#clojure – 2011-01-11

2011-01-11

00:22	qbg	Perhaps I can make an alpha2 release of syntax-rules tonight
00:44	amalloy	$mail cemerick it occurs to me your first-draft benchmark was a nice satirical comparison of clojure to java: clojure takes N seconds to get something done; java takes 2N and returns null :)
00:44	sexpbot	Message saved.
00:56	qbg	Yes, new release out
01:27	Berengal_work_	Is there a way to apply a java method the way you apply regular functions?
01:27	Berengal_work_	As in the function "apply"
01:37	bortreb	what's the latest clojure resource for reading from a file blazingly fast in map-reduce style?
01:38	bortreb	something that can do parallel line processing of a 50 Gb log file for example?
01:39	amalloy	Berengal_work_: i don't think so
02:10	Berengal_work_	int
02:10	Berengal_work_	oops
02:23	vIkSiT	hi all
02:23	vIkSiT	i'm currently using the randomAccessFile java class to tail a file - but it stops working if the log file rotates. Does anyone know how to ensure that the class keeps track of the new file?
02:24	vIkSiT	https://gist.github.com/774156 is what i use
02:25	amalloy	vIkSiT: do you want it to follow the new, rotated-in file, or the rotated-out file with the new name?
02:26	vIkSiT	the rotated-in file.
02:26	vIkSiT	so if i tail current.log, then it should always follow current..
02:27	amalloy	not that i really have any idea how to solve your problem :P
02:27	vIkSiT	hehe
02:27	amalloy	well, once java has the file open, it has a pointer to the file's id, not its name
02:27	vIkSiT	i'm trying to figure out what to google for - rotating files or following files
02:27	amalloy	just like any other process
02:27	vIkSiT	amalloy, right
02:27	vIkSiT	except - the inode changes
02:27	amalloy	vIkSiT: i just tried both of those searches
02:27	vIkSiT	so there has to be some way to keep track of the linked to file
02:27	vIkSiT	amalloy, hehe
02:28	amalloy	perhaps see what log4j does?
02:31	vIkSiT	hmm yes thats an idea
02:34	amalloy	or actually the unix tail command. i think it just polls/stats the filename every so often
02:36	amalloy	chuckle $ tail -F mylogfile \| java -cp .....
02:36	LauJensen	Is there a tool to convert a standalone jar to an .exe ?
02:37	amalloy	LauJensen: why? don't windows installs come with a file association for jars these days?
02:38	LauJensen	For some reason its a requirement of USB vendors to have applications delivered as .exe files in order to split it and hide parts of it on the drive itself
02:38	amalloy	that is priceless
02:39	amalloy	int main(void) {exec("java -cp ...");}
02:41	LauJensen	hehe
03:41	amalloy	raek: exceptions are making me weep at the moment. are we getting your map exceptions, or something similar, soon?
03:42	tomoj	map exceptions?
03:44	amalloy	tomoj: https://github.com/raek/map-exception
03:44	amalloy	contributed as an idea to clojure.core, but from the sound of http://dev.clojure.org/display/design/Error+Handling+Comparisons?focusedCommentId=950796 we'll end up with something else
03:45	amalloy	(that is, i already know we won't be getting these things in a huge hurry, but i want raek to make soothing noises)
03:46	Raynes	amalloy: Don't count your chickens before they hatch. raek has been rewriting Irclj's message parser for the better part of 2010. ;)
04:09	LauJensen	Raynes: Always great with some of that Alabama wisdom :)
04:11	ejackson	LauJensen: you're right about my blog. All the ideas I've had for it would take too long to do and I'm being lazy
04:12	LauJensen	ejackson: well, at least you're man enough to admit it :)
04:13	ejackson	i have a few kettles in the fire, and at least one of them will end up on the blog
04:13	LauJensen	well, dont sell your kettles before they burn as Raynes would say
04:14	ejackson	or my teapots will come home to roost...
04:14	LauJensen	hehe
04:14	bortreb	anyone know of a good resource to learn about processing huge log files in clojure?
