#clojure – 2011-10-05

2011-10-05

00:03	popinigga	Hello
00:04	popinigga	I'm new to Clojure and I've written some code in a file helloworld.clj on my desktop.
00:04	popinigga	How do I run the code?
00:14	sean_corfield	I see #(some-fn % arg1 arg2 ..) quite often and wondered if there was some variant of something like partial that could remove the #(.. % ..) noise?
00:34	amalloy	(defn kinda-like-partial [f & args] #(apply f % args))?
00:34	amalloy	that is: i'm not aware of one, but of course you can write your own
00:37	amalloy	sean_corfield: ^
00:38	ibdknox	yeah, it's a good personal utility one
00:39	ibdknox	though, I think a good case could be made for it to be a function in core
00:39	ibdknox	I want such a thing all the time
00:39	amalloy	ibdknox: core is full to overflowing. if i were a meme-generator kinda guy, i'd be pasting you a link to an image full of something looking ready to explode
00:39	ibdknox	lol
00:40	ibdknox	yeah
00:40	ibdknox	I said a good case could be made, didn't say it should make it ;)
00:40	ibdknox	I mean we couldn't risk crowding things like with-bindings :p
00:40	ibdknox	that one still gets me
00:41	amalloy	ibdknox: with-bindings exposes functionality you can't get with plain-old binding
00:41	ibdknox	like?
00:41	amalloy	you don't always have var symbols; sometimes you only have var objects
00:42	amalloy	ibdknox: for example, https://github.com/cognitivedissonance/clojail/blob/master/src/clojail/core.clj#L172 - without with-bindings, this whole function would have to be a macro
00:42	ibdknox	is there a reason that binding couldn't also allow that?
00:43	ibdknox	ah
00:43	ibdknox	I misunderstood
00:43	ibdknox	fair enough
00:43	amalloy	how could it? you can't tell, at macro time, whether the symbol 'foo refers to the var #'foo, or the var referenced by the local foo
00:44	ibdknox	yeah, I misunderstood what you were saying
00:45	amalloy	mkay
00:45	ibdknox	that makes sense
00:46	ibdknox	I almost would've expected that to be like binding*
00:46	ibdknox	maybe that need arises much more than I think though
00:47	amalloy	ibdknox: no, i think it's a fairly uncommon need
00:52	brehaut	amalloy: that kinda-like-partial is called 'dispatch' in moustache
00:54	brehaut	sorry i lied
00:54	brehaut	delegate
00:54	brehaut	https://github.com/cgrand/moustache/blob/master/src/net/cgrand/moustache.clj#L172-176
00:56	michaelr525	is there a web browser built-in in emacs?
00:56	brehaut	theres a browser for emacs called W3 but i don't know if its built in
00:58	michaelr525	it's not in package-list-packages
01:04	no_mind	I have couple of files in a directory,. I want all these files to be part of a single namespace, identified by the directory path. For example, if the directory path is xyz/applications/accounts then all files in this directory should belong to namespace xyz.applications.account . How do I do this ?
01:11	michaelr525	no_mind: why do you want to do it like that?
01:12	no_mind	ease of managing packages
01:14	brehaut	no_mind: http://stackoverflow.com/questions/4690758/splitting-a-clojure-namespace-over-multiple-files
01:33	no_mind	brehaut: that doesnt work. I get an error saying "Could not locate xyz//applications/accounts__init.class or xyz/applications/accounts.clj on classpath" the directory xyz/applications/accounts is in the classpath
01:34	brehaut	no_mind: i have no idea sorry. i just remembered the link and thought it might be relevant
01:44	symbole	no_mind: I'm not too sure, but you might have to run compile on the class you're generating.
03:11	sivajag	Hey All
03:54	ejackson	morning folks
04:22	Blkt	good morning everyone
04:29	amalloy	anyone (like seancorfield) know if there's a way in congomongo to say "give me a document except for keys X and Y?" i know you can ask for only keys X and Y, but i don't see the opposite
04:30	pyr	i wanted to tryout msgpack rpc but it seems as though it relies on annotations
04:31	pyr	is there a way to provide annotations in clojure ?
04:31	amalloy	mongodb allows it with db.coll.find(query, {X: false, Y: false})
08:40	Fossi	hmmm
08:41	Fossi	battlenet? i'm intrigued
08:42	wink	I wouldn't b :P http://github.com/winks/battlenet
08:43	ejackson	aaaah, a black hole.
08:44	Fossi	wink: :)
08:44	Fossi	already found that
08:45	Fossi	look too WoWy for me ;)
08:46	wink	I'm just now wondering what slows a page request down to only 3req/s.
08:46	wink	probably not jetty, so maybe h2
08:46	wink	a bit of html generated by noir loads at 123 req/s
08:51	eirikb	Hello. I'm trying to clojure:test with the maven plugin, it finds my test but no deftest
08:51	edw	Is it possible to make a List<String> in clj? I'm working on a thrift server right now.
08:52	eirikb	I understand it will (run-tests) on the test
08:53	napping	edw: You should be able to make a List and fill it with Strings
08:53	edw	Will that 'satisfy' the interface's signature?
08:54	napping	generics are erased, so it should
08:55	edw	OK. Thank you.
08:57	napping	Is it not easy to test?
09:16	eirikb	Anyone know about any example projects with the maven plugin and some tests with larger namespaces (not just src/main/clojure/main.clj)
09:26	eirikb	I'm guessing I have messed up ns, but it still finds the file. Although it should (run-tests 'some.ns)
09:38	ejackson	eirikb: hmm.... I have made this work in the past. Lets see.
09:38	ejackson	Have you got the main and test directories properly set up ?
09:38	eirikb	ejackson: https://github.com/eirikb/comicparser/tree/clojure
09:39	eirikb	Don't mind the java-parts and the resources
09:39	ejackson	I'll poke around for you a bit...
09:39	eirikb	Thanks a lot
09:39	mcoffbyone	hi, could anyone recommend me some clojure library which does a lot of file/directories/regex/parse related stuff?
09:40	mcoffbyone	(need to learn how to do it elegantly not to use it)
09:40	ejackson	eirikb: ah, you're not using the clojure-maven-plugin
09:41	eirikb	ejackson: Using the first hit on google
09:41	eirikb	talios/clojure-maven-plugin
09:41	ejackson	that's the one
09:41	ejackson	about halfway down it talks about testing
09:42	eirikb	I took the pom setup from here: https://github.com/talios/clojure-maven-plugin
09:45	ejackson	eirikb: your pom doesn't have the plugin listed https://gist.github.com/1264459
09:45	ejackson	that gives gives you clojure testing in the test goal
09:46	eirikb	I do have the goal now though, and it find the file. It just does not run any deftest.
09:46	eirikb	I updated my pom, let me try
09:48	ejackson	you also need to put your tests into a test namespace
09:50	ejackson	so to test the namespace no.eirikb.comicparser.comicparser you need your tests to live in the namespace no.eirikb.comicparser.test.comicparser
09:51	ejackson	or at least that was the arrangement back when I was using maven
09:57	ejackson	eirikb: there's something up with your fork as well, if I try to fork off of it I only get the java :)
10:04	duck1123	ejackson: I believe they can live in any namespace in the test directory. The plugin shoul pick up all of them
10:11	ejackson	duck1123: OK, cool. I was just looking at a working project of mine and spewing forth.
10:12	duck1123	Personally, I prefer to name my test namespaces a.b.c-test
10:12	duck1123	in part because it makes it easier to tell the buffers apart in emacs
10:14	ejackson	fact !
10:24	gfredericks	using clojure-maven for the first time, repl is primitive. Anybody know what a man gots to do to get rlwrap or whatever hooked up?
10:28	ejackson	gfredericks: id suggest using swank and getting at the repl via emacs
10:28	gfredericks	ejackson: what about the persecuted minority of vim users? :)
10:29	algernon	run vim from emacs, and get the best of both worlds
10:29	ejackson	there's a nailgun plugin too
10:29	ejackson	(what exactly that means I don't know, but vim folk seem to like it)
10:29	gfredericks	algernon: vim from emacs works with swank and all that?
10:29	eirikb	ejackson: Did you clone the correct branch?
10:30	Guest66735	I started using vsclojure in vs2010 and I'm having a problem with my solution suddenly not compiling any more. if i check out the project again to a different location, it compiles, but eventually will fail repeastedly (BootstrapCompile has stopped working). Is this a known problem - file corruption or something?
10:30	eirikb	duck1123: But why are the tests not run?
10:30	ejackson	eirikb: must be the issue... you having any luck ?
10:30	ejackson	eirikb: have you explicitly tried mvn clojure:test
10:30	eirikb	ejackson: Nope
10:30	eirikb	Yes
10:31	Guest66735	btw, the exact same code compiles fine from command line or ccw in eclipse
10:34	eirikb	ejackson: It says: Testing no.eirikb.comicparser.comicparser \n\n Ran 0 tests containing 0 assertions.
10:35	ejackson	yeah, I'm having a look now
10:35	eirikb	Have to go for a short bit, be back in an hour
10:36	duck1123	eirikb: your test files are in src/test/clojure/ right?
10:36	duck1123	the clojure-maven-plugin creates a test runnier in /tmp might whant to check out what namespaces it picked up
10:36	eirikb	duck1123: Yes, https://github.com/eirikb/comicparser/tree/clojure
10:37	eirikb	duck1123: yeah, I can see it uses require
10:37	duck1123	you need to configure the plugin for tests
10:38	duck1123	let me see if I can find my old poms
10:39	eirikb	duck1123: Namespace configuration under https://github.com/talios/clojure-maven-plugin ?
10:39	eirikb	Seems to only be for limiting
10:40	eirikb	And additional sources should not be needed either, as it finds my file
10:40	duck1123	Here's one of the files I was using https://github.com/duck1123/jiksnu/blob/5a124e864e2a98089d14ae56a798fc851277ebcc/pom.xml
10:40	duck1123	so what happens when you do mvn clojure:test
10:41	eirikb	duck1123: https://gist.github.com/1264580 (the ns is probably wrong, but same with no.eirikb.comicparser)
10:42	eirikb	I see you have <execution>, same as ejackson had. It did not seem to help me
10:43	ejackson	eirikb: if I move your test into a test dir it finds it
10:43	eirikb	ejackson: src/test/clojure/test ?
10:43	ejackson	i sec and I'll commit it back up for you
10:43	eirikb	cool
10:44	eirikb	Ah, no.eirikb.comicparser.test.comicparser ?
10:45	eirikb	That seems to do the trick at least
10:45	ejackson	eirikb: yes that's the one
10:46	eirikb	ejackson: Thanks a bunch! you even mentioned this earlier but I didn't test it out, sorry :(
10:46	ejackson	no worries, enjoy.
10:47	vinibaggio	hello, I'm still beginning Clojure and I am having a hard time to grasp some concepts, such as doseq. Do anyone know (besides clojure's own doc) any resource that could help me with that?
10:47	eirikb	What version of the maven plugin is the one I should use? On the github page it was 1.3.4, while I see 1.3.8 is in the repo and duck1123 had 1.3.7
10:47	ejackson	eirikb: no idea I'm afraid.
