisps left
02:45	Raynes	leifw: I'd probably use Haskell. But once I got sick of monads, I'd go to Ruby.
02:46	vegai	while it's quite nice for writing code, I've been suffering a bit from its lack of design
02:46	Raynes	vegai: I'm not going to defend Ruby or anything. I was just curious. :>
02:46	leifw	Raynes: I've tried to figure out haskell so many times it's just getting depressing.
02:46	Raynes	leifw: I figured it out once.
02:46	vegai	Raynes: I'm not rabidly against it... we have couple of much much worse languages in use here :P
02:46	leifw	that must've been fun
02:47	Raynes	But I went to Clojure because I couldn't grok monads enough to find a use for them in my own code.
02:47	vegai	perhaps it's more rails that bites me. It doesn't scale.
02:47	Raynes	I understood how monads worked, but I didn't understand how they were useful outside of the general stdlib monads.
02:47	Raynes	I think I need to do a Clojure monad tutorial.
02:47	vegai	people like to blog how rails scales and everyone who says it doesn't are idiots
02:47	Raynes	cgrand: ohai
02:47	vegai	and I fear that readers believe that
02:48	Raynes	I don't care for Ruby because nobody seems to use it except for rails.
02:48	vegai	yeah, I was heavily into haskell for several years also. I guess I still am in a way
02:48	Raynes	._.
02:48	leifw	vegai: if you don't mind, where do you work?
02:48	Raynes	And I don't do webdev.
02:49	vegai	leifw: in a small company that makes billing systems for telcos and such
02:49	cgrand	good morning #clojure!
02:50	leifw	ugh...good late late late night, cgrand
02:51	Raynes	(defplugin (:hicgrand "Says hi to cgrand." ["hicgrand" "hiChristophe"] [{:keys [irc channel]}] (ircb/send-message irc channel "cgrand: "HAI!")))
02:51	Raynes	Aw, damn.
02:51	Raynes	Misplaced a quote.
02:51	Raynes	That would have been clever, otherwise.
02:52	cgrand	Raynes: tank you for the comment in respond.clj
02:52	Raynes	<3
03:04	replaca	morning cgrand!
03:50	LauJensen	Morning all
03:51	Raynes	LauJensen: Moin.
03:52	Raynes	That's used in Denmark, right?
03:52	LauJensen	Well How'dy Mr. Alabama :)
03:56	LauJensen	Raynes: Its used in the south tip of Jutland (Denmarks main piece of land), but its German in origin
03:56	LauJensen	btw, could we rename your bot to The-Richinator ?
03:56	Raynes	I don't think I can consent to that.
03:56	Raynes	._.
03:58	Raynes	Yesterday's exposure got me 4 new sexpbot watchers on Github. :D
03:58	LauJensen	Its a popular project?
03:59	Raynes	It's my most popular project. It's not popular though.
03:59	Raynes	I've just never had a popular project to compare it to.
04:00	LauJensen	Ok
04:29	LauJensen	hmm, where's philjure? still nighttime in the US I guess
04:34	spariev	it's suspiciously quiet here ... Raynes finally got some sleep ?
04:34	Raynes	>.>
04:34	Raynes	I slept from 11:00am yesterday until 9:00pm last night. It's 3:34am now.
04:37	spariev	I envy your schedule :)
04:38	Chousuke	LauJensen: I think you might be the only *jure proponent on this channel ;P
04:38	LauJensen	Chousuke: Im not though
04:38	LauJensen	I just hate communism and *jure in equal measure
04:39	Chousuke	wait, whaT?
04:39	Chousuke	where does communism come into this?
04:39	LauJensen	Chousuke: When my build tool has an oppinion about my naming strategy, thats communism, dont you recognize?
04:39	Chousuke	no
04:39	Chousuke	just, no,.
04:39	LauJensen	Then you're a commujure yourself
04:40	Chousuke	:P
04:40	Chousuke	you shouldn't joke about things like this, someone might take it seriously
04:40	Raynes	Let's just rename Clojure, rumblethunderchickenfinger, and nobody will ever want to use the name as a pun.
04:41	LauJensen	Raynes makes a good point
04:41	LauJensen	Chousuke: I'm not kidding - Unless you can pursuade me otherwise I think its down right communistic to restrict name in a build-tool because of personal taste
04:42	LauJensen	And I'll be happy to rantjure about it
04:42	Chousuke	LauJensen: I think you might have some really distorted ideas about communism
04:43	Chousuke	a build tool discouraging you from stupid naming schemes is simply opinionated, not communist :P
04:43	Chousuke	by your logic, clojure itself is communist because it's definitely opinionated. And that's just silly
04:44	LauJensen	No your logic is that oppinionated = communistic which I dont think. This is purely a matter of style and thats why I think its a bad idea
04:44	Chousuke	but I think bringing a political ideology of any kind to the discussion is a mistake to begin with
04:45	Chousuke	LauJensen: right, it's okay to think that explicitly discouraging bad naming schemes is a bad idea, but please don't call it communist.
04:45	Raynes	#politics
04:45	LauJensen	Chousuke: np, if it offends you or anybody else I wont
04:45	Chousuke	because that's just an association fallacy
04:46	Chousuke	LauJensen: You'll find many people are put off by faulty logic :)
04:46	Raynes	I'm renaming sexpbot botjure
04:47	LauJensen	Woohoo!
04:47	Chousuke	gah.
04:47	LauJensen	pwnagejure
04:47	vegai	Raynes: how about 'RAMIREZ'?
04:47	LauJensen	Chousuke: But now that we're beyond the name calling, where do you see the logic failing?
04:47	vegai	and force all commands te be uppercase.
04:47	Raynes	:>
04:48	Chousuke	LauJensen: the fallacy is simply associating an opinionated feature with a political ideology :P
04:48	Raynes	Nobody ever called anybody any names.
04:48	Raynes	At least, what I read of it.
04:48	LauJensen	Raynes: I called leinjure communistic
04:48	Raynes	Did you call Chousuke a commie?
04:49	Chousuke	LauJensen: I mean, it's just an easter egg. You can still name your project whatever you want.
04:49	Chousuke	LauJensen: leiningen just wont help you if you choose to show a lack of taste :P
04:50	LauJensen	Chousuke: Lemme just quote Wikipedia for you "Communism is a social structure in which classes are abolished and property is commonly controlled and where the ruler decides what you can name your projects"
04:50	Chousuke	:D
04:50	SynrG	#politics++
04:50	Chousuke	yeah, I think this is getting a bit out of hand
04:50	LauJensen	Alright alright, go back to your coffee and news sites :)
04:51	SynrG	coffee and clojure101 exercises
04:51	Raynes	Such arguments can only be solved with a fight to the death.
04:51	SynrG	no, to the pain!
04:51	Raynes	SynrG: New ones?
04:51	Chousuke	no. this must be solved by a robot battle
04:51	SynrG	Raynes: i'm very slow and deliberate :) also i only have scraps of time for it.
04:51	Raynes	SynrG: I did the last ones in my head in about 60 seconds. :>
04:51	Chousuke	and the robots must be controlled by an AI written in Clojure
04:52	SynrG	Raynes: ah, but you weren't a clojure virgin
04:52	Raynes	$google robocode
04:52	sexpbot	First out of 34900 results is:
04:52	sexpbot	Robocode Home
04:52	sexpbot	http://robocode.sourceforge.net/
04:52	sexpbot	First out of 26 results is:
04:52	sexpbot	Robocode - Wikipedia, the free encyclopedia
04:52	sexpbot	http://en.wikipedia.org/wiki/Robocode
04:52	Chousuke	Raynes: hm, that's a bit floody :P
04:52	Raynes	Huh
04:52	Raynes	It isn't normally.
04:52	Raynes	A small bug.
04:53	Raynes	I introduced a few when I tried to rewrite 26 plugins in an hour last night.
04:53	SynrG	Raynes: everything i read raises questions, and more likely as not takes me on a trip through doc and some experiments in repl, so ... i'm only just now on the last exercise as a result
04:55	LauJensen	$google define: floody
04:55	sexpbot	First out of results is:
04:55	sexpbot	First out of 3 results is:
04:55	sexpbot	Tully (Irish sept) - Wikipedia, the free encyclopedia
04:55	sexpbot	http://en.wikipedia.org/wiki/Tully_(Irish_sept)
04:55	LauJensen	hmmm :)
04:56	Raynes	Mismatched parens that weren't really mismatched.
04:56	Raynes	Just in the wrong place.
04:58	Raynes	$google wtf
04:58	sexpbot	First out of 4530000 results is:
04:58	sexpbot	The Daily WTF: Curious Perversions in Information Technology
04:58	sexpbot	http://thedailywtf.com/
04:58	Raynes	Fixed.
04:58	LauJensen	Raynes: is the first line necessary ?
04:59	Raynes	LauJensen: It's the exact same thing clojurebot does.
04:59	LauJensen	yes, but if nobody is reading it, why not just let it go ?
04:59	Raynes	Because I think it's generally useful.
04:59	LauJensen	ok
04:59	Chousuke	Raynes: you could combine the first and second lines
05:00	Raynes	That would just be horizontal spam instead of vertical.
05:00	Raynes	:p
05:00	Raynes	I'll do that later.
05:00	Raynes	Working on another branch right now.
05:00	Chousuke	excellent!
05:02	LauJensen	(chousuke is really Mr Burns btw)
05:38	stilkov	I can't seem to find a way to invoke set! on any kind of var
05:38	stilkov	I must be missing something plainly obvious
05:39	_ato	stilkov: it needs to be thread-locally bound first
05:39	stilkov	you mean using binding?
05:39	_ato	yep
05:39	_ato	eg (binding [foo 1] (set! foo 2))
05:40	stilkov	ah ? now it makes sense; should have been more than obvious in hindsight
05:40	stilkov	esp. since I'm using binding all over the place
05:42	Chousuke	hmm, running git gc on my clojure clone once in a while seems to be beneficial
05:43	Chousuke	just shrank it from 55MB to 8.6MB :P
05:44	Raynes	We herd u liek Clojure, so we put some recurshen in ur recurshen so you can (apply str (repeat "recurshen while you "))
05:45	Raynes	Should probably be 'rescursh', but nobody is perfect.
05:45	Raynes	Except Rich.
05:45	Chousuke	:P
05:45	Chousuke	no-one is perfect. Some just are better at making others think so.
05:48	Raynes	Chousuke: In that case, wouldn't the perfect person be the one who could is perfect that making others think he/she is perfect?
05:58	Intensity	I'm wondering what the elegant way is to get a random element from a sequence. I'm using (nth result (int (rand size))), where "result" needs to be calculated, and "size" is either redundantly calculated or uses a very non-idiomatic workaround such as a (def temporary ...) within a (do ...). I'm quite new.
05:59	LauJensen	Intensity: rand-nth
06:01	Intensity	LauJensen: Ok, great. Thanks. In general, if I want to compute something like an infix f(g(x), h(g(x))), is there a most elegant way to avoid computing g(x) twice?
06:03	spariev	(let [g-res (g x)] (f g-res (h g-res))) maybe
06:05	Chousuke	let is your friend
06:06	Intensity	spariev: Ok, I'll work with that idea. Thanks. The code I am interfacing with is Java, and it's using (. this MethodName) repeatedly. Both myself and the one who wrote this temporary Clojure interface to that Java code are very new to Clojure. And I am also generally new to Java as well, although I'm not interested in picking up Java (only Clojure).