04:15	LauJensen	bortreb: depends on what you mean by good. Alex Osbourne has a blog that shows how to parse 45 Gb in about 8 minutes. I have a blogpost about parsing 50Gb weather data
04:16	gregh	bortreb: try http://stackoverflow.com :)
04:16	bortreb	I was wondering if there was some clojars library or something I am overlooking
04:16	bortreb	is 45gb in 8 minutes "good" ?
04:16	LauJensen	bortreb: no, its fast
04:16	ejackson	bortreb: nathanmarz's cascalog is a tool that might be useful for you
04:17	LauJensen	Ruby took 45 hours I think
04:17	LauJensen	ejackson: casalog for parsing logs?!
04:17	ejackson	depending on how sophisticated your definiton of parsing is :)
04:17	ejackson	and how big your logs
04:17	LauJensen	..and how much syntax you want to learn
04:18	raek	the plan seems to be to define the "map exception"/condition class first
04:18	LauJensen	Isnt casalog that query language that looks like [?age <> ?person >-@> ::filter] ?
04:18	amalloy	LauJensen: that looks more like clojure than most things do :P
04:18	LauJensen	For you know. When you what the age of the person to be confused for a fish swimming towards a filter
04:18	ejackson	yeah,
04:19	gregh	I've had good success recently processing big file by memory mapping them (on a 64-bit machine)
04:19	ejackson	it allows pretty arbitrary logic statements to be executed against gigantic files on a hadoop system
04:19	LauJensen	gregh: Thats nice - It sucks when you have 'bad success' doing that
04:19	LauJensen	ejackson: Man you'll earn your money as a consultant if you point people to Hadoop and Casalog when all they want is to parse a log file
04:19	bortreb	maybe bad success is getting handed more log parsing jobs because you're sooooo good at it?
04:20	afekz	naah, "bad success" is probably some jazz term or something
04:20	ejackson	ok, i was misled by the 'huge' in bortreb's query. I'll admit its probably overkill :)
04:20	LauJensen	jazz is bad sucess :)
04:20	bortreb	thanks, I'll read the blog then :)
04:20	LauJensen	ejackson: no no - While you're at it, get him an internship at Google so he can parse the logs on their hardware, thats really the only way to go
04:21	LauJensen	Everytime I need to skim a logfile, I always get an internship somewhere :)
04:21	ejackson	LauJensen: naaah, that's to 2009 - Twitterface or something.
04:21	TobiasRaeder	morning :)
04:22	LauJensen	Morning :)
04:22	ejackson	yello
04:23	TobiasRaeder	Anyone here using align in emacs?
04:23	jwr7	Is there a way to specify a timeout on an agent action? My agents perform network I/O which (as it turns out) may hang — I'd like to have a way of getting an exception if I/O takes too long.
04:24	jwr7	…and the obvious solution (use futures in agent actions) seems overly complex.
04:24	raek	jwr7: why not use (future ....) instead of send?
04:24	amalloy	jwr7: the obvious solution is to set a timeout on the network io itself
04:25	raek	future objects have a .get method that acccepts a timeout value
04:25	jwr7	raek: well, the agent metaphor works really really well for me.
04:26	jwr7	amalloy: well, this would mean using futures to perform I/O operations, lots of additional complexity.
04:27	amalloy	jwr7: there's no reason it should mean that
04:27	raek	ok, if you need to asynchronously keep track of some state rather than synchronously (ref/atom), then I guess agents are the best fit
04:27	jwr7	I was actually surprised there isn't a way to specify a timeout when calling send/send-off — I thought people would have ran into this already.
04:27	amalloy	jwr7: http://download.oracle.com/javase/1.4.2/docs/api/java/net/Socket.html#setSoTimeout%28int%29
04:27	raek	agents are not the main mechanism for "sending off stuff for running in another thread"
04:28	raek	s/main/most basic/
04:28	sexpbot	<raek> agents are not the most basic mechanism for "sending off stuff for running in another thread"
04:28	jwr7	amalloy: the network I/O I do accesses Amazon S3 using the jets3t library…
04:29	jwr7	raek: in my case agents are a perfect fit. I need serialized async operations on a data structure.