10:48	ejackson	vinibaggio: I'd really suggest you get a book :)
10:48	eirikb	ejackson: ok =)
10:48	eirikb	ejackson: Tried updating to 1.3.8 to see if I still had to put it under a test-folder, and that I did
10:49	ejackson	yeah, maven is persnickity.
10:49	eirikb	hehe. Well it works perfectly now
10:50	vinibaggio	ejackson: I am trying to follow The Joy of Clojure, but it's being difficult to me, I have to admit. It assumes I've understood what it does
10:51	vinibaggio	ejackson: do you recommend any other?
10:52	ejackson	vinibaggio: yeah, JoC is not the point of departure, more the goal :)
10:53	ejackson	I really like Programming Clojure, by Halloway
10:53	vinibaggio	ejackson: cool to know. I've been enjoying the read, it's amazingly well written, but quite hardcore for a beginner hehe
10:53	ejackson	but the others are all good too
10:54	vinibaggio	ejackson: okay, I will start with that. Thanks!
10:54	ejackson	enjoy.
10:54	vinibaggio	:)
10:54	ejackson	be aware that its a bit out of date though
10:56	vinibaggio	hmm I think it may serve as a good start nevertheless. It doesn't cover Clojure 1.3, right?
10:57	vinibaggio	i think the problems I have are not very version-aware, but getting the whole LISP state of mind in my mind
10:57	duck1123	Programming Clojure covers 1.0, but I think there's a second edition in the works
10:58	rbranson	just keep trying on JoC
10:58	rbranson	it's dense, but very rewarding
11:03	gfredericks	duck1123: I was just looking at a pdf of the second edition, coincidentially
11:04	vinibaggio	hmm ok, I think I will fallback a little bit on Programming Clojure and try to follow JoC after I have a better grasp on the language. Thanks guys
11:04	gfredericks	one of the beta versions I guess
11:05	rbranson	I have no previous LISP experience and found JoC to be a fantastic introduction
11:06	duck1123	so, it you bought the first edition, you have to pay for the second ed?
11:07	vinibaggio	rbranson: yeah, I am enjoying it, but i am having trouble to have a complete understanding of it. sometimes i feel i am leaving something out
11:08	rbranson	vinibaggio: it's a relatively challenging book
11:08	vinibaggio	rbranson: that's awesome :)
11:09	rbranson	vinibaggio: I found that it helped to stop and actually start hacking through some 4clojure to get my "watermark" up so that I really understood the previous material
11:09	vinibaggio	rbranson: ah yes, i am trying to do that. I am solving some project euler problems, and also spoj.pl
11:09	rbranson	no pain, no gain :)
11:10	ejackson	you guys are on the right track, as the rapper laureate himself said, "Lisping aint easy"
11:10	rbranson	i have been pleasantly surprised by the challenge, usually the challenge in learning a new language is learning the tools/environment
11:10	vinibaggio	rbranson: what was troublesome to me is that i am trying to read input from STDIN, and the example I've found used doseq, and I didnt understand what it does, and I, as a total beginner, am not understanding how to work with it
11:11	vinibaggio	rbranson: exactly! that's how i feel about JoC, and that's why i am loving reading it
11:11	rbranson	vinibaggio: i don't really fully grasp the lazy-seq / doseq stuff yet either
11:12	dnolen	vinibaggio: doseq is simply a way to write imperative code on sequences.
11:12	gtrak	for me, the challenge is it feels like I'm doing less work, since it's less keypresses, and you can maybe express a few java classes worth of code in a couple of lines of clojure, but the trick is to know the best way to express it
11:13	rbranson	gtrak: yeah, clearly there's the "silly unwashed clown" way of solving problems :)
11:13	dnolen	vinibaggio: (doseq [x ...] (draw-to-screen x)) as opposed to (map inc xs) which returns a new sequence
11:14	gfredericks	dnolen: and importantly map is lazy so it won't necessarily do anything
11:14	rbranson	I don't feel like doseq is super duper important at the beginning stage, most of the time you can build on-top of the library
11:14	dnolen	gfredericks: that too.
11:14	rbranson	recur feels really important though
11:14	gfredericks	not so important is the rarely-used do-not-seq
11:15	rbranson	donut-seq?
11:15	vinibaggio	dnolen: hmm, okay, I will try to play with it here, thanks
11:15	gfredericks	doughnut-seq
11:15	gfredericks	rbranson: my experience in #clojure is that beginners often use recur a lot more than they need to
11:15	rbranson	gfredericks: i agree
11:15	gfredericks	when map or reduce would do what they need
11:15	rbranson	gfredericks: ignorance of the core library
11:16	gfredericks	coming from another language you're used to having to do algorithms by hand
11:16	gfredericks	so loop/recur seems like the best way to do nearly anything
11:16	rbranson	recur is a special form though, so i feel it's worth knowing
11:17	Chousuke	recur is fine, but loop tends to be a code smell
11:17	gfredericks	rbranson: just like monitor-enter? :)
11:17	rbranson	gfredericks: :)
11:17	gfredericks	Chousuke: you mean in contrast to recuring to a fn?
11:17	Chousuke	yes
11:18	Chousuke	lots of newbies using loop just think they want a looping construct
11:18	Chousuke	so they end up using loop
11:19	gtrak	maybe we should change the name of loop to something less recognizeable
11:19	rbranson	I think people make the same mistake in learning python/ruby
11:19	gfredericks	I was about to say it sounds like Chousuke just doesn't like the name
11:19	Chousuke	nah, that's just obfuscation
11:19	rbranson	so it's not clojure's fault ;)
11:19	Arafangion	rbranson: Python, no, ruby, yes.
11:19	gfredericks	Chousuke: what kind of refactoring do you do when you see a loop?
11:19	Chousuke	gfredericks: the name is irrelevant. it's just what the newbie will find when they search for loops :P
11:19	zerokarmaleft	vinibaggio: restricting yourself to working inside the REPL will help you get used to thinking about functional composition
11:19	gtrak	maybe people could say... i want to loop, but there's no loop keyword, how the f do I loop?
11:19	Arafangion	rbranson: The right way to loop /is/ to do a loop in python.
11:19	gtrak	ah... maybe map does what I need
11:20	Chousuke	gfredericks: I'm not saying all loops are bad but you can get away without an explicit one in most cases
11:20	rbranson	Arafangion: list comprehensions replace a good number of cases though
11:20	gtrak	maybe instead of loop, call it label
11:20	rbranson	Arafangion: I see what you're saying though, list comprehensions are kind of a special case
11:20	Arafangion	rbranson: Indeed.
11:21	Chousuke	gfredericks: that's why it's always somewhat suspicious if you see lots of looping
11:21	Arafangion	rbranson: Also, note that recursion is Bad in python.
11:21	gtrak	label-recur makes people think of gotos, and makes them feel worse about it
11:21	rbranson	haha
11:21	vinibaggio	Chousuke: I felt this way, I build a fibonacci totally with loops, and almost every implementations I've seen on clojure does not use loops nor recursions (at least in a clear manner). Maybe it would be nice for beginners to have some sort of comparison, and how to achieve that?
11:21	dnolen	Chousuke: gfredericks: unless of course it's a tight inner loop ;)
11:22	Chousuke	vinibaggio: maybe
11:22	rbranson	vinibaggio: I feel like some of those become "golfing"
11:22	rbranson	there's definitely a threshold where code clarity takes a backseat to compactness
11:23	Chousuke	vinibaggio: the main issue is learning to use laziness
11:23	vinibaggio	Chousuke: i think so
11:23	gfredericks	and parentheses
11:23	arohner	I'm seeing an issue where noir is listening on ipv6 8080, but not ipv4 8080. Any ideas?
11:23	Chousuke	people with no functional background think in loops so using lazy sequences is something they can't even consider
11:24	Chousuke	until they encounter it somewhere
11:24	rbranson	are lazy seqs pretty much the norm?
11:24	zerokarmaleft	rbranson: coming from ruby, i find code clarity in lisps have a different paradigm...the shape of the code matters more than names of symbols
11:24	ejackson	I want to create a macro that does does some defs. I want to use that macro in various other namespaces. Clojure says "can't def in another namespace" Is there a workaround?
11:24	clojurebot	macro are just a game with symbols
11:25	ejackson	thanks clojurebot
11:25	Chousuke	rbranson: Clojure provides you with a whole library of dealing with sequences, so if you use sequences, you get lots of useful tools for free.
11:25	rbranson	Chousuke: right, but i'm talking about idiomatic code
11:25	arohner	ejackson: with-ns, in contrib
11:25	Chousuke	whereas looping always requires manual work.
11:25	ejackson	arohner: thanks, i'll check it out
11:25	Chousuke	sometimes you need a loop for performance reasons but yeah lazy seqs are almost always more idiomatic
11:26	gtrak	laziness should be invisible unless you're dealing with side-effects
11:26	raek	ejackson: maybe you are using syntax-quote for the def name.
11:26	clojurebot	defmulti doc is (defmulti #^{:doc "docs for foo"} foo class)
11:27	raek	, `(def foo)
11:27	clojurebot	#<Exception java.lang.Exception: SANBOX DENIED>
11:27	ejackson	raek: yeah, I am... ha ! let me try do it raw. thanks.
11:27	gfredericks	ejackson: I've done what you described and had no trouble. I think I was using regular backquote: `(defn ~thing-name ...)
11:27	raek	ejackson: you don't want the name you pass to be qualified with the namespace that the macro was defined in
11:28	raek	*the name you pass to 'def'
11:29	vinibaggio	this was a pretty nice discussion guys, thanks, learned a ton :)
11:29	raek	you can use `(def ~'my-name ...) to circumvent syntax-quote
11:29	ejackson	raek, yeah, gfredericks, I think I understand you guys. Lemme play a bit, thanks.
11:29	gfredericks	ejackson: when in doubt, sprinkle punctuation liberally
11:30	ejackson	:D
11:30	gfredericks	`~'~~('def ``~'puddle-ducks @@~'`&args)
11:30	gfredericks	^ expands to whatever you need your code to do
11:31	ejackson	raek: Winner, winner, chicken dinner.
11:31	TimMc	throw in some # just to make sure
11:32	ejackson	gfredericks: you're a perl exile i see.
11:32	gfredericks	ejackson: only in jest
11:33	zerokarmaleft	gfredericks: Q(`_')-o
11:33	zerokarmaleft	recognize!
11:33	gfredericks	that's just an emoticon, right?
11:33	gfredericks	maybe for something obscene?
11:34	zerokarmaleft	fairly certain it's not valid perl, but i don't think it's obscene
11:34	zerokarmaleft	s/but/and/
11:34	lazybot	<zerokarmaleft> fairly certain it's not valid perl, and i don't think it's obscene
11:35	Arafangion	If it were perl, then it would be obscene.
11:36	gfredericks	fun fact: Every obscenity is a valid perl expression, and vice versa.
12:11	bsod1	I want to create an NxN grid with vectors, what is the idiomatic way to do this in clojure?