06:06	Chousuke	Intensity: prefer (.methodName obj) to (. obj methodName)
06:07	Intensity	Chousuke: Ok. And in chains of methods, then would I prefer (.. obj method1 method2)?
06:07	Chousuke	(and AClass/staticMethod also to explicit . forms)
06:07	Chousuke	well, actually I prefer (-> obj .method1 (.method2 args))
06:08	Chousuke	since it allows you to use clojure functions in the pipeline too
06:09	Chousuke	,(-> "foobar" .toUpperCase (subs 3)) for example
06:09	clojurebot	"BAR"
06:10	spariev	,(doto (java.util.HashMap.) (.put "a" 1) (.put "b" 2))
06:10	clojurebot	#<HashMap {b=2, a=1}>
06:11	Intensity	Chousuke: Ok. I'm told that "this" was bound to the local class. The code is wrapped in a (proxy ...). And the corresponding Java code is calling methods like GetCurrentState(), not this.GetCurrentState().
06:12	Chousuke	within a proxy form, this is bound to an instance of the proxy.
06:12	Chousuke	and I think the GetCurrentState() call in that case is just a shortcut for this.GetCurrentState()
06:13	Chousuke	or a static method, in which case you need to do (ClassName/GetCurrentState)... But it probably isn't.
06:13	Intensity	Chousuke: Ok. Is it a great idea to use "this" within the Clojure code also? Presumably I'd use (-> this .method1 (method2)), or maybe (method2) doesn't need parenthesis.
06:13	Chousuke	Intensity: you need to use this if you want to call methods on the object
06:14	Chousuke	Intensity: and yeah, parentheses aren't needed if the method/function doesn't any other parameters besides the object itself.
06:15	Intensity	Chousuke: Ok. Thanks for clarifying. Given that I'm using "this", then, I'd prefer using it once. (doto ...) and (-> ...) or maybe also ".." seem to be powerful tools I can use here.
06:15	Chousuke	Intensity: From the clojure perspective it's good to think of java instance methods always taking the object they're called on as the first parameter
06:15	Chousuke	Intensity: yeah. They're all macros :)
06:15	Chousuke	Intensity: Once you get used to clojure you'll be able to write such tools yourself.
06:16	Chousuke	but for now, enjoy using the existing ones :D
06:17	Intensity	Chousuke: Great. Does the fact that those are macros mean that (-> obj .method1 (.method2)) and the same thing with .method2's parenthesis removed are considered identical? I'd guess if -> was not a macro, then surrounding .method2 with parenthesis would make a difference.
06:17	Chousuke	Intensity: yes.
06:17	Chousuke	Intensity: the -> macro expects list forms eg. "(foo)" as its parameters but if it sees just a symbol "foo", it wraps it in a list first
06:19	Chousuke	what it does is pretty simple. it takes the first parameter (a piece of code) and puts it in the second place in the following piece of code (a list). That's then recursively repeated until there are no arguments left
06:19	Chousuke	and as I said, if the second parameter is not a list, it's wrapped in one
06:20	Intensity	Ok. I'm trying to simplify: (. (. this (getStateMachine)) (getLegalMoves (.. this (getCurrentState) (getRole)))). I'm guessing I can make the highest level "." into a ".." and assume that the parenthesis can indicate to .. what structure I want to apply to the sequence of calls. Perhaps (doto this) or "->" could be useful here too.
06:20	Chousuke	so in (-> obj .method) it first internally "becomes" (-> obj (.method)) and then (.method obj)
06:21	Chousuke	yeah
06:21	bozhidar	I was wondering something about this snippet of code from clojure's site
06:22	bozhidar	http://gist.github.com/383420
06:22	bozhidar	why is the logging line outputted 5 times?
06:22	bozhidar	shouldn't it be just one
06:22	Chousuke	I think that would be something like (-> this .getStateMachine (.getLegalMoves (-> this .getCurrentState .getRole)))
06:23	Intensity	Ok. It's my guess that I might turn that into: (.. this getStateMachine (getLegalMoves (getCurrentState getRole))), somehow inferring that each method is meant to operate on "this" in a recursive way. I'm wondering if I can avoid using "this" twice.
06:25	Chousuke	Intensity: I'm not sure if that works.
06:25	Chousuke	Intensity: I don't think there is a way in this case that isn't unnecessarily contrived
06:26	Intensity	Chousuke: It seems that I could describe a macro to work in the way to support a statement like the above. But then it would need to differentiate between Clojure functions imbedded within the structure and Java methods that were meant to be called in this pyramid fashion.
06:27	Chousuke	bozhidar: hm I can't even test that in my repl due to binding being ineffective on str
06:27	Chousuke	Intensity: You could write a macro to interpret C++ code in Clojure. :)
06:27	Chousuke	Intensity: but yeah, you can't with ..
06:28	bozhidar	Chousuke: in my repl it show this exact output as in http://clojure.org/concurrent_programming
06:28	Chousuke	-> can differentiate between methods and functions because it just deals with the symbols
06:28	bozhidar	so I think that binding works on str
06:29	Chousuke	bozhidar: in your version, but I'm running 1.2-master here
06:29	Chousuke	bozhidar: and it might just be that str is calling itself recursively
06:29	Chousuke	bozhidar: thus resulting in more logging prints
06:29	bozhidar	Chousuke: I assume so, I'm running 1.1
06:30	Raynes	I wish people would stop telling me that it's bad for my projects to depend on. 1.2 >:\|
06:30	Raynes	I live for the danger, man.
06:31	Intensity	Chousuke: Ok. Thanks for the help.
06:32	Chousuke	I think str calls itself once for each argument and then finally once for the final stringbuilder -> string conversion
06:32	Chousuke	which would nicely account for all five logging prints.
06:38	Raynes	Is there something similar to filter/remove that returns a sequence of what was removed after it's finished?
06:38	_ato	I doubt it
06:38	rhickey	Raynes: returns 2 sequences?
06:39	Raynes	rhickey: Indeed.
06:39	Chousuke	you can do (juxt filter remove)
06:39	Raynes	Chousuke: I planned on doing that. Just wanted to be sure there wasn't already a function for that in core.
06:39	Raynes	:>
06:40	rhickey	http://richhickey.github.com/clojure-contrib/seq-api.html#clojure.contrib.seq/separate
06:40	Chousuke	the problem with having two sequences like that is that they will both hold onto the head of the original sequence :/
06:40	Raynes	That isn't a problem.
06:40	Raynes	It's for my defplugin macro.
06:41	Raynes	Expanding it for some hooking stuff.
06:41	Raynes	I don't think I'm going to do it like this.
06:42	Raynes	Just running ideas around my head.
06:42	Raynes	:>
06:48	vu3rdd	partial can take arguments in the order, not any arbitrary args, am I right?
06:49	Chousuke	right. I think.
06:49	Chousuke	IF I understood you correctly
06:49	bozhidar	rhickey: can I bother you with a question about one of the example on clojure.org?
06:51	kzar	So how should I escape variables I put into the SQL when I use with-query-results?
06:52	vu3rdd	Chousuke: were you replying to me?
06:52	Chousuke	yes
06:52	vu3rdd	thanks
06:53	carkh	kzar : something like (with-query-results rs ["select * from customer where id=?" customer-id] (dostuff rs))
06:53	carkh	escaping occurs automatically
06:54	Chousuke	if you're going to work a lot with SQL take a look at ClojureQL
06:54	kzar	carkh: Awesome thanks carkh
06:54	kzar	Chousuke: OK, I avoid it generally but this project needed to create a MySQL database
06:54	carkh	Chousuke: clojureql sometimes won't go as far as needed
06:55	carkh	like when using database engine specific features
06:55	carkh	for instance (last time i checked) the returning statement from postgres
06:55	Chousuke	carkh: it does support raw SQL when needed doesn't it?
06:55	carkh	well then what's the point =P
06:55	vu3rdd	Is there a way to give default value to function parameters?
06:56	Chousuke	vu3rdd: not directly, but arity overloading is a common workaround
06:56	vu3rdd	Chousuke: oh.. I forgot about it. Thanks
06:57	cemerick	There's also :or in map destructuring, which is a nice route
06:58	Chousuke	a defn-with-defaults might be useful
06:58	carkh	i like the addition from hiredman for that : (defn bleh [& {destructuring here}] ...) you get real named parameters !
06:58	Chousuke	(blah somefn [foo = 2, bar = 3, something] ...)
06:58	cemerick	carkh: yeah, that's quickly becoming my default, outside of absolutely required params.
06:59	carkh	yes absolutely love it =)
06:59	Raynes	vu3rdd: Note that in 1.2, we actually have default keyword arguments.
06:59	Raynes	vu3rdd: It's map destructuring without the map.
06:59	vu3rdd	Raynes: yes, I have seen that.
06:59	Raynes	sexpbot uses it.
06:59	Chousuke	how did the syntax for that go again?
06:59	Raynes	Chousuke: Same as for map destructuring.
07:00	Chousuke	was it just & {map-form}
07:00	carkh	Chousuke: yes
07:00	Raynes	(defn func [& {:keys [omg wtf bbq] :or {omg "omg"}}] ..)
07:00	Chousuke	there aren't too many tickets left with 1.2
07:01	carkh	Raynes: i wonder what that function's doc string says ...
07:01	Raynes	"Omgs your wtf with your bbq."
07:04	vu3rdd	paredit slurp is really really awesome!
07:07	carkh	i eventually switched to paredit after several years, it takes a bit of time getting used to it =/
07:08	carkh	specially ctrl-arrow not doing what expected anymore
07:09	vu3rdd	switching back to arity, one problem with variable arity to handle defaults is that we need to repeat the body of the code in certain situations.
07:10	cemerick	vu3rdd: you mean with overloads? Yes -- another reason why kwargs are so much friendlier in such circumstances.
07:11	vu3rdd	hmm.
07:12	Chousuke	vu3rdd: that's why I had the idea for a defn-defaults
07:13	Chousuke	vu3rdd: it would be trivial to generate the extra arities in a macro
07:13	Chousuke	as long as you accept that (defn [foo = 1, bar, zonk = 3] ...) makes no sense
07:13	vu3rdd	Chousuke: yes, I was also thining like that.
07:14	vu3rdd	it is a very neat idea to have a macro for that.
07:21	Raynes	carkh: The docstring is "Omgs your wtf with your bbq."
07:21	Raynes	$(update-in {:wtf []} [:wtf] #(conj % "LOL"))
07:21	sexpbot	result: {:wtf ["LOL"]}
07:25	Raynes	$(update-in {:wtf []} [:wtf] conj "LOL")
07:25	sexpbot	result: {:wtf ["LOL"]}
07:25	Raynes	Cool.
07:33	SynrG	hmm
07:33	SynrG	$(update-in {:wtf {:omg []}} [:wtf :omg] #(conj % "LOL"))
07:33	sexpbot	result: {:wtf {:omg ["LOL"]}}
07:34	carkh	$(update-in {:wtf {:omg []}} [:wtf :omg] conj "LOL")
07:34	sexpbot	result: {:wtf {:omg ["LOL"]}}
07:35	carkh	why mùake it hard to read ?
07:54	SynrG	carkh: good call. i didn't take suitable note of "... and any supplied args" in the doc
07:54	carkh	=)
07:56	Raynes	SynrG: Neither did I at first.
07:56	carkh	question : if i have a var cark.a/a in namespace cark.a and a var in namespace cark.b defined as (def b cark.a/a), when i recompile cark.a, the value of cark.b/b doesn't change (of course).