04:29	raek	ok. then agents sounds perfect.
04:30	raek	(agents tend to be somewhat over-used)
04:30	jwr7	so I guess the answer to my original question ("Is there a way to specify a timeout on an agent action?") is "no"
04:30	jwr7	raek: yes, I know — but in this case they are really perfect. I would be reimplementing agents if I chose anything else.
04:30	raek	jwr7: yes, that is true. how should the action be aborted?
04:31	raek	the simplest way is to make the action cooperative and let itself be interrupted
04:31	jwr7	raek: I'd expect a java.util.concurrent.TimeoutException which would cause an agent error.
04:31	jwr7	…which the rest of my application already knows how to deal with.
04:32	raek	exception handling is tricky when doing asynch stuff
04:32	jwr7	put differently, I'd like an agent implemented using futures — and I actually thought that is exactly what send-off does
04:32	raek	you can of course catch it in the action, and store it somewhere
04:33	raek	yes, agents are implemented on top of two ExecutorServices
04:33	jwr7	raek: as I said, the application already has to deal with agent errors and it does. It's tricky, but all concurrent systems are.
04:34	jwr7	so the obvious way to implement what I need would be just to wrap my action in a future call with a timeout.
04:34	jwr7	but that means I'd be creating twice the necessary number of threads, for no good reason
04:35	raek	hrm, yes
04:35	jwr7	…and I might end up doing it anyway, as it is the only solution I can think of right now.
04:36	raek	what should the operation on the data structure be in case of a timeout?
04:36	jwr7	raek: nothing — no change at all.
04:37	jwr7	raek: agent semantics — either the action succeeds or it doesn't, you don't get states in between.
04:38	raek	and the lib does not provide any way of setting a timeout?
04:38	raek	(the network IO thing)
04:39	raek	everything would have been simple if the blocking IO call could just throw an exception if it takes too long time
04:40	jwr7	raek: I'm going through the docs right now. But I'd like a more reliable mechanism, as in "whatever happens, there is a global timeout".
04:40	jwr7	raek: socket timeouts won't kick in if the data starts dribbling at 5 bytes a second for 2 days.
04:40	jwr7	raek: so I'd much rather have an absolute limit on the amount of time an agent action may take.
04:40	raek	then you pretty much need a thread that you can inerrupt
04:41	raek	also, it's too bad that blocking socket operation do not throw an InterruptedException when the thread is interrupted
04:42	raek	but another thread can close the socket, which will stop the blocking op
04:42	jwr7	raek: I think I'll try a macro that wraps an agent action in a future with a timeout on deref. E.g. instead of defn'ing my agent actions I'll defagentaction them.
04:43	raek	don't forget to future-cancel it
04:44	jwr7	raek: hmm — what do you mean? I don't understand that?
04:44	raek	futures provide a way of signalling that you don't care about the result anymore
04:45	raek	that will try to interrupt the thead if possible
04:45	raek	most blocking operations (except sockets) will then throw an InterruptedException
04:46	raek	so if the network operation takes too long time, you'd want to signal it that it can stop doing what it does, since you don't care about the result anymore
04:48	raek	(let [ftr (future ...)] (try (.get ftr 1 TimeUnit/MINUTES) (catch TimeoutException e (future-cancel ftr) (throw e))))
04:49	raek	also, I really recommend the book http://www.javaconcurrencyinpractice.com/
04:59	shortlord	is there any elegant way to test if a coll contains at least one logical true value? It seems strange to use 'some' and then check for 'nil?' and 'false?'
05:00	mrBliss	shortlord: (some identity coll)
05:00	nathanmarz	LauJensen ejackson: cascalog can be used for local, in-process computation. I do that a lot when developing or when I have a small enough dataset to poke at
05:00	LauJensen	really? interesting
05:00	LauJensen	guess I dont get to make fun at ejackson then :(
05:00	nathanmarz	LauJensen ejackson: at some point it will have a fork-join implementation which will be even better
05:01	ejackson	nathanmarz: that's awesome news
05:01	ejackson	LauJensen, nathanmarz : I'm glad not to be totally wrong - I usually default thus.