12:17	eblood	i don't know the idiomatic way, but a quick way to create a seq of seq: (repeat n (repeat n 0))
12:23	zerokarmaleft	bsod1: (for [x (range n) y (range n)] [x y])
12:28	zerokarmaleft	eh, that should be (for [y (range n) x (range n)] [x y]) if you want them ordered by row
12:29	TimMc	zerokarmaleft: I don't think that's correct.
12:29	TimMc	That gives a seq of coordinates, not a 2D array (matrix).
12:30	zerokarmaleft	ah, i read his question as a NxN grid with a vector at each (N, N)
12:31	TimMc	,(for [y (range 2) x (range 3)] [x y])
12:31	clojurebot	([0 0] [1 0] [2 0] [0 1] [1 1] ...)
12:31	TimMc	(repeat 2 (repeat 3 0))
12:31	TimMc	,(repeat 2 (repeat 3 0))
12:31	clojurebot	((0 0 0) (0 0 0))
12:34	raek	,(for [y (range 2)] (for [x (range 3)] [x y]))
12:34	clojurebot	(([0 0] [1 0] [2 0]) ([0 1] [1 1] [2 1]))
12:35	zerokarmaleft	TimMc: how would you access the value at x,y for that matrix?
12:36	PPPaul	what is the proper way to execute some side-effect code on a list? i'm doing (doall (map))
12:36	raek	,(vec (for [y (range 1 4)] (vec (for [x (range 1 4)] (+ (* y 10) x)))))
12:36	clojurebot	[[11 12 13] [21 22 23] [31 32 33]]
12:36	raek	,(let [matrix (vec (for [y (range 1 4)] (vec (for [x (range 1 4)] (+ (* y 10) x)))))] (get-in matrix [1 1]))
12:36	clojurebot	22
12:37	raek	PPPaul: if you don't care about the return value of the function with side-effects, you should use dorun
12:37	raek	(dorun (map println ["foo" "bar"]))
12:37	PPPaul	ok
12:37	raek	doseq is also very useful
12:38	PPPaul	using map, instead of some other construct, is still considered ok?
12:38	raek	(doseq [x coll] ...) = (dorun (for [x coll] (do ...)))
12:38	raek	sure
12:38	PPPaul	for... i still don't understand for >_<
12:39	raek	it's like map, but with swapped arguments and an implicit anonymous function...
12:39	PPPaul	hmm
12:39	raek	(map (fn [x] (+ (inc x) (dec x))) [1 2 3 4])
12:39	PPPaul	i've seen for do magic
12:40	raek	(for [x [1 2 3 4]] (+ (inc x) (dec x)))
12:40	PPPaul	ooooooooh
12:40	raek	I usually pick for when I find myself writing (fn ...) in a map
12:40	PPPaul	i'm beginning to understand it
12:40	seancorfield	amalloy_: just saw your Q about congomongo... no idea so i'll go take a look
12:40	raek	it looks much prettier for bigger fns
12:41	raek	doseq has the same syntax as for
12:42	raek	but also allows multiple expression in the body (as if there was an enclosing 'do')
12:42	crazyFox	hi. i am dealing with a group of entities that all have the same kinds of properties. first approach: using a record data type. downsides: every access of some property requires repetitive code to get to the atomic values; slower. second approach: have a separate vector for each kind of property. downside: the different properties of an entity are only linked via their positions in the...
12:42	crazyFox	...vectors. Any thoughts about the matter?
12:44	TimMc	zerokarmaleft: get-in
12:44	duck1123	crazyFox: proper use of destructuring should take care of a lot of the repetitive field access for maps / records
12:44	TimMc	But we haven't heard back from bsod1 about the specific requirements.
12:44	seancorfield	amalloy_: nope, it forces { fieldname: true } but it would be a nice enhancement to let it take a map of fields instead of just an array and then you could set true / false for selecting / omitting fields
12:44	seancorfield	i'll open an issue for that
12:45	zerokarmaleft	raek: is your matrix snippet still lazy?
12:46	crazyFox	duck1123: right now it looks sth like this: (let [xs (map :key entities) ...] ...)
12:46	raek	zerokarmaleft: the 'vec' calls forces the sequences returned by for
12:46	zerokarmaleft	that's what i suspected
12:46	TimMc	PPPaul: `for` can also do filtering and other fancy stuff.
12:46	raek	but if you leave out the 'vec' calls you have a lazy sequence of lazy sequences
12:47	PPPaul	!
12:47	gilescb	hi all, in core.logic is there something equivalent to prolog's retract?
12:48	duck1123	crazyFox: ok, if you're trying to key the :key field from every entry, that's probably the best you're going to get. But you'll probably need map no matter what
12:49	raek	(for [x (range 10), :let [y (* x x)], :when (even? y)] ...)
12:49	duck1123	if you're parsing them one record at a time, you can use destructuring to populate the variables for that record
12:50	crazyFox	duck1123: the thing is, i often need one property from all the entities at once
13:52	jjido	,(char 33)
13:52	clojurebot	\!
13:58	gfredericks	,(int \!)
13:58	clojurebot	33
14:00	jjido	thanks g. I wanted the next char after space and couldn't find the syntax in lua
14:00	jjido	or python
14:01	jjido	(on my iPhone)
14:02	gfredericks	you mean they let people program without having memorized ASCII?
14:02	jjido	lol
14:09	gfredericks	I'm not letting my kid go to any school that doesn't teach ASCII alongside the alphabet.
14:10	tordmor	and at least superficial knowledge of the more important unicode areas
14:11	gfredericks	if I pick any unicode character, he ought to have at least _seen_ it by third grade.
14:11	scgilardi	bmp: it's not just for bitmaps any more.
14:11	gfredericks	scgilardi: what's it for?
14:11	jjido	lol again
14:12	scgilardi	gfredericks: the basic multilingual plane
14:12	gfredericks	scgilardi: ah yes.
14:13	gfredericks	I just read on the relevant wikipedia article: "Plane 3, the Tertiary Ideographic Plane (TIP), is reserved for Oracle Bone script, Bronze Script, ..."
14:14	gfredericks	and clicked on the "Oracle Bone" link wondering what weird oracle project it was
14:14	gfredericks	and why some dumb piece of proprietary technology got its own swath of characters
14:15	jjido	you should know 'ideographic'
14:19	cb`	Hi, hope you can help. I'm trying to setup Clojure on Emacs and am following the instructions on http://dev.clojure.org/display/doc/Getting+Started+with+Emacs. I'm on Windows 7
14:21	cb`	Kind of like Joshua Wood's comment at the bottom of that page I get an error when trying to run M-x clojure-jack-in which says: "sh: line 1: lein jack-in 63783: command not found". However lein is on the PATH. I can run lein from a running eshell without issue. Any ideas?
14:21	technomancy	cb`: running clojure-mode from git or marmalade?
14:22	hiredman	cb`: on osx by any chance?
14:22	cb`	technomancy: marmalade
14:22	cb`	hiredman: Windows 7
14:23	hiredman	ah, well, interesting choice there
14:23	technomancy	cb`: I just made some commits in git that should at least make it more clear what the problem is; maybe try that?
14:23	technomancy	if it works on windows I'll push the release to marmalade
14:24	cb`	technomancy: Alright, thanks. Will give it a shot and report back here.
14:24	cb`	technomancy: Best to remove using package.el first or it'll pick up the latest version of clojure-mode regardless?
14:25	technomancy	cb`: yeah, try that
14:25	technomancy	and restart emacs since it involves a defvar change
14:28	bsod1	I can't import clojure.contrib in my counterclockwise project, anyone know why? => clojure.contrib/priority-map #<CompilerException java.lang.ClassNotFoundException: clojure.contrib (NO_SOURCE_FILE:0)>
14:29	gfredericks	bsod1: I didn't think clojure.contrib was ever a namespace...
14:29	amalloy	it ain't
14:30	bsod1	so how can I use clojure.contrib.priority-queue? sorry I'm new at clojure
14:30	gfredericks	(use 'clojure.contrib.priority-queue)
14:30	gfredericks	or (require '[clojure.contrib.priority-queue :as pq])
14:30	bsod1	gfredericks: #<CompilerException java.io.FileNotFoundException: Could not locate clojure/contrib/priority_queue__init.class or clojure/contrib/priority_queue.clj on classpath: (NO_SOURCE_FILE:0)>
14:31	amalloy	seancorfield: what's the canonical congomongo repo? i'll have a look at permitting maps
14:32	gfredericks	bsod1: I don't see docs for priority-queue in clojure-contrib -- where are you seeing it?
14:32	amalloy	(i'm just going to assume it's aboekhoff's github repo)
14:32	cb`	technomancy: Now swank buffer looks alot better. There are lot of lines in there and last one is: ";;; proceed to jack in"
14:33	technomancy	cb`: but no repl?
14:33	cb`	technomancy: However, Messages has: "error in process filter: eval-buffer: Symbol's function definition is void: define-slime-contrib" and yes, no repl
14:33	technomancy	cb`: you've got some slime configuration that only works with CL slime
14:34	bsod1	gfredericks: yeah I was looking wrong place, I think right one is this. https://github.com/clojure/data.priority-map
14:34	bsod1	still I don't understand how can I use this
14:35	gfredericks	bsod1: it'd be (require '[clojure.data.priority-map :as pm])
14:35	gfredericks	where 'pm' is whatever prefix you'd like to use
14:35	bsod1	gfredericks: same error: #<CompilerException java.io.FileNotFoundException: Could not locate clojure/data/priority_map__init.class or clojure/data/priority_map.clj on classpath: (NO_SOURCE_FILE:0)>
14:35	gfredericks	bsod1: It has to be included in your project as a dependency
14:36	gfredericks	how you do that with CCW, I don't know. Maybe build the jar and include it on the classpath somehow? Someone else may be able to comment better.
14:36	gfredericks	if you're using leiningen too, it shouldn't be hard
14:36	gfredericks	no idea if leiningen and CCW go together
14:36	cb`	technomancy: a hah.. OK, I commented out any previous slime related stuff from init.el and fired up another Emacs instance and I get a REPL with M-x clojure-jack-in!
14:37	technomancy	cb`: great! is the config you removed important?
14:37	cb`	technomancy: Do you know if it's possible to shield the CL slime config from Clojure? It looked to me like clojure-jack-in used its own copy of Slime anyway?
14:38	technomancy	cb`: I'm not sure; I don't use common lisp
14:38	amalloy	anyone know if type-hinting helps the compiler with protocol lookup? like, (extend-protocol SomeProto MyType (impl [this args] ...)) ... (impl ^MyType foo "test")
14:38	cb`	technomancy: It's quicklisp's version of Slime. I'd have to investigate to figure out what part is causing it to brek. So far it "just worked"
14:38	technomancy	this comes up periodically, but none of the CL users have gotten back to me about a reasonable fix
14:38	technomancy	cb`: what's the offending config look like?
14:38	amalloy	it seems to me like it shouldn't help, but i don't know everything about what's going on in protocols
14:39	cb`	technomancy: I'll probably join that list of offending users as I'm not a CL expert either. Just dabbling in CL, Scheme and Clojure for now. But I will try and determine what subset of this config is causing the problem.