07:56	carkh	what would be an easy way to have the value of cark.b/b to follow the value of cark.a/a as i recompile the later
07:57	Raynes	You could reload cark.b.
07:57	SynrG	you know, clojure is so great, it's kinda depressing. because then i have to return to work on our huge, legacy codebase rife with mutable state, side-effects, and really grotty interfaces in my day job
07:57	carkh	so far i'm defining cark.b/b as this : (def b ((fn [] cark.a/a)))
07:58	carkh	but that's ugly
07:58	carkh	Raynes: i'm trying to avoid reloading
07:58	carkh	that's the whole point actually =)
07:58	_ato	that's weird
07:58	Raynes	^
07:58	_ato	how does that help?
07:58	carkh	err actually it doesn't
07:59	carkh	but what i'm doing is make everything functions and call them
07:59	carkh	tried to simplify the example, but it was too simplified
07:59	carkh	=P
08:00	Raynes	You could use a ref. ;)
08:00	carkh	or an atom yes
08:00	carkh	but then i need to sprinkle @ everywhere
08:00	carkh	which is ugly too
08:01	carkh	what i'm trying to achieve : i don't like compojure because it doesn't scale to hundreds of pages
08:01	cgrand	carkh: (refer '[cark.a :only [a] :rename {a b}])
08:01	carkh	mhhh
08:01	Raynes	We have :rename?
08:01	carkh	let me explain further
08:02	carkh	so i have this routing system that will route the request from namespace to namespace until it reaches a page
08:02	carkh	something like this : (def top (matchers customers users))
08:03	carkh	then in another namespace (def customers (mactchers ["add" add-customer] ["list" list-customers]))
08:03	carkh	and so on
08:04	carkh	but when i make a change in customers, it won't reflect in top unless i reload the whole thing
08:04	carkh	which defeats the purpose of image based hacking =P
08:06	Raynes	If I have a sequence of functions with side-effects whose values I do not care about, what is the best way to iterate through the sequence calling them all?
08:06	Raynes	(doseq [f fns] (f)) is simple enough, of course.
08:09	carkh	,(dorun (map #(%1 %2) (repeat print) [1 2 3 4]))
08:09	clojurebot	1234
08:09	carkh	returns nil
08:09	carkh	but yes doseq seems more appropriate
08:10	LauJensen	Raynes: doseq is better, because it implies that you're going for side-effects
08:18	cgrand	Raynes: yes doseq
08:18	Raynes	Thanks, friends. <3
08:23	AWizzArd	a nice thing could be a pdoseq
08:24	eevar2	AWizzArd: aren't side effects the whole point of doseq? how do you paralellize that?
08:25	eevar2	assuming that's what you ment by p
08:25	Chousuke	AWizzArd: I think (dorun (pmap ...)) is good enough for parallel execution of side-effects :)
08:25	Raynes	Rich could parallelize two 6 ton pipes at an angle from each other with his bare hands.
08:26	Chousuke	:P
08:26	AWizzArd	Chousuke: yes, i currently use pmap for that
08:27	AWizzArd	eevar2: the seq we traverse may be a seq of independend jobs.
08:30	spariev	so pdoseq would be basically syntactic sugar for (dorun (pmap ... ))
08:32	zakwilson	I want a function that, gives a seq and a predicate gives me a collection of the items matching the predicate and a collection of the items not matching the predicate. partition-by is almost it, but not quite
08:32	zakwilson	err... given a seq...
08:32	mmarczyk	there's separate in contrib, I think
08:32	mmarczyk	,(doc clojure.contrib.seq-utils/separate)
08:32	clojurebot	"([f s]); Returns a vector: [ (filter f s), (filter (complement f) s) ]"
08:34	zakwilson	Yes... that does what I want (and I pretty much wrote it myself already), but it calls f (or (complement f)) twice for each element, which I hoped to avoid.
08:34	carkh	~def clojure.contrib.seq-utils/separate
08:35	carkh	ah yes it does =/
08:36	carkh	http://github.com/richhickey/clojure-contrib/blob/master/src/main/clojure/clojure/contrib/seq.clj#L42
08:37	zakwilson	It probably won't become too speed-critical. I'll go with that until I have a problem.
08:38	jacortinas	test
08:38	mmarczyk	right
08:38	mmarczyk	well, you could (map #(vector % (f %)) your-seq)
08:39	mmarczyk	then filter on the second item in each vector
08:39	mmarczyk	then (map first ...)
08:39	mmarczyk	if (f %) is really expensive to make this worthwhile
08:40	cgrand	separate aka (juxt filter remove)
08:40	cgrand	,((juxt filter remove) even? (range 10))
08:40	clojurebot	[(0 2 4 6 8) (1 3 5 7 9)]
08:40	slyphon	hello?
08:40	clojurebot	BUENOS DING DONG DIDDLY DIOS, fRaUline slyphon
08:41	slyphon	hah
08:41	The-Kenny	Oo
08:42	The-Kenny	(Is hiredman just as crazy as his creation?)
08:42	mmarczyk	cgrand: that still calls even? twice per element, right?
08:43	cgrand	right
08:44	chouser	((juxt filter remove) (memoize even?) coll)
08:44	chouser	that doesn't. :-)
08:44	mmarczyk	but surely we don't need to have a `separate' function if (juxt filter remove) is available :-)
08:45	mmarczyk	chouser: oh, now you're making my convoluted solution look bad :-)
08:46	chouser	well, you have to be pretty careful with a plain memoize like that. can eat a lot of memory
08:46	_ato	heh, chouser's solution is also a huge memory leak though, yours is not ;-)
08:47	mmarczyk	well, I suppose, for a really large coll :-)
08:48	mmarczyk	btw, apparently it's possible to create a function which takes 20 positional arguments and rest args
08:49	mmarczyk	which then consistently returns nil without evaluating its body if applied to >= 20 args
08:49	mmarczyk	(throws arity for less)
08:50	chouser	>= 20 literal args, right? not 'apply' with a seq of >= 20
08:50	mmarczyk	right
08:51	mmarczyk	although it's the same with apply, just checked
08:51	_ato	,((fn foo [a b c d e f g h i j k l m n o p q r s t & z] z) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22)
08:51	clojurebot	nil
08:52	_ato	,((fn foo [a b c d e f g h i j k l m n o p q r s & z] z) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22)
08:52	clojurebot	(20 21 22)
08:52	_ato	interesting
08:52	chouser	not good
08:52	mmarczyk	it's the same with latest (close to latest?) checkout from github
08:52	bendlas	Hi folks
08:52	mmarczyk	someone on SO asked why Clojure limits arg number to 20, ran into this while replying
08:54	fogus	mmarczyk: Do you have a link to that SO question?
08:54	mmarczyk	I was going to write about & args, then about mixing that with 20 positional args, then had to change that to 19... haven't investigated the reasons yet
08:54	mmarczyk	yes, sure, just a sec
08:54	mmarczyk	http://stackoverflow.com/questions/2735600/why-is-there-a-limit-of-max-20-parameters-to-a-clojure-function
08:54	sexpbot	"Why is there a limit of max 20 parameters to a clojure function - Stack Overflow"
08:54	mmarczyk	latest one in the clojure tag
08:54	fogus	mmarczyk: thank you
08:56	vegai	anyone know how I get a new version of a MEAP book in Manning?
08:56	mmarczyk	vegai: you should receive an e-mail from Manning with a d/l link
08:56	vegai	I was going to check out if there's a later version of Joy of Clojure, but the link I got in mail is no longer valid
08:57	mmarczyk	vegai: if you don't, e-mail them
08:57	fogus	vegai: That link is only valid for 5 days
08:57	fogus	I think if you email them they will send another
08:57	fogus	(seems to have worked for others)
08:57	vegai	oh, now I see that I got the previous mail on 13th
08:57	vegai	no wonder there haven't been newer versions yet
08:58	fogus	vegai: We're hoping to have 2 more chapters out soon, but I wouldn't want to place any Vegas bets on that
09:00	vegai	righto
09:02	LauJensen	fogus: if I buy your book today, how many chapters can I read and how many are still pending ?
09:03	fogus	There are 5 chapters available right now and two more in the pipeline, leaving 5 remaning chapters (in various stages of completion)
09:03	LauJensen	ah ok ... I might just hold off for a liiitle bit longer then :)
09:04	LauJensen	One of my old colleagues bought Programming Clojure way too early, so he ended up having to re-read the whole thing once it was released
09:04	fogus	LauJensen: We'd love your feedback, but I can sympathize
09:05	LauJensen	Arent you guys getting a ton of feedback already fogus ?
09:05	fogus	indeed
09:06	LauJensen	Ok - If you really were in need of help I would happily pitch in, but it sounds like you guys are right on track
09:34	jacortinas	I'm wondering if I can use http-agent(in clojure.contrib) with an http proxy, anyone do anything like this?
09:35	ordnungswidrig	jacortinas: doesn't the java.net system properties work?
09:35	ordnungswidrig	s/doesn't/don'/
09:35	ordnungswidrig	s/'/t/
09:35	ordnungswidrig	sigh
09:35	jacortinas	ordnungswidrig: is that really all it would take?
09:36	ordnungswidrig	I'm not sure...
09:37	Raynes	Swank is worthless with compile-time errors.
09:37	Raynes	"Who needs a line number? Psh, let's just throw a bunch of random useless swank stuff into the stack-trace, and you can work with that. kthnxbai"
09:38	ordnungswidrig	it uses HttpURLConnection which should make use of http.proxyHost and http.proxyPort
09:39	ordnungswidrig	Raynes: are you uptodate? For me it reports the line number
09:39	zakwilson	,(contains '(:foo :bar) :bar)
09:39	clojurebot	java.lang.Exception: Unable to resolve symbol: contains in this context
09:39	Raynes	ordnungswidrig: Yes, I am.
09:39	Raynes	1.2-SNAPSHOT.
09:39	zakwilson	,(contains? '(:foo :bar) :bar)
09:39	clojurebot	false
09:39	zakwilson	why?
09:39	clojurebot	why not?
09:39	ordnungswidrig	Raynes: ok, I'm using clojure-1.1
09:39	Raynes	If I try to require something, and no dice, it gives me an error message and a random bunch of garbage.
09:40	Raynes	ordnungswidrig: I mean 1.2-SNAPSHOT swank.
09:40	zakwilson	Silly bot... why not? Because it obviously does contain :bar.
09:40	ordnungswidrig	,(doc contains?)
09:40	clojurebot	"([coll key]); Returns true if key is present in the given collection, otherwise returns false. Note that for numerically indexed collections like vectors and Java arrays, this tests if the numeric key is within the range of indexes. 'contains?' operates constant or logarithmic time; it will not perform a linear search for a value. See also 'some'."
09:40	Raynes	And 1.2 Clojure, but that's irrelevant.
09:41	ordnungswidrig	Raynes: ah, yes, the require issue. It likes to report line no: 0, right?
09:41	Raynes	It reports line number nothing.
09:41	Raynes	It doesn't even try to give me a line number.
09:41	zakwilson	Oh, I get it.
09:41	Raynes	Unable to resolve symbol: prepend in this context
09:41	Raynes	[Thrown class java.lang.Exception]
09:41	Raynes	And that's it, besides the worthless stacktrace. :p
09:41	ordnungswidrig	Raynes: or like that
09:42	ordnungswidrig	Raynes: java -jar clojure.jar mybrokenscript.clj will do better?