14:42	cb`	technomancy: It appears to be (slime-setup '(slime-fancy))
14:43	cb`	technomancy: That's a configuration for slime that tells it to use several of the common contrib packages. Perhaps cojure-mode just needs to grab slime contrib as well as slime when it's running clojure-jack-in?
14:45	technomancy	that's one way to do it
14:45	bsod1	argh, I can't use clojure.data.priority-map in la clojure too... it shouldn't be that hard.. I've already have clojure-contrib.jar imported in my project
14:46	technomancy	slime-setup-contribs could also just go inside an ignore-errors
14:47	cb`	A related question is, is it possible to get clojure-mode to save its copy of Slime somewhere? It seems to be downloading it each time and in that case downloading the contrib directory would obviously take a bit longer. I wouldn't really mind an extra bit of download time if it was a one off
14:47	technomancy	cb`: sure, swank-clojure 1.3.3 does that
14:47	technomancy	(it's not downloading it from the 'net, just unpacking it out of the jar though)
14:48	cb`	oh, cool. I didn't realize it was coming from the jar. That's actually not so bad. I was concerned about trying to kick it off when not connected to the net. A few seconds of loading doesn't bother me too much
14:49	technomancy	yeah, but it's ugly
14:49	cb`	is 1.3.3 good to use or should I stick to 1.3.2 for other reasons?
14:49	technomancy	1.3.3 is better
14:51	cb`	I don't notice a lein upgrade option. If I just ran "lein install swank-clojure 1.3.3" would that take of it?
14:51	technomancy	I think so
14:51	technomancy	you may need to uninstall the old one; I'm a bit fuzzy on how that works
14:52	cb`	is adding slime-setup-contribs inside ignore-errors a quick fix? I'd be happy to try it out and confirm if it works here
14:53	technomancy	should be; can you give it a try?
14:53	technomancy	if that works I'll add it to the swank readme
14:54	bsod1	can anyone help me? I created a lein project, added clojure 1.3 and clojure.contrib 1.2 to deps, run lein deps, run lein repl, and still can't use clojure.data.priority-map
14:57	ibdknox	bsod1: contrib won't work with 1.3, and I don't think there's an official release of clojure.data yet
14:57	technomancy	bsod1: priority map is not part of contrib 1.2; it's a separate project
14:57	technomancy	clojurebot: what happened to contrib?
14:57	clojurebot	excusez-moi
14:57	technomancy	clojurebot: where did contrib go?
14:57	clojurebot	well... it's a long story: http://dev.clojure.org/display/design/Where+Did+Clojure.Contrib+Go
14:58	cb`	For what it's worth the list of things in slime-fancy is: slime-repl, slime-autodoc, slime-c-p-c, slime-editing-commands, slime-fancy-inspector, slime-fuzzy, slime-presentations, slime-scratch, slime-references, slime-package-fu and slime-fontifying-fu. I think at least some of those are used/working in clojure-mode.
14:59	technomancy	cb`: yeah, that could possibly be added to the bootstrapped elisp in the jar
14:59	technomancy	but right now it loads everything; there's no way to configure which you want
14:59	bsod1	ibdknox: so how can I use this priority-map?
14:59	technomancy	maybe a separate swank-clojure-fancy jar could be used for that
15:00	ibdknox	bsod1: [org.clojure/data.priority-map "0.0.1"] it appears
15:03	sritchie	cemerick, have you had any trouble getting the clojure view server started (with clutch) in 0.2.5-SNAPSHOT?
15:03	cemerick	sritchie: no; though, 0.2.5 was released last week IIRC
15:03	cemerick	I saw your message; you probably have a classpath issue in your view server config.
15:04	sritchie	cemerick got it, I'll keep playing. The server command works fine when I copy it into the command line, it just fails when couchdb tries to run it
15:04	cemerick	ah, that's interesting
15:05	cemerick	Well, try 0.2.5 release. If you're still having issues, ping me (though probably not 'til next week? :-)
15:06	sritchie	cool, this is the homebrew couchdb, maybe I'll try couchbase or something
15:06	sritchie	1.1.0. thanks for taking a look, I'll ping you next week
15:13	cemerick	sritchie: I do all my testing with CouchDBX or couchbase single, FWIW.
15:14	cemerick	I'm seeing some very odd behaviour out of clojurescript macros.
15:14	cemerick	Which probably means I'm not using them properly. :-P
15:14	cemerick	a simple macro: (defmacro emit [k v] `(- k v))
15:15	cemerick	when used, it produces javascript like: com.K.N.M.O - com.K.N.M.P
15:15	cemerick	neither of which exist at runtime
15:16	chouser	cemerick: are you pulling it it with use-macros or require-macros?
15:16	cemerick	require-macros
15:16	chouser	hm. dunno why that wouldn't work.
15:17	cemerick	I'm just baffled by where those names are coming from.
15:17	cemerick	or, not 'just', perhaps 'additionally'
15:17	chouser	you're using gclosjure advanced?
15:17	cemerick	yup, {:optimizations :advanced :pretty-print true}
15:17	cb`	technomancy: I'd not tried to use ignore-errors before. I tried wrapping (slime-setup '(slime-fancy)) in my config with (ignore-errors ...) but it didn't seem to help. I guess because the error doesn't happen when emacs starts up anyway, it's later when clojure-jack-in runs. I could be using it wrong though
15:18	cemerick	what a fool
15:18	cemerick	chouser: I didn't put tildes in the macro body
15:19	cemerick	good grief, Charlie Brown.
15:20	amalloy	chouser: any idea what's up with data.xml? it looks like the indentation test you wrote is failing on jira, but it works when i run tets locally so i'm not sure how i can get it building
15:21	cemerick	hey all: I'm speaking on (experimentally!) using ClojureScript to build CouchDB views, tomorrow night at the Bay Area Clojure Meetup: http://bit.ly/pFjdVK
15:21	cemerick	seancorfield: you asked about my topic :-) ^^
15:21	amalloy	er, on hudson. or jenkins or whatever the dang build server is
15:22	chouser	amalloy: sorry, I still haven't given data.xml the love it needs.
15:23	amalloy	if it will help i can get out a pitchfork or other threatening-looking device
15:25	chouser	the xml emit stuff relies a lot on whatever xml processing classes are loaded into the jvm instance.
15:25	chouser	that might be related to whatever's causing it to fail.
15:25	chouser	what's the jenkins url?
15:27	amalloy	chouser: http://build.clojure.org/job/data.xml/ is in the general area
15:28	amalloy	chouser: and http://build.clojure.org/job/data.xml/lastUnsuccessfulBuild/console has the test output
15:28	chouser	yeah, it's clearly not indenting it
15:28	chouser	isn't it.
15:29	amalloy	indeed. but when i run the test myself, it passes
15:30	seancorfield	amalloy: thanx for the patch for congomongo - merged and i'll post an updated snapshot build in a few
15:30	amalloy	perhaps jenkins and i have a different instance of (javax.xml.transform.TransformerFactory/newInstance)
15:30	chouser	amalloy: yeah, probably. ugh.
15:34	chouser	I don't know if that can be forced for hudson, or at least reported on the console?
15:34	PPPaul	in slime what can i press to kill the current operation?
15:35	amalloy	chouser: http://stackoverflow.com/questions/1264849/pretty-printing-output-from-javax-xml-transform-transformer-with-only-standard-ja seems to be suggesting you can also try setting an attribute on the factory
15:36	amalloy	but it's equally implementation-dependent what properties the factory supports
15:37	amalloy	as for reporting on the console, you can print out the class of the transformer you get back, right?
15:41	chouser	amalloy: yeah, I'd think so.
15:53	kencausey	Why for lein deps (in this case for cascalog) do I see messages indicating that each jar is downloaded 3 times, once each for conjars, clojars, central?
15:55	kencausey	Should I assume no actual download occurs for each message unless I see a comparable Transferring message?
15:55	hiredman	kencausey: I believe those messages are for each attempt to download, it looks for jars from each repo
15:55	kencausey	I see, so it should really be a 'Looking for ... in ...'
15:56	kencausey	thanks
16:28	cb`	technomancy: I'd not tried to use ignore-errors before. I tried wrapping (slime-setup '(slime-fancy)) in my config with (ignore-errors ...) but it didn't seem to help. I guess because the error doesn't happen when emacs starts up anyway, it's later when clojure-jack-in runs. I could be using it wrong though
16:29	technomancy	cb`: hm... I guess what you really want is to defer that execution until you're connecting and you know it's a CL instance
16:29	cb`	I did however add (setq slime-setup-contribs nil) to the top of clojure-jack-in's definition and that worked (though of course with the side effect of wiping out part of your slime config if you do run it later)
16:29	technomancy	well, you can't go from jack-in to a CL slime session without restarting emacs anyway, so that shouldn't be a problem
16:30	cb`	How come it uses its own version of slime? Is it just to reduce the amount of things a user has to install?
16:31	technomancy	cb`: CL makes you install slime from CVS
16:31	technomancy	two reasons this is infeasible for clojure:
16:31	technomancy	0) my conscience won't allow me to suggest users install CVS and 1) slime from CVS is always breaking, and I don't have the time or inclination to fix it.
16:32	cb`	it's on ELPA now, or marmalade (I'm not sure how to get Package Menu to show the source of a particular package)
16:33	technomancy	the marmalade version is not compatible with CL
16:33	cb`	with some particular CL? It seems a bit redudant to have it there if it doesn't work
16:34	technomancy	it's there for Clojure
16:34	technomancy	bundling it via jack-in is a pretty recent development
16:34	technomancy	however, you can use swank-clojure with marmalade's slime, it's just more manual and error-prone
16:34	technomancy	useful if you're connecting to a swank server in production or something though
16:35	cb`	ok
16:35	cb`	well, the git version of clojure-mode worked better for me on Windows 7 than the one on marmalade currently.
16:36	technomancy	they're now the same version
16:36	technomancy	M-x package-refresh-contents to get your source list updated
16:36	cb`	I have a work around for the 'slime-fancy thing but do wonder if it might make sense to allow that to work automatically (and/or even support some of those things anyway).
16:38	technomancy	open to improvements there
16:39	cb`	Are there instructions anywhere for using swank-clojure with another copy of slime? Like the one that comes with marmalade. I'd quite to try using it with my quicklisp copy of Slime and current slime config and see if it "just works" when it has access to the contrib directory and all the .el files in there
16:40	srid	is there a websocket example using clojurescript?
16:40	technomancy	cb`: sure; the "Alternate Usage" section of the readme shows how to use it with M-x slime-connect and lein swank
16:41	duck1123	srid: I don't know if the Closure library has been updated since Clojurescript first came out, but the Websocket lib wasn't in the release
16:41	ibdknox	srid: the first result on google for clojurescript websocket is useful
16:42	srid	duck1123: https://github.com/clojure/clojurescript/blob/master/src/cljs/clojure/browser/net.cljs uses websocket; just wondering if there is any code that uses clojure.browser.net
16:43	srid	ibdknox: http://lab01.kungfoo.pl:8108/ ? it doesn't work for me. hitting ENTER doesn't do anything (doesn't send the message)
16:43	ibdknox	https://github.com/neotyk/ws-cljs/blob/master/src/cljs/websocket.cljs
16:43	ibdknox	srid: there's a link to the source at the bottom, but that link above is what you want
16:43	srid	... doesn't work only in chrome
16:44	srid	ibdknox: hmm ok. that source doesn't use clojure.browser.net though. but i'll copy paste it in my project
16:44	ibdknox	srid: clojure.browser.net didn't exist in master until about two weeks ago
16:45	ibdknox	so I suspect you won't find much that uses it quite yet
16:49	crazyFox	fastest (execution) way to make / return a double value? (/ a (double b))? (i know for numbers it would be (/ 2 3.0) but im dealing with vars, so...
16:52	srid	looks like I can't import goog.net.WebSocket at all -- ERROR - required "goog.net.WebSocket" namespace never provided
16:52	hiredman	,(/ 1.0 2 3)
16:52	clojurebot	0.16666666666666666
16:53	hiredman	,(/ 2 3 1.0)
16:53	clojurebot	0.6666666666666667
16:53	zippy314	has anybody successfully used cookies in clojurescript with goog.net.cookies? I keep getting this error: 'this.isValidName' [undefined] is not a function when I try to set a cookie via goog.cookies.set.
16:55	cb`	technomancy: it didn't quite work. Probably incompatible versions of slime and swank (I wasn't using marmalade slime I was using the one that quicklisp installs). M-x slime-connect worked but then I got exceptions from the "lein swank" process when I tried to execute anything. One thing I did notice though was that at least some of the slime-fancy things must have worked. One of them puts a title bar at the top of the REPL very similar
16:55	cb`	to the top of an ERC buffer and that appeared to work fine.
16:55	ibdknox	srid: ./script/bootstrap --closure-library-head
16:56	technomancy	cb`: yeah, people have reported problems with that before
16:56	ibdknox	ah actually
16:56	ibdknox	that wasn't added
16:58	ibdknox	I think the goog-jar.jar that I packaged for Pinot has websockets in it
16:58	TimMc	,((juxt identity double) (/ 2 3))
16:58	clojurebot	[2/3 0.6666666666666667]
16:58	srid	ibdknox: yes, it didn't help https://gist.github.com/1265695
16:58	TimMc	Is (double) definitely slower than including a 1.0?
16:58	crazyFox	ok, (/ (double a) b) is actually about 5 percent faster than (/ a (double b)). marginal but - just in case somebody wondered...
16:59	ibdknox	srid: right, the patch that would make that work isn't in master.
17:01	srid	ibdknox: bootstrap is using the latest release of clojure-library. is there a newer snapshot uploaded somewhere that I can use?
17:01	srid	its funny there are clojurescript websocket examples without clojurescript having a working websocket support
17:01	crazyFox	TimMc: (/ (double a) b) is about 27 faster than (/ a b 1.0). if u meant that case
17:02	TimMc	nice!
17:02	crazyFox	*27 times
17:02	crazyFox	y
17:02	TimMc	How big a test run did you do?
17:04	gfredericks	I once saw a video where a guy did some clojure performance tests, but when I looked at the code it turned out he was just testing the time it took to create lazy seqs, as they were never realized
17:04	crazyFox	TimMc: just a little microbenchmark. 100 000 iterations for the inner loop
17:04	zerokarmaleft	bet they were fast
17:04	gfredericks	zerokarmaleft: they were rather fast
17:04	zerokarmaleft	heh
17:05	gfredericks	I also remember another perfomance test that the jvm must have somehow optimized away, as the results indicated something like petaflop-equivalent performance
17:05	cb`	technomancy: Thanks for your help, this is working well enough :) I have got clojure-mode from marmalade again. Will keep an eye out for new versions.
17:06	technomancy	sure
17:06	technomancy	hopefully some CLer can get things sorted out with the quicklisp versions too
17:06	TimMc	crazyFox: Did some of my own (100K, repeated until times settled out) -- got a 20x improvement by using (double)
17:08	TimMc	Moral of the story: FLOPs are expensive.
17:08	cb`	I'm not sure what the steps would be to try lein-swank with the quicklisp version of slime but it would be nice yes
17:08	crazyFox	TimMc: u mean: ration op are expensive...?
17:08	crazyFox	*ratio
17:08	crazyFox	ops
17:08	technomancy	cb`: well that might be a bit much; it would involve changing swank and breaking everyone on the stable version of slime
17:09	technomancy	but getting the stable version and the cvs version to coexist would be nice
17:09	gfredericks	floating operations per second are expensive?
17:09	TimMc	Hmm, good point -- it doesn't actually do two FLOPs.
17:09	TimMc	gfredericks: FLOPs, not FLOPS. :-)
17:09	gfredericks	TimMc: :P
17:09	TimMc	FLoat OPs
17:09	gfredericks	acronyms are expensive
17:09	zerokarmaleft	there's a stable version of slime?
17:09	TimMc	AAE
17:09	gfredericks	\o
17:10	technomancy	zerokarmaleft: sure. I mean, I just picked a working version arbitrarily and called it "version 20100404"
17:11	TimMc	crazyFox: Hrm, you're right! It is definitely the rational ops.
17:11	technomancy	but in practice it has proven to be very stable
17:11	gfredericks	FLOPs are cheap, but they never give you the correct answer. no matter what math you're doing, the result is always wrong by 0.00000079
17:11	crazyFox	TimMc: y, THEY are the real expesive ones ^^
17:11	crazyFox	TimMc: good that ur result is about the same than mine - confirms my trust in this microbenchmark thingy im using
17:11	TimMc	(/ (double 2) 3 1.0) is only 3x slower than (/ (double 2) 3)
17:12	zerokarmaleft	technomancy: ah i thought you were referring to a stable release out of the CL devs
17:13	technomancy	zerokarmaleft: oh... wouldn't that be nice. =(
17:13	TimMc	I also tried hinting with (int) in case that was the problem, but it didn't make a difference.
17:14	crazyFox	could be even slower, since long is the default...
17:14	gfredericks	I'm using clojure-maven for a clojure/java project, and am not sure how to handle the clojure depending on the java and vice-versa
17:15	gfredericks	both the "compile" and the "clojure:compile" tasks fail for the other not having been run
17:15	gfredericks	I normally use leiningen, so I'm sure I'm just missing something obvious
17:16	brehaut	gfredericks: did you remember to slaughter the goat at the appropriate hour?
17:16	gfredericks	brehaut: CRAP
17:16	gfredericks	well there goes another 24 hours of productivity
17:16	technomancy	I bet you're missing the SCSI terminator at the end of your chain.
17:16	brehaut	gfredericks: its the first rule of new build systems
17:17	arkh	would the correct way to 'serialize' the value of a symbol be with pr, pr-str, etc. ? And if so, how does one deserialize the data back into a symbol?
17:17	gfredericks	I tried setting <packaging>war</packaging> to <packaging>clojure</packaging> as suggested in the clojure-maven readme, but that caused my tests to fail, so I set it back to war
17:17	TimMc	crazyFox: I'm on 1.2. You?
17:17	gfredericks	arkh: read-string
17:18	gfredericks	and yes
17:18	gfredericks	,(read-string ":foo")
17:18	clojurebot	:foo
17:19	arkh	gfredericks: thank you
17:19	TimMc	arkh: Standard warning: Understand the implications of read-eval if you use read-string on user-provided data.
17:19	gfredericks	,(-> 'famished pr-str read-string type)
17:19	clojurebot	clojure.lang.Symbol
17:19	hugod	gfredericks: I don't think the dependencies between java and clojure can be resolved automatically, so you will need to add an execution for whichever bits need to be built first
17:19	crazyFox	TimMc: 1.3.
17:19	arkh	TimMc: right! I planned on using it only on my stuff
17:20	gfredericks	hugod: as I've never done anything in maven that required creativity, could you suggest some terms to google to get me started with that?
17:20	zippy314	hoping for some love on my cljs cookie question: has anybody successfully used cookies in clojurescript with goog.net.cookies? I keep getting this error: 'this.isValidName' [undefined] is not a function" (Safari) or "Uncaught TypeError: Object [object DOMWindow] has no method 'isValidName'" (Chrome) when I try to set a cookie via goog.cookies.set.
17:21	zippy314	s/goog.cookies.set/goog.net.cookies.set/
17:21	lazybot	<zippy314> hoping for some love on my cljs cookie question: has anybody successfully used cookies in clojurescript with goog.net.cookies? I keep getting this error: 'this.isValidName' [undefined] is not a function" (Safari) or "Uncaught TypeError: Object [object DOMWindow] has no method 'isValidName'" (Chrome) when I try to set a cookie via goog.net.cookies.set.
17:21	TimMc	gfredericks: psilocybin
17:21	hugod	gfredericks: maybe http://maven.apache.org/guides/mini/guide-default-execution-ids.html
17:22	gfredericks	hugod: thanks
17:22	TimMc	It should help witht he creativity, although perhaps not with maven specifically.
17:22	hugod	gfredericks: never tried it it with clojure-maven-plugin…
17:23	dnolen_	zippy314: seems like some context is not being properly bound. ClojureScript is pretty new territory, might want to ask on the ML as well.
17:23	zippy314	yah.
17:23	gfredericks	hugod: do you use the clojure-maven plugin? if not, maybe it has a way to do this automatically...
17:23	zippy314	Ok. thanks.
17:23	gfredericks	I'd think it be pretty common in clojure-java projects
17:24	duck1123	I think in cases like that, you might want to split your project into multiple modules
17:24	hugod	gfredericks: I've used it - I mainly use zi now. Bi-directional dependencies aren't that common, I think
17:25	gfredericks	hugod: really? E.g., I need to provide an implementation of an interface. The java code needs to call my factory method, and the clojure code needs the definition of the interface. Is this simpler than I think it is?
17:26	hugod	as duck1123 said, separate modules may be easiest
17:26	gfredericks	oh hmm.
17:26	technomancy	gfredericks: you could use resolve to get the factory at runtime maybe?
17:27	technomancy	but yes, it's very uncommon to need bidirectionality
17:27	gfredericks	technomancy: if you wanted to implement an interface to supply to the java code, what would you do? break the project into modules? use resolve?
17:27	technomancy	I'd use resolve if the Java only needs a handful of clojure defns
17:28	technomancy	but I've never written any Java, so what do I know
17:28	gfredericks	ha.
17:28	technomancy	I have seen this problem in Mirah though
17:28	technomancy	solved imperatively in ant
17:29	gfredericks	I guess I'll break out a bit of java and create a ClojureBridge class
17:29	technomancy	so by comparison, runtime resolution is much less hideous
17:29	technomancy	is definterface out of the question?