09:43	Raynes	Yeah, I can track it down. It's just annoying sometimes. :<
09:43	ordnungswidrig	C-c C-k also does better reporting than C-c C-l
09:44	Raynes	I usually use that, but I'm not using either at the moment. My code loads all of the plugins whenever I use sexpbot.run.
09:52	Raynes	That's weird.
09:52	Raynes	paredit-mode randomly disabled.
09:52	Raynes	stuartsierra: ohai
09:52	stuartsierra	hi
09:55	vu3rdd	stuartsierra: the dotrace procedure you wrote in the c.c.trace is really nice. Very helpful to me. Thanks for that.
10:01	stuartsierra	wu3rdd: you're welcome
10:01	stuartsierra	actually, I think someone else added dotrace after I wrote the original lib
10:06	vu3rdd	stuartsierra: oh.. ok
10:06	vu3rdd	JoC new MEAP has arrived!
10:07	Chousuke	it has? where's my mail :(
10:07	The-Kenny	Chousuke: Mine hasn't arrived yet too
10:07	vu3rdd	May be the electrons travel faster to India. :-)
10:07	The-Kenny	*Hitting Refresh in Gmail"
10:08	The-Kenny	vu3rdd: There was the case of the 300 Mile Email
10:08	vu3rdd	:-)
10:08	The-Kenny	Some strange stuff with timing etc. which prevented emails from going farther than 300 miles
10:08	spariev	Postmen these days ...
10:08	vu3rdd	7 chapters now
10:08	vu3rdd	ch 6 functional programming and ch 7 macros
10:08	The-Kenny	http://www.csua.berkeley.edu/~ranga/humor/500_mile_email.txt There it is
10:09	The-Kenny	Ok, looks down. Here's a mirror if someone is interested: http://www.ibiblio.org/harris/500milemail.html?
10:09	sexpbot	"The case of the 500-mile email"
10:11	vu3rdd	When is the next MEAP? The last sentence of ch 7 says: "In the next chapter we will cover various powerful way to organize and categorize data types and functions using Clojure's namespaces, multimethods, types, and protocols."
10:11	vu3rdd	"
10:21	kzar	If you have something like this {:dumb "example"}, how do you destructure it to the key and value without knowing what the key is?
10:22	cYmen_	Does anybody know of sample code or tutorials on image processing using clojure?
10:22	mmarczyk	kzar: you don't
10:23	kzar	mmarczyk: Well how would I get the key and the value out of something like that without destructuring?
10:23	LauJensen	cYmen_: define 'processing'
10:24	mmarczyk	kzar: I'm not sure what you mean by that, but...
10:24	mmarczyk	,(seq {:dumb "example"})
10:24	clojurebot	([:dumb "example"])
10:24	cYmen_	LauJensen: Read image from $foo invert save to $bar...
10:24	mmarczyk	you can map a function across a map
10:24	mmarczyk	if you need to process entries one by one
10:25	mmarczyk	then you get [key value] pairs
10:25	LauJensen	cYmen_: I suggest you check out the ImageIO class which lets you read the image and work on the graphics context, then save it again, should be only a few lines
10:25	LauJensen	$google Best In Class Tribute to Steve Ballmer
10:25	sexpbot	First out of 1650 results is:
10:25	sexpbot	My tribute to Steve Ballmer \| BEST IN CLASS
10:25	sexpbot	http://www.bestinclass.dk/index.php/2010/02/my-tribute-to-steve-ballmer/
10:25	kzar	mmarczyk: Sweet doing seq on it is a good idea, that lets me do what I wanted. Thanks
10:25	spariev	horaay for JoC, received my MEAP letter :)
10:25	LauJensen	I think I do almost that in this post, only I dont invert the colors but scale it @ cYmen_
10:25	mmarczyk	LauJensen: that was one seriously cool blog entry! :-)
10:25	LauJensen	mmarczyk: Thanks! :)
10:25	mmarczyk	kzar: hth :-)
10:26	mmarczyk	(as in happy th)
10:26	eevar2	cYmen_: and do have a look the java2d api docs
10:26	cYmen_	LauJensen: Thanks.
10:26	eevar2	and tutorial
10:26	cYmen_	eevar2: and to you, too.
10:26	esj	hooray - Fresh MEAP !
10:36	Chousuke	ooh, now I got mine too
10:39	bsteuber	me not yet sniff
10:42	cYmen_	meap?
10:43	pjstadig	everytime i see meap i think they missed out by not calling it the Manning Early Access Trial
10:44	LauJensen	pjstadig: they actually tried that when ...
10:44	LauJensen	~suddenly
10:44	clojurebot	CLABANGO!
10:45	djpowell	do manning nicen up the typesetting a bit before release?
10:45	pjstadig	clojurebot: meap is manning missed the opportunity to call it Manning Early Access Trial
10:45	clojurebot	'Sea, mhuise.
10:45	pjstadig	~meap
10:45	clojurebot	meap is manning missed the opportunity to call it Manning Early Access Trial
10:45	pjstadig	bah
10:45	pjstadig	been too long since i've instructed clojurebot
10:46	chouser	djpowell: I hope so. At least move the footnotes to the bottom of the page...
10:47	fogus	djpowell: I believe they have a more formal layout process for the printed book.
10:48	djpowell	how do I do clojure patches? I posted #300 a while ago re newlines, but i worry I've made some patch process faux pas.
10:49	djpowell	do i have to post the mailing list about the patch before i create a ticket?
10:49	stuartsierra	that is preferred
10:50	stuartsierra	http://clojure.org/patches
10:52	djpowell	hmm, actually I did post back in november
11:00	Raynes	Shuffle is in core now? Cool.
11:02	bsteuber	oh, didn't know about the /patches documentation
11:03	bsteuber	would have kept me from doing things in a silly order :)
11:06	bsteuber	stuartsierra: so how are people supposed to find the clojure.org/patches page?
11:06	wthidden	stuartsierra: why does json/read return 0 for an element that is an empty string?
11:07	cgrand	stuarthalloway: looking at your patch for :added in vars metadata, no :added "1.0"?
11:07	stuarthalloway	cgrand: true
11:07	stuarthalloway	rich wanted to leave that implicit
11:09	cgrand	stuarthalloway: ok, so I guess the policy is to tag all new vars with :added as soon as they are created.
11:10	stuarthalloway	cgrand: as soon as they are public and have a docstring
11:10	stuarthalloway	which of course requires creation :-)
11:10	stuarthalloway	example corner case: definterface
11:11	stuarthalloway	it is public, but not doc'ed or metadata'ed yet
11:12	djpowell	by public, do you mean not-private or publicised?
11:13	stuartsierra	bsteuber: dunno
11:13	djpowell	is this to avoid clashes between clojure.core and your code?
11:13	stuartsierra	wthidden: dunno, maybe that's a bug, got a test case?
11:16	djpowell	i like cgrand's :as suggestion
11:17	cgrand	I'm trying to figure out the rules for adding :as-of in refer..., if nil means "1.0" and below, definterface will be refered even when refering core :as-of "1.1"
11:18	djpowell	i guess it would require even non-publicised vars in core to be given an :added tag
11:19	cgrand	should I check :doc too?
11:19	djpowell	which I don't think should be a big problem?
11:20	djpowell	I think var polution will be a problem in clojure if we don't do something
11:21	chouser	this is strictly about var-existence, right? If the behavior or arglist of a fn changes, that won't be reflected in the metadata at all?
11:21	djpowell	i wonder who it will benefit though?
11:22	cgrand	only about preventing nameclashes
11:22	djpowell	if you are going to the trouble of adding :as-of tags to your code, then you could probably fix the nameclash at the same time
11:24	djpowell	I got bitten by char-array in 1.1, cause I'd already created my own char-array function for 1.0
11:24	cgrand	for example Enlive is designed for 1.1 but worked fine with clojure 1.2 until the addition of flatten
11:26	djpowell	even C has some namespacing of global functions via _ prefixes to prevent compiler supplied functions from clashing with user ones
11:32	rhickey	stuarthalloway, cgrand : changed my mind on "1.0", I think it will be more useful with it
11:33	rhickey	cgrand: definterface is due for a docstring and then a "1.2" designation
11:33	stuarthalloway	rhickey: ok
11:43	cgrand	rhickey: ok, so all vars listed here http://gist.github.com/383775 should not be referred?
11:45	rhickey	cgrand: other than definterface, they all look like implementation details to me, yes
11:46	cgrand	ok, thanks
11:53	Raynes	Who was it that asked me to combine the first and second lines of $google results?
11:53	Raynes	Anyways
11:53	Raynes	$google omg
11:53	sexpbot	First out of 6190000 results is: omg! Celebrity gossip, news, photos, babies, couples, hotties, and ...
11:53	sexpbot	http://omg.yahoo.com/
11:55	cburroughs	Is there a clojure json library that can handle unquoted keys?
11:57	cburroughs	heh, google tells me I asked that here on 2009-12-09 and apparently did not find one
11:59	hiredman	cburroughs: you know thats not valid json, right?
12:00	cburroughs	oh yes, I want to read it, not write it
12:02	cemerick	cburroughs: sounds like you want to use rhino's eval or somesuch
12:13	cburroughs	cemerick, that is a special mixture of crazy and genius
12:14	cemerick	cburroughs: which?
12:15	cburroughs	I'll tell you after I get it to work
12:15	cemerick	heh
12:16	cemerick	chouser / rhickey: any chance "non-eager" AOT-compilation will make it into 1.2?
12:17	rhickey	cemerick: what's that?
12:17	cemerick	hrm, right, sorry
12:17	chouser	cemerick: you mean some way to turn off AOT for some 'require'd namespaces?
12:17	cemerick	rhickey: this, although it was filed in contrib for some reason: https://www.assembla.com/spaces/clojure-contrib/tickets/23
12:17	sexpbot	"#23 - Clojure AOT compilation will compile unwanted classes (Invalid) \| Clojure Contrib \| Assembla"
12:18	cemerick	chouser: or, better, have AOT performed only for explicitly-named libs
12:19	rhickey	cemerick: no
12:19	cemerick	bummer
12:20	rhickey	wasn't even on the radar, and we're trying to wrap up
12:20	cemerick	rhickey: Sure -- I just thought of it, and figured I'd see if it was on the table. Are you sympathetic to the issue for future?
12:20	cemerick	the* future, that is
12:21	rhickey	cemerick: yes, also want to tackle the generated names, now we have the possibility of stable, e.g. function names
12:21	cemerick	OK; I'll file a ticket for it in the clojure assembla.
12:22	cemerick	I'm going to start lobbying to have source distributions become the norm for clojure-maven-plugin-based builds. If this issue could be resolved, that would eventually make things a lot easier w.r.t. distribution concerns.
12:23	technomancy	btw, ticket #315 would help quite a bit for reducing dependence on AOT
12:23	technomancy	and it's only 7 lines
12:23	technomancy	(for stuff that is intended for CLI usage at least)
12:25	cburroughs	cemerick, I'm sticky with both: (.eval e (slurp* "http://www.matsblog.com/static/matsblog/files/JSON-js-prettyPrint.js"))
12:25	cburroughs	(read-json (.eval e "eval({foo:1}).toJSONString()"))
12:25	cburroughs	{"foo" 1}
12:25	chouser	#315 is also an nice reduction in the need to use gen-class. is good!
12:25	cemerick	cburroughs: ah, glad it worked out for you :-)
12:26	technomancy	ticket #315
12:26	stuartsierra	For lazytest: what do people think of having assertions inside a vector, like pre/post conditions?
12:26	chouser	hope that's trustworthy input
12:26	cemerick	cburroughs: though, do you need the round trip through JSON? Surely rhino's eval returns a java.util.Map given a js object literal.
12:26	technomancy	~ticket #315
12:26	clojurebot	{:url http://tinyurl.com/37u5rdb, :summary "Add support for running -main namespace from clojure.main without AOT", :status :test, :priority :normal, :created-on "2010-04-24T13:04:43-03:00"}
12:29	cburroughs	cemerick, I get #<NativeObject [object Object]>
12:33	tomoj	not sure where this question really belongs, but..
12:33	tomoj	if I 'sudo su', then run 'lein clean', it works fine
12:33	tomoj	if I sudo sh -c 'lein clean' it complains that leiningen is not installed
12:33	tomoj	I'm assuming this is an environment variable thing
12:33	slyphon	PATH
12:34	tomoj	not PATH, the error comes from lein
12:34	slyphon	hrm
12:34	tomoj	it's the "Leiningen is not installed. Please run "lein self-install"."
12:34	cemerick	cburroughs: I don't know jack about rhino, but it looks like NativeObject (or its superclasses?) has a .getIds and a .get(id-name, object) method...using those would surely be better than str -> eval -> json-string -> eval -> read-json
12:34	tomoj	it is installed, in /root/.m2
12:34	slyphon	one is running with root privs as your user
12:34	slyphon	i.e. your $HOME is the same
12:34	tomoj	right
12:34	slyphon	the other is actually running as root
12:34	tomoj	so which env var does lein care about, I wonder
12:34	slyphon	$HOME, i'll bet
12:34	sexpbot	Command not found. No entiendo lo que estás diciendo.
12:34	tomoj	oh, of course
12:35	cemerick	cburroughs: http://stackoverflow.com/questions/2559450/rhino-how-to-get-all-properties-from-scriptableobject
12:35	tomoj	well, except...
12:35	sexpbot	"Rhino: How to get all properties from ScriptableObject? - Stack Overflow"
12:35	tomoj	there's no ~me/.m2
12:35	cemerick	cburroughs: though, watch out for the .toString that guy is using on the return from .getProperty
12:35	slyphon	hum
12:35	cburroughs	interesting
12:35	stilkov_	send (on agents) and swap! (on atoms) inside a dosync get invoked only if the transaction is successful
12:35	tomoj	there is a /root/.m2 though
12:35	stilkov_	where is this documented?
12:35	slyphon	why are you running lein as root, anyway?
12:36	tomoj	I don't know a better way to do it yet
12:36	slyphon	that's a "shell script" you "downloaded off the internets"
12:36	tomoj	what I really want is a systemwide install
12:36	cburroughs	I find it ironic that so many hoops need to be jumped through to get javascript to have decent json support
12:36	tomoj	yeah, I trust technomancy though :)
12:36	slyphon	heh
12:36	slyphon	and somewhere, a unix sysadmin loses his wings
12:36	tomoj	maybe I should just have all this clojure build stuff happen offline
12:37	tomoj	I have hudson set up already, it can run the build and drop artifacts in S3
12:37	stilkov_	I don't really mean invoked, I meant the functions passed to them get executed
12:37	stilkov_	I can't seem to find this documented in the official documentation
12:37	tomoj	... except the output of my build is mixed in with my source checkout because it's solr :(
12:38	tomoj	maybe I need to use maven
12:38	chouser	a rhino NativeObject is apparently not at all a java.util.Map
12:38	slyphon	how would you say Object[] in clojure?
12:38	cemerick	Yeah, it's too bad that there's no easy way to get there.
12:39	chouser	roundtripping through another json reader is surely not necessary, but probably the shortest path to a correct solution.
12:39	cemerick	chouser: I'm guessing it's a quick (into {} (for ...))) away.
12:40	slyphon	or, i need to know how to call a varargs via wall-hack
12:40	chouser	yes, plus recursion, key-type coersion...
12:40	cemerick	cburroughs: FWIW, unquoted keys are not JSON.
12:40	xeqi	,(into-array Object [])
12:40	clojurebot	#<Object[] [Ljava.lang.Object;@12a5896>
12:40	slyphon	xeqi: that's an instance
12:40	slyphon	i guess
12:40	cburroughs	I meant the "toJSON" part of javascript support
12:40	slyphon	,(class (into-array Object []))
12:40	clojurebot	[Ljava.lang.Object;
12:41	slyphon	"ew"
12:46	tomoj	slyphon: you were right, it's $HOME
12:46	slyphon	yay!
12:47	tomoj	when you sudo sh -c '...', $HOME is your user's home
12:47	tomoj	when you sudo su, it's /root
12:47	tomoj	it looks like lein doesn't respect M2_HOME though..
12:47	tomoj	or maybe I misunderstand M2_HOME's purpose
12:48	chouser	I'm so looking forward to Clojure's built-in abstractions being protocols.
12:49	stuartsierra	chouser: me too
12:49	chouser	(extend-class NativeObject PersistentMap ...)
12:49	cemerick	tomoj: isn't M2_HOME only used by the mvn bootstrap scripts?
12:50	slyphon	variadic args are really crap in java
12:50	cemerick	FYI: http://www.assembla.com/spaces/clojure/tickets/322
12:50	sexpbot	"#322 - Enhance AOT compilation process to emit classfiles only for explicitly-specified namespaces (New) \| Clojure \| Assembla"
12:51	stuartsierra	chouser: would that work if NativeObject were mutable?
12:51	tomoj	cemerick: you're probably right
12:51	tomoj	I just guessed from the name
12:52	tomoj	that M2_HOME would point to wherever your .m2 is
12:52	tomoj	it always looks in $HOME/.m2, I guess?
12:52	chouser	stuartsierra: yeah, you'd just have to implement COW
12:52	cemerick	by default, yeah
12:52	stuartsierra	chouser: or choose a less restrictive protocol, like Associative
12:52	chouser	good point
12:54	cemerick	chouser: bleh, I totally forgot about the clojure.xml/emit patch you tasked me to bundle up. (http://www.assembla.com/spaces/clojure/tickets/277) Is that something you want to fold in to 1.2?
12:54	sexpbot	"#277 - Making clojure.xml/emit a little friendler to xml consumers (New) \| Clojure \| Assembla"
12:54	npoektop	hi! how to use clojure.contrib.http.agent with ssl? i want to do (http-agent "https://..." ...)
12:55	stuartsierra	npoektop: I think you need supporting SSL libraries for Java.
12:55	stuartsierra	Or a better HTTP client, like Apache's.
12:56	npoektop	supporting libraries?
12:56	stuartsierra	SSL libraries
12:56	stuartsierra	SSL is a pain almost everywhere, including Java.
12:57	npoektop	how to get and use them?
12:57	stuartsierra	no idea, sorry
12:58	stuartsierra	Try the Apache HTTP client, it may have built-in support.
12:59	cemerick	npoektop, stuartsierra: the JRE includes a perfectly well-functioning SSL HTTP client.
13:00	chouser	zipmap still doesn't use transients
13:01	mmarczyk	any reason in particular why seq-contains? isn't defined as (fn [coll x] (some #(= % x) coll)), without the if?
13:01	mmarczyk	(looking at the newly promoted functions...)
13:02	mmarczyk	ahhh
13:02	mmarczyk	,(some #(= % 5) [])
13:02	clojurebot	nil
13:03	mmarczyk	right.
13:05	chouser	http://gist.github.com/383892 -- includes at least 50% premature optimization!
13:07	chouser	cburroughs, cemerick ^^^
13:08	tomoj	cemerick: on the one that's in the JVM, don't you have to do funky stuff with certs?
13:08	tomoj	at least for ssl sockets
13:09	cburroughs	oh my
13:10	mmarczyk	chouser: 1s/set/set!/ :-)
13:10	cemerick	chouser: nice :-)
13:11	chouser	mmarczyk: oh no!
13:11	cemerick	everyone: FYI, there's a good chance that if you talk about some work you're doing, chouser will just do it for you. ;-)
13:11	chouser	heh. not a reputation I need. I deny it!
13:11	mmarczyk	cemerick: spot on :-D
13:12	chouser	huh, I thought extend-protocol would hint my arg for me
13:18	chouser	oh, of course it's the calls in the inner fn that aren't hinted. bleh.
13:19	chouser	there, that's better: http://gist.github.com/383892
13:19	chouser	now, really, back to work...
13:25	abedra	technomancy: you around?
13:26	abedra	technomancy: I have some time today to tackle the leiningen exit code issue and wanted to see if you had started any work on it yet
13:28	stuartsierra	Does swank-clojure-binary work with current swank-clojure?
13:29	tomoj	abedra: lein test exit code? or is there another problem?
13:29	abedra	tomoj: that's the one
13:40	cemerick	tomoj: you shouldn't have to mess with any certs to use the JRE's SSL client -- at least in JDK 1.4+ -- I think you might have had to do something special to enable SSL in 1.2 or 1.3.
13:40	cemerick	e.g. (slurp* "https://paypal.com") works just fine.
13:43	joshua-choi	I'm looking at the promoted seq functions. What is the difference between the intended uses of keep-indexed and map-indexed?
13:47	technomancy	abedra: no, I haven't had a chance to look into it. would be great if you could; thanks.
13:48	tomoj	cemerick: oh, just ssl sockets then, maybe
13:48	technomancy	stuartsierra: it should, but I don't know if anyone uses it, so there's a decent chance it's bitrotten
13:48	tomoj	I was looking into not-yet-commons-ssl
13:48	cemerick	tomoj: you mean if you were going to implement an ssl sever?
13:48	cemerick	server*
13:49	tomoj	I mean for example doing APN
13:50	tomoj	no server, just open a socket to some server over ssl
13:50	tomoj	I read there's something funky you have to do to use what's in the jvm
13:50	cemerick	tomoj: is that not what https is doing?
13:50	cemerick	says cemerick, exposing his ignorance...
13:50	tomoj	isn't it?
13:51	tomoj	but there's something funny
13:51	cemerick	tomoj: what's APN, btw? http://en.wikipedia.org/wiki/Access_Point_Name ?
13:51	sexpbot	"Access Point Name - Wikipedia, the free encyclopedia"
13:51	tomoj	apple push notification
13:51	cemerick	ah
13:51	tomoj	you have to open an ssl socket with some cert and push raw bytes
13:52	abedra	technomancy: cool, just forked and started digging in. When you create the java task i don't see where you would set failonerror=true
13:52	tomoj	but I just found a java project for it, so I will just do whatever they do
13:52	abedra	technomancy: that would make ant exit non zero if the tests didn't pass
13:52	abedra	but i'm looking further
13:52	replaca	technomancy: does swank work with the latest clojure 1.2? I thought I saw something yesterday that indicated that it might not.
13:53	cemerick	tomoj: if you're going to use "unlimited" strength ciphers, you do need to install the appropriate policy files in order to get those (e.g. 256-bit AES, etc)
13:53	tomoj	and what do you with the cert?
13:53	technomancy	replaca: it should be fixed, yeah
13:53	replaca	technomancy: do I need to pull a new one then?
13:53	cemerick	tomoj: well, the policy files are essentially orthogonal to certificates
13:54	technomancy	replaca: the fix is on clojars; that would be the easiest
13:54	replaca	technomancy: got it. thanks!
13:54	Raynes	technomancy: Ohai, philjure.
13:54	Raynes	:D
13:54	technomancy	...!