17:30	gfredericks	technomancy: to duplicate the definition of the interface? The java code is not mine, so I'm trying to avoid disturbing it
17:30	technomancy	oh, I figured you owned the java
17:30	gfredericks	It's mine in the sense that I can do what I want, but I didn't write it
17:30	gfredericks	it's sorta forked off another project
17:30	gfredericks	so e.g., when they create another release, I'd like to be able to merge in their changes painlessly
17:30	gfredericks	if I rewrite half of the code...
17:32	gfredericks	technomancy: hugod: duck1123: thanks for the help
17:32	gfredericks	(doseq [handle [technomancy hugod duck1123]] (inc handle))
17:33	gfredericks	lazybot: your karma evaluator is worthless???
17:33	lazybot	gfredericks: How could that be wrong?
17:33	lazybot	karma: It's 'you're', you fucking illiterate bastard.
17:34	gfredericks	lazybot: what kind of word has three apostrophes in it?
17:34	lazybot	It's AWWWW RIGHT!
17:35	cgray	does (empty? (filter foo bar)) short-circuit?
17:36	gfredericks	cgray: yes, since filter is lazy
17:36	gfredericks	,(empty? (range))
17:36	clojurebot	false
17:37	cgray	gfredericks: cool example :)
17:37	gfredericks	I guess that was kind of an irrelevant example
17:37	gfredericks	,(empty? (filter even? (range)))
17:37	clojurebot	false
17:37	gfredericks	there we go
17:38	cgray	yeah, that makes sense
17:38	gfredericks	cgray: the term "short-circuit" sounds a bit funny. It's normally just phrased as laziness.
17:39	cgray	gfredericks: yeah, I was thinking of it more like the way `and' short-circuits...
17:39	ibdknox	gfredericks: that is a commonly accepted term for stopping as soon as a value is known
17:40	gfredericks	ibdknox: I know, I've just never heard it used when laziness is a sufficient description
17:40	ipostelnik	gfredericks, I don't think empty? is lazy
17:40	gfredericks	ipostelnik: in what sense?
17:41	ipostelnik	actually, it probably is lazy, in a sense that computation is not performed until the result is needed by something
17:41	gfredericks	ipostelnik: I think the computation is performed in that it calls first on the input
17:42	ibdknox	it's actually (not (seq x))
17:42	gfredericks	,(let [dont-care (empty? (map #(println "foo") (range 2)))] "constant")
17:42	clojurebot	#<ArityException clojure.lang.ArityException: Wrong number of args (1) passed to: sandbox$eval79$fn>
17:42	gfredericks	augh
17:43	gfredericks	ibdknox: would seq not call first?
17:44	gfredericks	,(let [dont-care (empty? (map (fn [_] (println "foo")) [1 2]))] "constant")
17:44	clojurebot	foo
17:44	clojurebot	foo
17:44	clojurebot	"constant"
17:45	ibdknox	not lazy :)
17:45	gfredericks	I suppose chunked seqs could mess up expectations
17:45	ibdknox	it escapes into java pretty quickly and I'm too lazy to look what it does :p
17:46	gfredericks	I didn't mean that empty? itself is "lazy", I meant that filter is lazy and empty? only checks the front of the list
17:46	gfredericks	so as demonstrated earlier, you can pass an infinite seq to filter and check if it's empty and get a return
17:46	ibdknox	yep yep
17:47	gfredericks	but chunking can screw that up
17:47	arohner	what are the advantages of deploying a ring/noir app to say.. tomcat as opposed to nginx or apache?
17:47	arohner	or is there any reason to not just use jetty in production?
17:48	ibdknox	arohner: anecdotally I've had no troubles with Jetty, but presumably someone has a real answer to that question :)
17:50	ipostelnik	empty? seems to force the first chunk of the lazy seq
17:50	gfredericks	ipostelnik: right
17:51	ipostelnik	so it's not lazy in same way as map is
17:51	gfredericks	(let [dont-care (empty? (map (fn [_] (println "foo")) '(1 2)))] "constant")
17:51	gfredericks	dangit I always mess up codez
17:51	ipostelnik	,I think it reduces the weight of the small rare tz
17:52	clojurebot	#<CompilerException java.lang.RuntimeException: Unable to resolve symbol: I in this context, compiling:(NO_SOURCE_PATH:0)>
17:52	gfredericks	ipostelnik: I think that's the chunked seq's fault though. empty? only cares about the first element
17:52	ipostelnik	,(let [x (map prn (range 2))] "foo")
17:52	clojurebot	"foo"
17:52	arohner	ipostelnik: if you want to see whether there is a first item, you always have to force it. map doesn't force, because it doesn't care whether there is a first node (until map's consumer forces)
17:52	ipostelnik	sorry
17:52	ipostelnik	,(let [x (empty? (map prn (range 2000)))] "foo")
17:52	clojurebot	0
17:52	clojurebot	1
17:52	clojurebot	2
17:52	clojurebot	3
17:52	clojurebot	4
17:52	clojurebot	5
17:52	clojurebot	6
17:52	clojurebot	7
17:52	clojurebot	8
17:52	ibdknox	lol
17:52	clojurebot	9
17:52	rbranson	uh oh
17:52	clojurebot	10
17:52	TimMc	>_<
17:52	clojurebot	11
17:52	ipostelnik	woops
17:52	clojurebot	12
17:52	clojurebot	13
17:53	clojurebot	14
17:53	clojurebot	15
17:53	ipostelnik	it only goes to 31
17:53	clojurebot	16
17:53	clojurebot	17
17:53	clojurebot	18
17:53	TimMc	timeout, plz
17:53	clojurebot	19
17:53	clojurebot	20
17:53	clojurebot	21
17:53	clojurebot	22
17:53	clojurebot	23
17:53	clojurebot	24
17:53	clojurebot	25
17:53	clojurebot	26
17:53	clojurebot	27
17:53	clojurebot	28
17:53	clojurebot	29
17:53	clojurebot	30
17:53	clojurebot	31
17:53	clojurebot	"foo"
17:53	gfredericks	this is the worst disaster #clojure has seen since the great infinite seq of 2007
17:53	TimMc	haha
17:53	ibdknox	haha
17:53	rbranson	hah
17:53	ipostelnik	sorry about that
17:53	dnolen_	chunked seqs strike again.
17:54	ipostelnik	anyway, if empty? were lazy you wouldn't see any numbers printed here
17:54	gfredericks	why didn't rich chunk them to a more IRC-appropriate size?
17:54	TimMc	ipostelnik: Was that "sorry" at __:50 in advance? :-P
17:54	ipostelnik	noidi, that one was a copy-and-paste error with ,
17:54	gfredericks	ipostelnik: I never meant lazy in that sense. That sort of lazy wouldn't be possible without using a delay object and having to deref the result
17:54	ipostelnik	I'm not using x anywhere
17:55	TimMc	&(let [x (map prn (range 2000))] "foo") But this is safe.
17:55	lazybot	⇒ "foo"
17:55	ipostelnik	because map is lazy
17:55	TimMc	Fully.
17:55	ipostelnik	empty? computes the value right away
17:55	gfredericks	no, because seqs can be lazy but atomic values can't be
17:56	TimMc	&(let [x (first (map prn (range 2000)))] "foo")
17:56	lazybot	⇒ 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 "foo"
17:56	TimMc	lazybot: botsnack
17:56	lazybot	TimMc: Thanks! Om nom nom!!
17:56	ipostelnik	lazybot is better at long output :)
17:56	ibdknox	is there some significance for it being 32 elements in a chunk?
17:57	gfredericks	something about 5 bits I'm sure
17:57	amalloy	ibdknox: probably because vectors have a branching factor of 32
17:57	ibdknox	amalloy: ah
17:57	TimMc	and 32 is 2 to the whatever
17:57	ipostelnik	amalloy, what is branching factor?
17:58	TimMc	There was just a good link about this.
17:58	TimMc	ipostelnik: http://blog.higher-order.net/2009/02/01/understanding-clojures-persistentvector-implementation/
17:58	ipostelnik	TimMc, thanks
17:59	TimMc	Long story short, the vectors are implemented as trees.
18:20	cgray	what's the timeout for solutions to the 4clojure.com exercises?
18:21	amalloy	cgray: 10s for each testcase
18:22	cgray	amalloy: interesting, i wonder why i'm timing out... i have a solution that's running in 6s on my (not very powerful) machine
18:22	amalloy	cgray: sandboxing takes extra time. more time than we can really understand, tbh
18:22	devn	TimMc: That is a fantastic post. I'm building a list of stuff for people of my Clojure users' group to read, and that is definitely now on the list.
18:23	cgray	amalloy: ok, good to know... i'll just have to do more optimization i guess :)
18:25	amalloy	cgray: solving levenshtein distance? not many of the other problems have performance issues
18:25	cgray	amalloy: no, balanced primes
18:26	cgray	amalloy: my first solutions have were atrocious, but i should be getting close now
18:26	amalloy	cgray: hm. i didn't make any special effort to optimize, and my testcases run in 1-2s total
18:29	amalloy	cgray: mind if i take a look?
18:29	cgray	amalloy: no, should i paste it somewhere?
18:29	amalloy	i guess that makes it sound like i can look at it without you gisting it or something, which i can't
18:30	amalloy	yeah, please
18:30	cgray	http://pastebin.com/twX4B7A5
18:33	amalloy	cgray: well, you're recomputing the list of primes a zillion times, it seems to me
18:34	cgray	amalloy: i'm trying not to
18:34	amalloy	eg, (is-prime n) has to compute all the primes up to n, and then if it returns true you immediately call next-primes again
18:34	cgray	right, that's two computations :)
18:34	amalloy	my solution was to let a lazy seq of all the primes outside the function itself, and then just use that for reference without recomputing it
18:35	amalloy	(letfn [(primes [] (...compute primes...))] (let [all-primes (primes)] (fn [n] (...solve actual problem...))))
18:35	cgray	ok, makes sense
18:36	amalloy	which is especially useful for the last test case where your function gets called once on every number up to 1103
18:36	cgray	yeah, that's the one that's killing me
18:37	cgray	cool, thanks for the hint
18:43	srid	closure doesn't have a prependChild? what a shame.
18:44	patchwork	hey all, is there a quick way to get the path of the project root in a lein project?
18:44	patchwork	so I have a project in "/Users/xxxxx/projects/lein-project"
18:45	patchwork	and I want that as a string
18:45	srid	what would this be if your app is run as a jar?
18:45	patchwork	to do various programmatic things
18:45	patchwork	srid: sure?
18:45	patchwork	although that is a good question, I still have access to the filesystem from a jar right?
18:46	patchwork	or is it some kind of encapsulated system?
18:46	srid	compojure uses resources/ directory for css, img, js and things. they get wrapped in a jar. so maybe you can look at compojure's route/resources implementation
18:46	patchwork	as you can tell, I have much ignorance towards the practical side of jars and java deployment
18:46	patchwork	srid: Aha
18:47	patchwork	I have users uploading images
18:47	patchwork	so I want to write those images somewhere in the project dir
18:47	patchwork	should those go in resources too?
18:47	patchwork	I thought resources were more for application resources, not user generated content
18:47	Arafangion	Generally yes.
18:47	srid	probably not. why not $HOME/.project-uploads
18:48	Arafangion	As in, Generally "yes, you're correct".