13:55	cemerick	tomoj: if you download this, you'll find info about them in the readme: https://cds.sun.com/is-bin/INTERSHOP.enfinity/WFS/CDS-CDS_Developer-Site/en_US/-/USD/ViewProductDetail-Start?ProductRef=jce_policy-6-oth-JPR@CDS-CDS_Developer
13:57	tomoj	cemerick: I'm just going to save that link and hope I don't have to look at it
13:57	tomoj	:)
13:57	cemerick	yeah
13:58	cemerick	it's absolutely absurd that those policy files aren't the default yet
13:58	cemerick	the export controls were lifted like, 3 years ago
13:58	cemerick	ha
13:59	cemerick	technomancy: only if I absolutely had to, but yeah.
13:59	technomancy	I have yet to see with my own eyes a difference between openjdk and sun's. I've heard stories, but I require empirical proof.
14:01	cemerick	they're different codebases, so I take it as a given that there are differences. shrug
14:01	rcg	technomancy just as you mention it .. i have to use some strange client application for a kvm switch which refuses to run with icedtea but runs with sun-jdk ..
14:02	technomancy	rcg: bah, hearsay. I'll believe it when I see it.
14:02	technomancy	=)
14:02	cemerick	hrm, right, I remember now: glassfish v2 wouldn't run with openjdk, but did just fine under sun's
14:02	rcg	hehe
14:03	cemerick	'course, that could be bad packaging for ubuntu's openjdk, but that was enough of an experience. :-)
14:05	abedra	technomancy: are there expected failures in the test suite for -stable?
14:05	LauJensen	I've been burned badly by openjdk as well
14:11	slyphon	hiredman: hey, is that colorized paren thing for emacs you were talking about "cparen"?
14:22	stuartsierra	does gist.el work?
14:22	slyphon	stuartsierra: i was wondering the same thing
14:22	stuartsierra	I get "Symbol's value as variable is void: login"
14:22	headius	cemerick: they're the same codebase
14:23	headius	some devs still use internal working copies, but it all pushes to the openjdk repos
14:23	cemerick	headius: well, now they are, but not for JDK 6
14:23	cemerick	?
14:23	headius	there's not much difference between 7 and 6
14:23	headius	recent updates of 6 have invokedynamic bits hidden in them
14:24	headius	just disabled
14:24	cemerick	hrm. Maybe it's all in the packaging. shrug
14:24	headius	there's only a few months lag between the hotspot that's in 7 and what gets into current 6 updates
14:24	cemerick	in any case, it's hard to argue with installing sun-java-6, and then having everything work after a couple hours of pain :-)
14:25	headius	we get almost weekly reports against our openssl library because of those missing policy files
14:25	cemerick	headius: is that going to be resolved for JDK 7?
14:25	cemerick	or better, a JDK 6 update...
14:25	headius	beats me
14:25	headius	it would be nice
14:28	hiredman	slyphon: rainbow-paren-mode
14:28	slyphon	yeah? i couldn't find that
14:28	slyphon	i found highlight-paren-mode
14:28	hiredman	best bet may be searching clojurebot's url log
14:28	slyphon	which does colorization based on the position of the point
14:28	slyphon	kk
14:28	slyphon	ty
14:29	hiredman	http://delicious.com/search?p=rainbow&chk=&context=userposts%7Cclojurebot%7C&fr=del_icio_us&lc=0
14:29	sexpbot	"Search results for rainbow on Delicious"
14:29	slyphon	ooh
14:30	slyphon	thanks :)
14:35	tcepsa	If I have a Clojure namespace with :gen-class that I compile into Java classes, and that namespace also depends on other namespaces (which don't have :gen-class), it looks like the other namespaces get compiled into classes as well. Is that correct?
14:36	tcepsa	(depends on == :requires)
14:36	chouser	tcepsa: yes, though it has nothing to do with :gen-class
14:37	tcepsa	All right, so far so good (though I'd like to know more about what it does have to do with, if not :gen-class, I think that's tangential to my actual question)
14:38	karchie	I just wrote a function take-after [xs x], which returns the subsequence of xs starting after the first appearance of x (or nil if x isn't found). is there a core (or contrib) equivalent that I should have seen but missed?
14:39	tcepsa	My actual question is if I then change one of those namespaces that the first namespace :requires and recompile the first namespace, the changed namespaces do not appear to be recompiled. Is that a bug, and is there something else that I have to do to force the changed namespace to recompile?
14:39	cemerick	tcepsa: AOT compilation occurs any time a namespace is loaded, and the flag that controls AOT is thread-bound, so transitively-loaded namespaces get caught up in the compilation when they're :require'd
14:39	cemerick	tcepsa: are you doing this interactively in a REPL?
14:40	tcepsa	cemerick: I am calling (compile 'namespace
14:40	tcepsa	) in a REPL, yes (sorry about the extra linebreak!)
14:40	cemerick	tcepsa: right -- so, the namespaces required by your top-level namespace are already loaded at that point, so they won't get AOT-compiled.
14:40	chouser	,(next (drop-while #(not= 4 %) (range 10)))
14:40	clojurebot	(5 6 7 8 9)
14:40	chouser	karchie: ^^^
14:41	abedra	technomany: soooooooo close, just need a way to actually do the exit
14:41	karchie	chouser: there it is, drop-while. thanks much
14:42	cemerick	tcepsa: if you add a :reload-all to your :require in the ns declaration, then the dependency(ies) should be reloaded when you call compile
14:42	cemerick	(I think)
14:43	tcepsa	cemerick: Ah, okay, thanks--and I may have spoken too soon, as it looks like the classes for the :required namespace were actually recompiled, but it just took a couple of minutes for that to happen. I'll experiment with it a bit more, and I'll try that
14:44	tcepsa	:reload-all flag as well
14:44	cemerick	tcepsa: AFAIK, compiling interactively is quite atypical
14:44	cemerick	You really want to have a tool do repeatable builds for you, e.g. maven or lein
14:45	tcepsa	cemerick: Ahh, yeah, that makes sense. How is Lein for Windows at this point? I've shied away from it so far because I thought I saw something that said it wasn't really stable.
14:46	cemerick	tcepsa: No idea. I'm of the maven persuasion. :-)
14:46	tcepsa	(And I'm completely unfamiliar with Maven. If I have to learn it for this project I will, but it doesn't seem like this is complicated enough to really require something that involved)
14:47	tcepsa	(At least I get the impression that Maven is involved. Perhaps that is a misconception on my part as well.)
14:47	cemerick	It's pretty simple to get started with IMO, but you'll find a lot of disagreement there.
14:48	tcepsa	All right, I'll have a look at both of them and see whether I can get it working. Thanks!
14:48	cemerick	tcepsa: here's a summary of my perspective, entirely FWIW: http://muckandbrass.com/web/display/~cemerick/2010/03/25/Why+using+Maven+for+Clojure+builds+is+a+no-brainer
14:48	sexpbot	"Why using Maven for Clojure builds is a no-brainer - Chas Emerick - Muck and Brass"
15:01	stuartsierra	Starting to think a real I/O library should use java.nio instead of java.io.
15:02	chouser	sounds right
15:02	danlarkin	io and nio have different use cases I think
15:02	stuartsierra	nio is lower-level
15:02	chouser	I've got a pile of code here using java.io and a sneaking suspicion it would be a good deal better on nio abstractions instead
15:03	danlarkin	nio is async, so when that's not what you want, you don't want nio
15:03	chouser	nio has sync methods
15:04	chouser	http://java.sun.com/javase/6/docs/api/java/nio/channels/FileChannel.html#read(java.nio.ByteBuffer)
15:04	sexpbot	"FileChannel (Java Platform SE 6)"
15:04	chouser	returns number of bytes read.
15:05	stuartsierra	The combination of asynch I/O and Clojure's thread protections might be very powerful.
15:06	danlarkin	maybe I misjudged nio
15:06	arkahn	has anybody seen an async way to 'tail' a file (e.g. a logfile) in a separate thread? If not, I think I'll write one
15:11	abedra	technomancy: We've got something that is "working" http://gist.github.com/384079 but I fear that we've missed something here
15:11	abedra	especially with shutdown-agents
15:12	abedra	can you take a quick look and tell me what is wrong?
15:14	arohner	stuartsierra: what kind of shape is clojure-rdf in?
15:15	stuartsierra	none
15:15	stuartsierra	it is unfinished, unmaintained, and unloved
15:15	arohner	stuartsierra: hah, ok. thanks
15:15	stuartsierra	np
15:16	stuartsierra	I need to flag most of my github projects as "disowned"
15:16	slyphon	as a general rule, when you have multiple arities of a function providing default values, should you "call your way up the chain"?
15:17	chouser	slyphon: not necessarily, though that removes duplicated default value.
15:17	chouser	Other than that, not much wrong with just calling the real implementation.
15:18	slyphon	k
15:19	chouser	depending on various factors, it's often nicer to just use destructuring and a single arity. (fn [a & [b c d]] ... (or b 25) ... (or d "hello")...)
15:19	slyphon	hmmm
15:20	slyphon	i hadn't thought of that
15:20	slyphon	destructuring the & more
15:20	LauJensen	chouser: Doesn't that go directly against the code-standard which says its better to unroll args ?
15:20	slyphon	interesting :)
15:20	chouser	LauJensen: heh, could be. where is that doc again?
15:20	LauJensen	assembla I think
15:21	chouser	http://www.assembla.com/wiki/show/clojure/Clojure_Library_Coding_Standards
15:21	sexpbot	"Clojure Library Coding Standards \| Clojure \| Assembla"
15:21	LauJensen	hmm
15:21	LauJensen	Unroll optional named arguments. Callers should not have to wrap optional named arguments in a map literal:
15:21	LauJensen
15:21	LauJensen	(release-sharks 2 :laser-beams true) ; good (release-sharks 2 {:laser-beams true}) ; bad
15:22	tcepsa	Is there a way to read in data structures (e.g. vectors or maps) from some kind of config file at runtime?
15:22	chouser	LauJensen: that's not what I'm talking about
15:22	LauJensen	chouser: oh, sorry
15:22	chouser	the only item that may apply is "only destructure in the arg list if you want to communicate the substructure as part of the caller contract"
15:23	slyphon	yeah, i was just using this
15:23	chouser	which is related, but doesn't advocate exclusively one way or the other.
15:23	slyphon	er, reading this
15:23	LauJensen	$(read-string "{:one 1 :two 2}")
15:23	sexpbot	result: {:one 1, :two 2}
15:23	LauJensen	@ tcepsa
15:23	slyphon	hmm
15:24	tcepsa	Perfect!
15:24	slyphon	if blah is only meant for rebinding, is there a convention for vars that are meant to be constants?
15:24	chouser	slyphon: yep. all lower-case, dash-separated words.
15:25	slyphon	oh
15:25	cemerick	I like the +constant+ convention
15:25	chouser	for example: clojure.core/drop-last is a constant. don't change its value to some other function please.
15:25	cemerick	most others don't, I suppose
15:25	slyphon	heh
15:25	slyphon	cemerick: is that CL?
15:25	cemerick	slyphon: I think that's its origin, yeah.
15:26	cemerick	Though I'm no CL expatriot. :-)
15:26	cemerick	I just like to distinguish what are really, honestly constants, vs. fns and other defs that you could rebind if you needed to.
15:28	chouser	there are some interesting non-Cartesian maths out there!
15:29	abedra	chouser: others have tried http://en.wikipedia.org/wiki/Indiana_Pi_Bill
15:29	sexpbot	"Indiana Pi Bill - Wikipedia, the free encyclopedia"
15:29	pjstadig	I might like the +constant+ convention more if there was actually a way to make constants
15:31	LauJensen	Clojure: The Lisp that makes Perl dynamic
15:31	cemerick	defonce should probably have prevented rebinding
15:31	pjstadig	not only for constantness, but also for speed
15:31	pjstadig	like symbol macros or something
15:32	cemerick	well, direct linking is already there, although I don't know how to use it.
15:32	pjstadig	yeah i may be out-of-date...ignore me
15:34	chouser	direct linking isn't yet available for general use
15:34	technomancy	no, pretty sure you can't opt-in to direct linking
15:34	cemerick	ah, I thought it was floating around in there somewhere
15:34	technomancy	even so you can always with-ns+redef
15:34	slyphon	LauJensen: huh?
15:34	technomancy	not to mention alter-var-root
15:35	LauJensen	slyphon: Just saying, you and me read -> ->> -?> @ #() % etc as plain english, but new-comers will stumble
15:35	slyphon	ah
15:35	slyphon	indeed
15:36	slyphon	a little confusion is good every once in a while
15:46	remleduff	rhickey: Hi, was just reading irc logs and saw you mention that you were hoping to wrap up 1.2 soonish. Was hoping you could look at ticket #317, it was driving me batty having my repl unexpectely quit like that while debugging a 3rd party library.
15:48	technomancy	abedra: sorry, headed off for a while; will try to look at that tonight.
16:52	wthidden	Is there a known issue with partial in the latest clojure build?
16:54	chouser	I've not heard of one
16:55	wthidden	hmm the issue is as follows
16:56	wthidden	i have a function x that takes 3 args so I make a partial of it and supply 2 args so i get a function taking one arg.
16:56	abedra	technomancy: no worries
16:56	wthidden	but the issue may be that within this function x I make another partial
16:57	wthidden	that is this the current function, sorta of a self-reference partial.
16:57	wthidden	it used to work, pre-1.2
16:58	chouser	wthidden: can you paste a small example that generates the error?
16:58	clojurebot	api examples is http://en.wikibooks.org/wiki/Clojure_Programming/Examples/API_Examples
16:58	wthidden	no blows up NoClassDefFoundError
16:58	wthidden	sure
16:58	wthidden	to where should i paste?
16:59	chouser	http://paste.lisp.org/new/clojure
17:01	hiredman	fyi, http://delicious.com/clojurebot/pastbin
17:01	sexpbot	"clojurebot's pastbin Bookmarks on Delicious"
17:02	technomancy	chouser: I wonder if you could get lisppaste back in here if you set it so unregistered nicks weren't banned
17:02	chouser	technomancy: ooh! could be...
17:02	chouser	Or if lisppaste8 registered itself
17:02	chouser	I have the authority to do neither.
17:03	wthidden	http://paste.lisp.org/+241S
17:03	technomancy	oh, you can't change the channel policy just by being an op?
17:03	wthidden	is the paste
17:03	chouser	I don't have op.
17:03	technomancy	oh, bummer
17:03	chouser	I had it for a few days, but whoever gave it to me apparently did it manually rather than through ChanServ, so next time I logged out I lost it.
17:04	technomancy	=\
17:04	chouser	wthidden: thanks, but that refers to several unreferred vars
17:05	wthidden	oops
17:05	wthidden	ok..
17:07	wthidden	hmmm here is the def for substitute-sym http://paste.lisp.org/+241T
17:08	chouser	wthidden: do you get the error if you just leave out that call? the smaller the example, the easier it will be to find the problem.
17:08	wthidden	hmmm trying to make it smaller.
17:08	chouser	thanks
17:09	wthidden	so what is somewhat strange is the i can call resolve-task-symbols fine
17:10	wthidden	and i think it creates a partial of itself without issue.
17:10	wthidden	but the closer i look at the def it may not call the parital with the test data i give it.
17:13	wthidden	when test data is a sequence gives the same error.. working on a simpler example.
17:14	arohner	is there another way to create the symbol "^" aside from (symbol "^")?
17:15	arohner	\^ doesn't seem to work
17:15	arohner	or '^
17:15	chouser	arohner: that's not a valid symbol
17:16	arohner	so then why does (symbol "^") work?
17:16	chouser	'symbol' is overly permissive
17:16	arohner	chouser: ok, thanks
17:17	chouser	,(read-string (pr-str (symbol "(+ 2 3 4)")))
17:17	clojurebot	(+ 2 3 4)
17:17	arohner	nice
17:17	chouser	,(read-string (pr-str (symbol "foo")))
17:17	clojurebot	foo
17:17	chouser	,(read-string (pr-str (symbol "^")))
17:17	clojurebot	java.lang.RuntimeException: java.lang.Exception: EOF while reading
17:18	chouser	that a somewhat better way to check, though really you're only safe using the chars listed at http://clojure.org/reader
17:24	wthidden	ok how about this paste: http://paste.lisp.org/+241X
17:24	unlink	am I missing something here? (reify IFoo (bar ([] 0) ([a] 1)))
17:24	unlink	Why can't I do this?
17:27	wthidden	chouser: i have to go catch the train.. will be back in a few hours.. maybe anotate will work, maybe not.... I'll restate my issue at a later date.
17:28	chouser	wthidden: ok
17:29	chouser	wthidden: that last paste produces your error?
17:42	remleduff	Why is it read-"string" and pr-"str"?
17:42	remleduff	Vowel shortage?
17:44	remleduff	,(read-string (pr-str (symbol "\u2038")))
17:44	clojurebot	‸
17:45	remleduff	,(read-string (pr-str (symbol "\u02c4")))
17:45	clojurebot	˄
17:45	remleduff	So many hats
17:51	unlink	,(reify Object (toString [] ""))
17:51	clojurebot	java.lang.Exception: Unable to resolve symbol: reify in this context
17:57	DavideAngelocola	hi all
17:58	unlink	ah. reify is different from proxy.
18:01	DavideAngelocola	I'm trying to use swank with maven but .. http://gist.github.com/384329
18:01	DavideAngelocola	any idea?
18:07	remleduff	I think the updates moving stuff from contrib to core last night broke the newest swank, you have to use the versions with timestamps that are listed in http://groups.google.com/group/clojure/msg/07eacb58a63de596
18:08	remleduff	Hmm, but you have contrib set to 1.1, so that shouldn't be the problem you're seeing
18:09	remleduff	Maybe you have an extra contrib 1.2-SNAPSHOT jar lying around somewhere
18:13	DavideAngelocola	I'm trying updating to clojure 1.2-SNAPSHOT
18:15	remleduff	I'm not sure swank works with the newest snapshot, I think it doesn't
18:18	DavideAngelocola	indeed, still the same error :(
18:18	arohner	remleduff: I'm getting a lisp reader error
18:19	remleduff	arohner: Those were silly play examples, the ^'s you get out of them won't actually be useable for anything resembling ^{metadata}
18:27	arohner	I was able to get swank-clojure to start and pass the tests by replacing #^ with ^. Give me a minute and I'll tell you if I can connect
18:31	arohner	ok, I'm having problems building my new changes
18:31	arohner	Caused by: java.lang.ClassCastException: clojure.lang.Cons cannot be cast to clojure.lang.Named
18:38	technomancy	clojure-test-mode may also need an update
18:40	arohner	technomancy: what is the deal with src/winston/rpc.clj? its formatting is all kinds of screwed up
18:41	arohner	sorry, src/swank/rpc.clj
18:45	etate	hey, why aren't monads idiomatic?
18:48	technomancy	don't look at me, I'm just the maintainer. =\ that's seriously ancient code
18:55	ataggart	etate: who said they weren't and in what context?
18:57	etate	ataggart: stuarthalloway at disclojure.org 28th april
18:58	etate	"Number of monads in Clojure 1.0: Zero. In 1.1: Zero. In 1.2: Zero. Monads are not idiomatic "
18:58	ataggart	link?
18:58	clojurebot	your link is dead
18:59	etate	ataggart: http://disclojure.org/2010/04/28/today-in-the-intertweets-april-28th-ed/
18:59	sexpbot	"Today in the Intertweets (April 28th Ed) \| disclojure: all things clojure"
18:59	etate	i just don't get whats not idiomatic about monads in a mostly purely functional programming languge
19:02	mmarczyk	etate: I prefer to take that to be a statement of the present state of affairs
19:02	zakwilson	I'm inclined to say that most of the things you'd want to do with monads are easier by other means in Clojure.
19:02	mmarczyk	after all, people don't seem to be writing lots of monadic ("explicitly monadic"?) code in Clojure
19:02	zakwilson	That said, I'd really like it if that port of Parsec was usable.
19:02	ataggart	the quote is "Number of monads in Clojure 1.0: Zero. In 1.1: Zero. In 1.2: Zero. Monads are not idiomatic #clojure.:
19:02	mmarczyk	say goes for Scheme, even when it's being used in a purely functional manner (which it often is)
19:02	ataggart	which I take to mean Stuart is drawin a conclusion based on the lack of usage of monads
19:03	ataggart	it's twitter, don't read too much into it
19:03	etate	okay :)
19:03	mmarczyk	if "idiomatic" means "widely used by the community", then he's right
19:03	mmarczyk	if it were to mean "shouldn't be used cause it doesn't mesh with the language"
19:04	mmarczyk	well, he's entitled to his opinion :-)
19:04	ataggart	meh, I don't care either way. IT's a tweet. Use monads if you want.
19:04	ataggart	there are more important things to get worked up over
19:05	etate	i was more just curious as to why he said it, rather than getting worked up
19:05	etate	as in i didn't know whether it was referring to monads in the implementation of clojure, or in daily usage or ?
19:06	mmarczyk	must be the impl given the 1.0-1.1-1.2 tags
19:06	etate	i thought maybe he had something else in mind than monads which solve the same problems
19:06	etate	but is more idiomatic
19:07	mmarczyk	yeah, like impure functions and ref types for mutable state :-)
19:07	etate	lol, not exactly what i had in mind :)
19:08	ataggart	I have hopes that some day I'll grok monads.
19:08	zakwilson	Lists, or seqs in Clojure are pretty obviously monads - just not explicitly so.
19:08	arohner	ataggart: a monad is just a set of rules for chaining together function calls
19:09	etate	yeah was just going to say that i thought let etc was basically monadic
19:09	ataggart	"monad" and "just" don't belong in the same sentence
19:09	ataggart	so comp is a monad?
19:10	etate	well, if stuart says there are no monads, i guess not?
19:10	ataggart	smells like monad is more adjective than noun
19:10	etate	:p
19:10	mmarczyk	monad is a way of making a more funky comp :-)
19:10	arohner	ataggart: sort of. monads have an "interface" for how to chain function calls
19:10	arohner	comp is a way of chaining together that doesn't follow the monad interface
19:10	arohner	huh. That would be an interesting blog article
19:10	arohner	monads for java programmers
19:10	arohner	define monads as java interfaces
19:12	zakwilson	I think the definition of return for Clojure seqs would be #(apply concat (map %1 %2))
19:12	zakwilson	err... bind
19:14	zakwilson	return is #(seq [%])
19:21	arohner	to the people who are having problems with swank-clojure + clojure-1.2-SNAPSHOT, try out my branch
19:21	arohner	http://github.com/arohner/swank-clojure
19:21	arohner	http://clojars.org/org.clojars.arohner/swank-clojure
19:21	sexpbot	"swank-clojure \| Clojars"
19:21	arohner	I had some problems, and then it magically started working, so I think it works, but I'm not 100%
19:22	arohner	if you have problems, tell me on github
19:27	technomancy	hang on, if you have trouble with my fork tell me too. =)
19:28	arohner	technomancy: you already have 1.2 fixes?