18:48	patchwork	So, do I have access to $HOME from inside clojure?
18:48	srid	&(System/getenv "HOME")
18:48	lazybot	java.security.AccessControlException: access denied (java.lang.RuntimePermission getenv.HOME)
18:48	patchwork	all of these practical issues I haven't had to deal with yet from clojure
18:48	patchwork	nice
18:48	amalloy	yes, though you shouldn't get at it that way
18:49	patchwork	amalloy: so where should I put all of these images? : )
18:49	patchwork	I guess that is a better question
18:49	amalloy	(System/getProperty "user.home")
18:49	amalloy	more portable than getenv of "HOME"
18:49	srid	right, windows has no HOME by default
18:50	srid	that's the first thing I do on every windows install
18:50	patchwork	amalloy: Aha, that is helpful
18:50	srid	$g gridfs filesystem
18:50	srid	$google gridfs filesystem
18:50	lazybot	[mongodb - Is GridFS fast and reliable enough for production ...] http://stackoverflow.com/questions/3413115/is-gridfs-fast-and-reliable-enough-for-production
18:50	patchwork	still though, is there a best practice for where to store user files in a lein project?
18:50	amalloy	patchwork: just store them in .
18:51	amalloy	or ./users/images, whatever
18:51	amalloy	you can still construct java.io.File objects, and the cwd is whatever the cwd was when you started the jvm
18:51	patchwork	So if I reference '.' from a file, it gives me the root of the project?
18:51	patchwork	okay, that is good
18:51	patchwork	that is what I was wondering
18:52	amalloy	&(-> (java.io.File. "test") .getAbsoluteFile)
18:52	lazybot	java.security.AccessControlException: access denied (java.util.PropertyPermission user.dir read)
18:52	amalloy	feh
18:58	sritchie	hey guys, is there some equivalent to defalias in clojure 1.3?
19:56	ibdknox	Steve Jobs died :(
19:57	duck1123	we all knew it was coming. Didn't think it would be this soon
19:57	duck1123	Of course, it makes sense that he would wait to step down til the last moment
20:03	srid	aleph + clojurescript + websocket awsomeness.
20:04	duck1123	srid: you got it working?
20:04	srid	duck1123: yup, in a company project. so I'm considering doing it again in a minimal example app and writing a blog about it.
20:05	jayunit100	aleph?
20:05	duck1123	look forward to seeing that
20:05	srid	$google aleph lamina
20:05	lazybot	[ztellman/aleph - GitHub] https://github.com/ztellman/aleph
20:05	jayunit100	srid what is this stack that you put together>
20:07	srid	jayunit100: see above. lamina is also ztellman's project. i may write about it in the weekend.
20:07	srid	i was afraid of using clojurescript, for I didn't how easy it would be to debug cljs. in practice, it seemed as if i was able to develop client side code /faster/ in cljs than js.
20:08	jayunit100	oh ok.
20:09	jayunit100	so your application is a bunch of web clients talking to each other ?
20:12	jayunit100	leiningan guy !
20:13	technomancy	IRC user with a number in his nick!
20:13	ibdknox	lol
20:14	jayunit100	yuppppp
20:24	cemerick	ibdknox: FYI, I've not yet hit any big snags using ClojureScript as a long-lived (very relatively speaking at this point) library.
20:25	ibdknox	cemerick: awesome :)
20:25	ibdknox	cemerick: what tools have you been using?
20:25	cemerick	tools?
20:25	clojurebot	eg, https://github.com/clojure/tools.logging is the new version of clojure.contrib.logging
20:25	ibdknox	lol
20:25	ibdknox	cemerick: browser repl, cljs-watch, repl compiler?
20:26	cemerick	no, no, and no
20:26	cemerick	Just regular REPL, compiling cljs forms down to js strings, stuffing that into couchdb
20:26	cemerick	Outside of my flailing around with cljs, it's worked without a hitch.
20:27	cemerick	(maybe that's what you meant by 'repl compiler')
20:27	jayunit100	what does <== do
20:27	ibdknox	yep :)
20:28	cemerick	I guess I didn't categorize that as a tool. :-)
20:28	cemerick	'tool' implies artifice to me. Code is…just code.
20:28	sritchie	cemerick, just wanted to let you know that Couchbase Single worked great
20:29	sritchie	thanks for the help
20:29	ibdknox	yeah, I guess so. I didn't think about the fact that you were outputting to place it into couchdb
20:29	ibdknox	that makes most of those not useful lol
20:29	cemerick	sritchie: That's interesting! I'd be curious if the stock CouchDBX for 1.0.2 or whatever would work for you too.
20:29	cemerick	e.g. maybe it's the homebrew build getting in the way
20:30	sritchie	cemerick: I'll go ahead and test out couchdbx as well; I'll also take a look at that homebrew build and see if I can track down what it's sending to clojure, to start up the jvm
20:30	cemerick	ibdknox: just different use cases. I'm not a web guy, so it's properly stereotypical of me to totally ignore all the nifty things going on in those spaces.
20:31	duck1123	does anyone know of a good way to keep firefox from popping up and stealing focus when doing tests with clj-webdriver?
20:31	ibdknox	cemerick: yep yep. Good to go for tomorrow then? :)
20:31	duck1123	it makes it impossible to get anything done when I'm running features
20:32	cemerick	ibdknox: yup, it'll be nifty. Some definite TODOs, but it'll be a nice demonstration, and another good cljs use case.
22:00	glob157	is it idiomatic to do (class? Seq) to test before calling cons or distinct or other seq funcs ?
22:00	brehaut	glob157: seq? would be more idiomatic
22:01	glob157	oh ok. would mine work ?
22:01	gfredericks	no
22:01	glob157	&(class? Seq [2 1])
22:01	lazybot	java.lang.Exception: Unable to resolve symbol: Seq in this context
22:01	gfredericks	glob157: you're probably wanting instance?
22:01	brehaut	glob157: depending on what you are doing (seq thing) is more idiomatic still
22:02	glob157	why instance?
22:02	dnolen	glob157: why do you need to test before calling seq functions?
22:02	brehaut	glob157: class? tells you if something is a class
22:02	glob157	well , clojure is loosely typed. so seems like you would want to test such things
22:02	khaliG	he probably means just plain class
22:02	brehaut	,(map class? [(list 1 2) java.util.Date 1 :a])
22:02	clojurebot	(false true false false)
22:03	dnolen	glob157: it is not idiomatic to include such tests
22:03	glob157	a list isnt a class ?
22:03	glob157	nor is :a ? thats odd.
22:03	glob157	how do they work in the jvm then
22:03	brehaut	glob157: no, only classes are classes.
22:03	gfredericks	glob157: they're objects, not classes
22:03	brehaut	glob157: they are objects, aka instances, of classes
22:03	glob157	duh oops
22:04	gfredericks	,(class? Class)
22:04	clojurebot	true
22:04	gfredericks	I'm not sure why I did that, that was probably just confusing.
22:04	glob157	(class? (java.util.Date new))
22:04	gfredericks	,(class? (new java.util.Date 2840824))
22:04	clojurebot	false
22:04	brehaut	glob157: also, clojure is not loosely typed, it is dynamically typed.
22:04	glob157	&(class? (java.util.Date new))
22:04	lazybot	java.lang.Exception: Unable to resolve symbol: new in this context
22:05	glob157	&(class? (new java.util.Date))
22:05	lazybot	⇒ false
22:05	gfredericks	&(instance? java.util.Date (new java.util.Date))
22:05	lazybot	⇒ true
22:05	khaliG	glob157, i think you meant (class (new java.util.Date))
22:06	glob157	yup
22:06	brehaut	glob157: loosely typed implies that the data can be reinterpreted, such as in C where you can cast things about, or in some interpreted languages where coercion happens all over the place and is implicit
22:06	dnolen	glob157: if you want to guarantee invariants, use pre-post conditions. those at least can be turned on / off
22:08	hugod	dnolen: did you manage to get warn-on-reflection into your pom?
22:09	glob157	yeah, pre/post is nice. whats the point of the @symbol
22:09	gfredericks	glob157: it's short for deref
22:09	gfredericks	,(deref (atom 8))
22:09	clojurebot	8
22:09	gfredericks	,(let [a (atom 8)] @a)
22:09	clojurebot	8
22:09	brehaut	,(read-string "@(atom 8)")
22:09	clojurebot	(clojure.core/deref (atom 8))
22:10	gfredericks	slick.
22:10	dnolen	hugod: no, but thanks for the tip, will use that in the future.
22:11	amalloy	&'@(atom 8) ;; brehaut - no need for read-string
22:11	lazybot	⇒ (clojure.core/deref (atom 8))
22:11	brehaut	amalloy: i know three no need, but read string makes it clearer that it occurs in the reader
22:13	gfredericks	amalloy: that had as many punctuation marks as it did alphanumerics.
22:13	amalloy	gfredericks: welcome to perl
22:13	gfredericks	aw I hate these dreams
22:14	brehaut	amalloy: not quite, perl is more punctuation than alphanumeric
22:14	amalloy	brehaut: i carefully did not do it inline with ##'@(atom 8), see
22:14	lazybot	⇒ (clojure.core/deref (atom 8))
22:15	brehaut	wait, am i in #perl? when did that happen‽
22:15	gfredericks	brehaut: anything can happen when you're dreaming
22:48	dnolen	match 0.2.0-alpha4 out. Mostly involved deleting code, YAY
22:54	toxmeister	hiya… could someone pls tell me briefly what the correct pattern is for when using primitive type hints ^double or boxed ones ^Double (w/ 1.3)? am encountering all sorts of weird compiler errors like: "Unable to resolve classname: clojure.core$double@1f3a34af"
22:55	srid	what is it with CL/scheme kids looking down upon people learning clojure? http://news.ycombinator.com/item?id=592502
22:56	brehaut	srid: the narcissism of minor differences
22:56	srid	the only argument I can accept is the utility of works like SICP.
22:57	srid	they tell me to study every dialect of lisp, their specs, libraries, etc.
22:57	srid	which is crazy
22:57	srid	"play around with all the dialects and implementations it supports"
22:58	brehaut	srid: racket's environment is a good argument for learning racket / scheme over another lisp
22:58	brehaut	srid: you realize that thread is two and a half years old right?
22:59	srid	yes. someone I know forwarded it, along with his arguments to learn CL over clojure, to me recently.
22:59	srid	fwiw, i did study most of sicp (and that affected the way i programmed / thought-about-programs in other languages).
23:01	dnolen	srid: mahmud doesn't know anything about Clojure (as far as I know) so informed opinion that is not.
23:02	dnolen	srid: but the gist of what he says is true. While it's not necessary, surveying the various languages from Rich Hickey's Amazon book shelf will enrichen your understanding of Clojure.
23:03	srid	I agree. I have some of them in my wishlist, just need to find time and energy to get to them.
23:03	brehaut	dnolen: his biggest mistake is probably in that thinking only one language could possibly have influenced the design of clojure
23:04	srid	... though how exactly would 'enriching' my understanding of Clojure help (if I'm not going to design a new programming language)?