19:28	technomancy	arohner: I believe so, are you having issues?
19:28	arohner	technomancy: yeah, with #^ in swank source
19:28	technomancy	arohner: jochu's fork?
19:29	arohner	technomancy: yours
19:29	technomancy	interesting... I thought that had been cleaned up a long time ago.
19:30	arohner	it's not in master. maybe another branch?
19:31	arohner	ok, my branch is not working, but I have to run
19:31	technomancy	arohner: oh wait, #^... I was thinking of using ^ as meta
19:31	technomancy	which has been removed
19:31	technomancy	what's wrong with #^?
19:32	arohner	technomancy: http://github.com/richhickey/clojure/commit/a5ca8243e3ea45e6387b2b00d0e9b2624d16adbf
19:33	arohner	getting rid of old ^'s was the first step towards #^ being replaced with ^
19:33	arohner	I gotta run
19:33	technomancy	I think it's just that ^ and #^ become the same thing
19:34	technomancy	pretty sure #^ is not going anywhere
19:34	technomancy	since we haven't seen any deprecation warnings for it.
19:53	mmarczyk	technomancy: not yet
19:53	mmarczyk	http://clojure-log.n01se.net/date/2010-04-26.html#08:57
19:53	sexpbot	"#clojure log - Apr 26 2010"
19:53	technomancy	I hope that's pretty far down the line
19:53	technomancy	since it will make it impossible to be backwards-compatible
19:54	technomancy	maybe for 2.0?
19:54	mmarczyk	right... though ^String is so much prettier!
20:00	rhudson	Looks like the logbot is dead -- the last entry for today is about 7 hrs ago
20:10	alexyk	so... is ^ already the replacement for #^ in the trunk?
20:14	chouser	alexyk: yes
20:14	alexyk	nicee
20:51	alexyk	what's the shortest wy to test that ^ works instead of #^, in repl?
20:52	alexyk	way
20:52	mmarczyk	^#'+
20:52	qbg	Call meta on something read in using it?
20:52	mmarczyk	if the above doesn't work, then presumably ^String foo will
20:52	_ato	,'#^inc
20:52	clojurebot	EOF while reading
20:53	_ato	,'#'inc
20:53	clojurebot	(var inc)
20:53	mmarczyk	,(meta '^foo foo)
20:53	clojurebot	java.lang.Exception: Unable to resolve symbol: foo in this context
20:53	_ato	,'^inc
20:53	clojurebot	(clojure.core/meta inc)
20:53	mmarczyk	,(meta '#^foo foo)
20:53	clojurebot	{:tag foo}
20:53	_ato	,(= (name (first '^inc)) "meta") ; heh, ugly
20:53	clojurebot	true
20:54	alexyk	so, clojurebot knows about ^ already?
20:54	alexyk	~ how do you know anout ^?
20:54	mmarczyk	no, it still uses it to mean (meta ...)
20:54	clojurebot	Excuse me?
21:04	tomoj	anyone happen to be around who's made a custom assembly descriptor?
21:13	unlink	It seems that leiningen didn't expect the promotion of seq-utils to core.
21:14	unlink	That is, it will happily run an up-to-the-minute build of clojure with software which doesn't understand the changes introduced in 1.2.0.
21:24	unlink	How do I tag a function as returning void? (for reify)
21:32	eslick	When using 'lein swank' to launch a clojure process, the files in the /src directory are compiled, but the namespaces don't appear to be loaded
21:32	eslick	So I have to manually compile the files
21:33	eslick	Is there a method by which this should happen automatically?
21:37	leifw	any clojure-hadoop users around?
21:37	leifw	I need some help figuring out how to submit a jar
21:52	mmarczyk	eslick: actually you only have to (require ...) your namespaces at the REPL
21:52	mmarczyk	eslick: or (use ...) them, or do your work (in-ns your-namespace)
21:53	eslick	I do in-ns
21:53	eslick	But the namespace is not defined after connecting to clojure via swank
21:54	eslick	So the /src directory is properly on the jvm classpath
21:54	eslick	But none of the functionality in the .clj files are loaded
21:54	mmarczyk	seems to work for me
21:54	mmarczyk	with the latest (unstable) leiningen
21:55	eslick	How do I upgrade to unstable?
21:55	mmarczyk	you mean you can't call your functions?
21:55	eslick	Yes
21:56	mmarczyk	um, strange
21:56	mmarczyk	and which version of Lein are you using?
21:56	mmarczyk	the repo's here, btw:
21:56	mmarczyk	http://github.com/technomancy/leiningen
21:57	mmarczyk	see the README for instructions on how to set it up
21:57	eslick	Better to stick with the leading edge?
21:58	eslick	I was on 1.1.0
21:58	mmarczyk	well, no, I thought it was more reasonable to stay on 1.1
21:59	mmarczyk	if it does its job fine
21:59	eslick	Except of course it's not (although more likely it's operator error)
21:59	mmarczyk	but if it doesn't, the checkout way might be worth a shot
21:59	eslick	Been 5 years since I've used Java and I've got common lisp hardwired in my head
21:59	mmarczyk	classpath issues certainly are a pain
22:00	mmarczyk	not sure this is one of them, mind you
22:00	eslick	So if I have a file in src/com/project/core.clj
22:00	somnium	encountered a problem once where (swank-clojure-project) worked but lein swank + slime-connect didnt
22:00	eslick	with (ns com.project.core (:use ...))
22:00	eslick	When I connect, I should be able to do in-ns and then call the functions from that file, correct?
22:01	mmarczyk	somnium: do you know what the problem was?
22:01	slyphon	you have to require them
22:01	somnium	and at the repl (use 'com.project.core)
22:01	slyphon	eslick: (require 'com.project.core) first
22:01	mmarczyk	you should be able to do (in-ns 'com.project.core), if that's what you want
22:01	hiredman	eslick: in-ns just creates a namespace with the name you give it and puts you in it
22:01	mmarczyk	tried that a moment ago and it worked for me
22:01	hiredman	it doesn't load any files
22:01	eslick	So if I have a large system, I have to (require all the namespaces I may depend on?)
22:01	somnium	I have no idea, we kind of shotgunned it
22:02	mmarczyk	hiredman: I just started lein swank on Leiningen itself
22:02	mmarczyk	typed in (in-ns 'swank.swank)
22:02	mmarczyk	then I could do (start-repl) just fine
22:02	slyphon	eslick: i've found having a foo_init.clj file is useful, with the requires for a given level of the project
22:02	hiredman	mmarczyk: lein may load your stuff for you
22:02	mmarczyk	but not for eslick? :-)
22:03	hiredman	at work we gen-class something with a java main and do java some.class.here
22:03	slyphon	and then in user.clj i have a repl-init function that does requires of those _init.clj files
22:03	hiredman	mmarczyk: regardless of what lein may or may not do, I know what in-ns does
22:03	mmarczyk	hiredman: no argument about that
22:03	slyphon	:)
22:03	hiredman	mmarczyk: which was what I told eslick
22:04	eslick	Does require recursively load everything that the namespace so required depends on?
22:04	mmarczyk	hiredman: just pointing out that the thing it does should be sufficient for functions from foo.core to be available when lein swanking the foo project, after nothing more than (in-ns 'foo.core)
22:04	mmarczyk	eslick: yup
22:05	hiredman	well, it's not that require is recursive, it is that those files also call require
22:05	eslick	slyphon: where do you put the foo_init.clj file so it's run on startup?
22:05	eslick	Ah, of course.
22:05	eslick	It does take a bit to rebuild ones mental model of the world.
22:06	slyphon	well, for the repl, if you have a user.clj in the classpath, it's automatically required
22:06	eslick	Any clojure gurus in the bay area looking for full time jobs or a short term consulting contract? :)
22:06	slyphon	i dunno if that's true in other cases
22:07	mmarczyk	hm, I wonder if I'm wrong about lein vs. in-ns actually
22:08	mmarczyk	apparently I am
22:09	mmarczyk	sorry for introducing confusion then
22:10	leifw	I submitted a job to my hadoop cluster, as a jar, with main class clojure_hadoop.job and a mapper my.namespace/mapper-function
22:10	leifw	and I get a classnotfoundexception for my.namespace
22:10	leifw	helpz!
22:11	eslick	So the answer to my original question is 'require
22:11	hiredman	do you have a class named clojure_hadoop.job?
22:11	eslick	Thank you everyone for the help
22:11	eslick	!
22:11	leifw	hiredman: that's part of the clojure-hadoop library
22:11	eslick	And I was serious about the job opportunity...
22:11	leifw	I built this jar with lein uberjar, so it's finding that
22:12	leifw	it doesn't get the classnotfound until it tries to load the mapper function in my namespace
22:12	hiredman	leifw: and do you have a class named my.namespace?
22:13	hiredman	(if you don't that would be the reason you get the class not found exception)
22:13	leifw	no, I don't know why it thinks that's a class
22:13	mmarczyk	eslick: try posting your offer to the ggroup maybe :-)
22:13	leifw	oh hang on maybe I remember this
22:13	hiredman	…
22:14	eslick	mmarczyk: good suggestion. Lispjobs too
22:14	mmarczyk	right
22:14	leifw	riiiiiiiiiiight I have to add a :gen-class
22:14	leifw	no clue why this was working on my machine then
22:15	leifw	thanks, hiredman
22:15	leifw	oops
22:15	leifw	nope
22:15	leifw	:-/
22:15	leifw	I guess I'll bang on this later
22:20	chessguy	is there a typical way to raise/throw an error in clojure?
22:21	slyphon	clojure.contrib.except/throwf is pretty handy
22:24	_brian2_	noon question> what do I do with third party java jar files in leiningen projects?
22:24	rhudson	,(throw (RuntimeException. "yowza"))
22:24	clojurebot	java.lang.RuntimeException: yowza
22:24	slyphon	_brian2_: either install them locally, or make a clojars acct
22:25	_brian2_	slyphon: how do I install locally?
22:26	slyphon	maven will tell you when it can't find a dependency
22:26	slyphon	looks something like:
22:26	slyphon	mvn install:install-file -DgroupId=org.powermock.modules -DartifactId=powermock-module-testing -Dversion=1.3.7 -Dpackaging=jar -Dfile=/path/to/file
22:26	slyphon	(that one was due to a typo)
22:27	_brian2_	so I need to run maven?
22:27	slyphon	"run maven"?
22:27	slyphon	maven keeps a cache of stuff in ~/.m2
22:28	slyphon	it's like rubygems or CPAN
22:28	_brian2_	yes, I sort of familar with maven , but leiningen doesn't handle third party jar files by itself
22:29	slyphon	well, leiningen is a thin wrapper of awesome around maven
22:29	slyphon	at least in the "handles your dependencies" regard
22:30	_brian2_	well, if I have a jar file, I shouldn't need maven
22:31	_brian2_	its already built
22:31	_brian2_	I just want to call it in the leiningen project
22:33	alexyk	how do I make lein work with clojure/c.c. trunk?
22:36	lithper1_	when you run swank-clojure-project, is it possible for the project directory to be created, and all the proper files moved there in order to start a new REPL? for now, i patched swank-clojure.el to do it, but maybe there's a better way? i'm trying to avoid any massive build tools.
23:53	vu3rdd	/quit