23:04	ibdknox	to command a thing you have to know it inside and out
23:05	ibdknox	enriching your understanding of Clojure gives you greater control over it
23:05	srid	for example?
23:06	ibdknox	performance is a great example
23:07	dnolen	srid: have you done much work w/ code that might callback at anytime creating subtle race conditions? (even in JavaScript)
23:07	sjl	if anyone feels like tearing apart my first Noir webapp and telling me how to make it better, I'd love advice: https://github.com/sjl/newseasons
23:08	dnolen	sjl: you're on a Clojure roll.
23:08	srid	dnolen: not much - in the distant past i did a bit of python twisted deferred programming, don't remember having to debug race conditions.
23:08	sjl	dnolen: hah, not really. I've paused Clojurecraft until Notch stops breaking the damn wire protocol all the time :)
23:08	amalloy	toxmeister: should be ^double afaik
23:09	srid	is callback-based programming related to enriching clojure understanding through studying various works based on lisp?
23:09	brehaut	sjl: not strictly clojure, but your email regex is invalid
23:09	sjl	brehaut: I used the one here: http://www.regular-expressions.info/email.html
23:10	ibdknox	sjl: there's one in noir
23:10	sjl	brehaut: I mostly agree with the Trade-Offs section on that page
23:10	sjl	ibdknox: yeah, but I needed it for the route, and the only one I saw in noir is a function, not a raw regex
23:10	ibdknox	sjl: ah, yep :)
23:11	dnolen	srid: Mozart/Oz has a pretty amazing declarative programming model that makes programming under concurrency trivial - it eliminates the problems around ordering entirely.
23:11	ibdknox	sjl: fwiw you don't need to explicitly state a route is [:get], it's that by default
23:11	brehaut	sjl: http://commons.apache.org/validator/api-1.3.1/org/apache/commons/validator/EmailValidator.html
23:11	toxmeister	amalloy: thx, that's what i thought, but here's the deal:
23:11	sjl	ibdknox: yeah, I did that to make it easier to change as I worked with the pages, but I could go ahead and remove them now
23:11	toxmeister	this doesn't work: (def ^double PI 3.14) (defn rad ^double [^double x] (/ (* PI x) 180.0))
23:12	dnolen	srid: if you understand that, promise/deliver makes a lot more sense.
23:12	toxmeister	but this one (boxed) does work: (def ^Double PI 3.14) (defn rad ^double [^double x] (/ (* PI x) 180.0))
23:12	ibdknox	sjl: is there a reason you use (form-to) sometimes and not others?
23:12	amalloy	toxmeister: that's not how typehints on defs work
23:12	dnolen	toxmeister: if you're on 1.3.0 you don't want vars for things like PI, you want constants anyhow.
23:12	sjl	brehaut: I need the regex as an actual Clojure regex for the Noir URL, not just a function/validator
23:13	brehaut	fair enough
23:13	amalloy	if you hint a def's name, you're marking it as "a function which returns this when called"
23:13	amalloy	or some such
23:13	srid	dnolen: I see. that's specific enough, for me. :-) i wish there was "matrix" of programming languages where the cells represent the benefit of learning Y in understanding some features of X.
23:13	sjl	ibdknox: nope, I think I learned about form-to halfway through, hah
23:14	ibdknox	sjl: having an email in the url is a bit weird to me, but I guess there's nothing technically wrong with it
23:14	ibdknox	sjl: look up pre-route :) You shouldn't need to put your (login-required) wrapper all over the place
23:15	brehaut	ibdknox: not being an expert with clout / webnoir, does the url rules allow you to specify any callable as a validator?
23:15	toxmeister	i see, didn't know that def's can't be dealt with the same way (sorry if this has been a case of RTFM, but i went in circles an thought I've read everything i could find about that topic)
23:15	sjl	ibdknox: yeah, I didn't want to have usernames and I don't trust myself to slugify URLs perfectly. It's a bit weird though, you're right.
23:15	sjl	ibdknox: I saw pre-route but couldn't figure out how to make it work with the email url
23:15	sjl	ibdknox: pre-route seems to only take a string, not a regex
23:16	ibdknox	sjl: same routing mechanism as defpage
23:16	ibdknox	sjl: the only difference is that the thing you destructure isn't params, it's the entire request
23:16	ibdknox	sjl: so you should be able to do anything you could with defpage
23:16	alandipert	toxmeister: (def ^:const PI 3.14) (defn rad [x] (/ (* PI x) 180)) ; PI a constant, rad won't box
23:17	ibdknox	brehaut: compojure only allows regexs I believe
23:17	sjl	ibdknox: hmm, okay, I'll take a closer look at that
23:17	brehaut	ibdknox: ok. I'm used to mustache allowing fns
23:17	ibdknox	brehaut: though, it's trivial to do that check in the body
23:18	brehaut	of course
23:18	ibdknox	sjl: wait, why does the email route make a difference? From what I can tell everything other than / requires a login
23:19	ibdknox	sjl: so you could just do (pre-route "/*" [] (when-not authed ...
23:21	sjl	ibdknox: just tried that, it first on / too
23:21	sjl	s/first/fires/
23:21	lazybot	<sjl> ibdknox: just tried that, it fires on / too
23:21	sjl	ibdknox: so going to / when not logged in goes into an infinite redirect loop
23:23	ibdknox	ah
23:24	ibdknox	hm, I'm not sure if there's some trickery we could do there or not
23:24	glob157	how would i make a future write to a queue with a "listener" on it .... like for an a synchronous web request/process
23:24	ibdknox	most of the time restricted urls are under their own path
23:25	sjl	ibdknox: yeah, on a bigger site I would have done that
23:26	ibdknox	sjl: lol since all your urls are only one segment, you can do [:get ["/:something" :something #".+"]]
23:26	ibdknox	lol
23:26	sjl	ibdknox: so I tried this, and it doesn't work: (pre-route [:get ["/:anything" :anything #".+"]] {:keys [anything]} ...)
23:26	ibdknox	that would work
23:26	sjl	ibdknox: java.lang.ClassCastException: java.lang.String cannot be cast to clojure.lang.IPersistentVector
23:26	ibdknox	hm
23:26	ibdknox	sjl: sounds like I have a bug to look into, can you add an issue in github
23:26	sjl	ibdknox: sure
23:28	glob157	just found "full disclojure" ... very high quality videos.
23:31	ibdknox	sjl: looks pretty solid :)
23:32	sjl	ibdknox: cool, thanks. now I need to figure out how to add tests, hah
23:32	ibdknox	:)
23:32	ibdknox	sjl: what's this about? https://github.com/sjl/newseasons/blob/master/src/newseasons/loops/refresh.clj#L11
23:32	sjl	ibdknox: I'll have to consolidate the redis stuff in one place and use a separate db for testing I guess
23:33	sjl	url 1
23:33	sjl	derp
23:34	sjl	,(sort < ["b" "a"])
23:34	clojurebot	#<ClassCastException java.lang.ClassCastException: java.lang.String cannot be cast to java.lang.Number>
23:34	sjl	ibdknox: ^ that
23:34	ibdknox	(sort ["b" "a"])
23:34	ibdknox	,(sort ["b" "a"])
23:34	clojurebot	("a" "b")
23:34	sjl	wtf
23:34	ibdknox	lol
23:35	sjl	ibdknox: actually I don't need to sort, I actually need to compare two
23:35	ibdknox	(> "a" "b")
23:35	sjl	ibdknox: https://github.com/sjl/newseasons/blob/master/src/newseasons/loops/refresh.clj#L36
23:36	ibdknox	,(> "a" "b")
23:36	clojurebot	#<ClassCastException java.lang.ClassCastException: java.lang.String cannot be cast to java.lang.Number>
23:36	ibdknox	ah
23:36	sjl	ibdknox: I could just inline the compare call, but I think pulling it out makes it easier to read
23:37	ibdknox	hm, I've got to believe there's something that does that correctly
23:37	sjl	ibdknox: if there is, let me know :)
23:39	trptcolin	java interop noob question: i have a thread that i'd like to raise an exception on from a different thread. (.interrupt thread) doesn't appear to be sufficient - any hints?
23:41	trptcolin	i can verify that the thread gets isInterrupted set, but can't seem to trigger an exception. i'm also very open to a more clojure-y solution if there is one as i suspect
23:42	alandipert	trptcolin: in the target thread, are you in a loop of some kind?
23:42	duck1123	trptcolin: My guess is you're going to need to have that thread monitor some shared state and then throw
23:42	amalloy	trptcolin: threads can only throw InterruptedExceptions at particular blocking-points, such as while waiting on a socket
23:42	trptcolin	alandipert: yup
23:43	trptcolin	amalloy: ahh, that explains that
23:43	alandipert	trptcolin: are you blocked on a queue in your loop? or socket/io
23:44	sjl	ibdknox: Oh, another thing, when I'm working in the auto-reload mode and make an error, I get the traceback the first time I visit the page,
23:45	trptcolin	io - i'm still digging into the internals of jline here to see where the blocking is
23:45	sjl	ibdknox: but if I refresh it seems like it goes back and uses the old version of the code
23:45	sjl	ibdknox: until I make another change, and then it reloads again
23:45	trptcolin	of course: InputStream read
23:45	ibdknox	sjl: that's because it was a compile error and wasn't actually able to load the file
23:46	alandipert	trptcolin: i cooked up https://gist.github.com/1214282 to deal recently with a similar problem, approach may or may not be useful to you
23:46	sjl	ibdknox: Hmm, I don't suppose there's a way to make it not do that? When I'm developing I always want to see the result of the current state of the code
23:46	ibdknox	sjl: not that I can think of :/
23:46	ibdknox	sjl: I look into it thoug
23:48	sjl	ibdknox: cool, thanks
23:48	alandipert	trptcolin: fill-queue (used to be in seq-utils, dunno where is now) is another way, may be applicable as well. beats explicit loop/block, anyhow
23:49	trptcolin	yeah, unfortunately the loop itself is in code i'd prefer not to control
23:50	alandipert	trptcolin: maybe you can feed it a dummy val to unblock, then interrupt
23:51	alandipert	hack city though :-\
23:51	trptcolin	interesting.
23:51	trptcolin	it's already hack city
23:51	alandipert	then let's paint the city purple
23:51	trptcolin	i'm essentially trying to trap SIGINT to have a REPL that doesn't quit on ctrl-c, but just skips to the next prompt
23:52	trptcolin	totally hotspot-specific, i know, but it's a dream
23:53	alandipert	trptcolin: running from cli? maybe you can wrap with shim that traps and doesn't propagate
23:55	trptcolin	right. so here's the other thing: i'd like to be able to, say, kill running infinite loops as well
23:56	trptcolin	basically what irb does
23:56	trptcolin	i can "stty intr undef" and skip all ctrl-c's, but that makes the other stuff impossible
23:57	duck1123	trptcolin: that's also what swank does. (in emacs, at least)
23:59	trptcolin	hmm, i don't know much about swank - wonder if it's in clojure or elisp
23:59	trptcolin	clones