05:02	nathanmarz	it just runs hadoop in "local mode", which just means it's in process and there's no daemons to deal with
05:02	LauJensen	nathanmarz: Have you got some docs on this somewhere?
05:03	nathanmarz	on running it in local mode?
05:03	LauJensen	On the whole setup
05:04	nathanmarz	there's nothing to running it in local mode, just include it as a dependency in leiningen and open up a repl
05:04	nathanmarz	this post: http://nathanmarz.com/blog/news-feed-in-38-lines-of-code-using-cascalog.html talks about using it on a real cluster
05:05	nathanmarz	oh and you'll want hadoop as a dev dependency: [org.apache.hadoop/hadoop-core "0.20.2-dev"]
05:05	nathanmarz	setting up a good wiki is a priority over the next month :)
05:07	ejackson	nathanmarz: I saw from some tweets that you were playing with redis. Any joy ? I've been using it recently in clojure and have been v. impressed with it.
05:07	ejackson	persisting a sorted set directly is magic
05:08	nathanmarz	ejackson: i haven't played with redis at all... or tweeted about it
05:08	ejackson	hmmm......
05:08	nathanmarz	i know the runa guys use it
05:09	ejackson	i must be losing my mind, going to examine the logs.....
05:09	ejackson	(not with cascalog...)
05:10	ejackson	ugh, whatever.
05:10	LauJensen	nathanmarz: thanks
05:20	nathanmarz	LauJensen: sure thing
05:31	shortlord	it's not possible to set 2 validators for one single ref, is it? which means I would need to 'and' both functions, right?
05:40	ejackson	shortlord: that's what i'd do.
05:44	LauJensen	shortlord: Likely you'll use cond or condp and make a list of requirements for validation
05:45	LauJensen	or you could take ejacksons advice and simply let the validator log the event, upload it to Google Docs and then inspect the logs manually with your coworkers via a VPN connection implemented in C
05:47	shortlord	LauJensen: the google docs thingy is a hell of a great idea! thx a lot, I'll do that :)
05:47	LauJensen	well, thank ejackson :)
05:47	ejackson	lmao
05:47	shortlord	thx, ejackson :)
05:47	ejackson	your invoice in the post - BOTH of you
05:48	LauJensen	yea ejackson's a _real_ consultant :)
05:48	shortlord	^^
05:48	ejackson	LauJensen: i've got a _special_ rate for you buddy.
06:21	shortlord	is it possible to validate a ref based on the old and the new ref value?
06:23	bsteuber1	,(doc set-validator!)
06:23	clojurebot	"([iref validator-fn]); Sets the validator-fn for a var/ref/agent/atom. validator-fn must be nil or a side-effect-free fn of one argument, which will be passed the intended new state on any state change. If the new state is unacceptable, the validator-fn should return false or throw an exception. If the current state (root value if var) is not acceptable to the new validator, an exception will be...
06:25	bsteuber1	so I guess no
06:29	raek	shortlord: pre- and post-conditions for the state transition function could perhaps do something similar
06:30	jwr7	The docstring for future says "…yields a future object that will
06:30	jwr7	invoke the body in another thread, and will cache the result…" — cache it until when? What if my future returns a 1GB object and my heap is 3GB?
06:30	shortlord	raek: ok, that will do, thx
06:30	jwr7	assuming I don't hold on to the ref, will the result still be stored somewhere?
06:31	raek	jwr7: what it means is that evert time you dereference it you will get the same value (it is only computed once)
06:31	raek	it will only be around as long as you hold a reference to the future
06:31	jwr7	raek: oh, that's cool then.
06:32	raek	caches in java are often implemented with some variant of weak references, whose entries will be GC'ed when no-one else has a reference to the object
06:32	raek	so it probably wouldn't be a problem even if it was holding it in a cache
06:33	jwr7	raek: thanks. That sounds ok.
06:54	raek	I've dropped my old .emacs file and started to use emacs starter kit, but I'm getting stuck on one detail:
06:55	raek	when doing C-x C-f, it enters some fuzzy matching mode (whis is very nice) but sometimes I want to visit exactly what I have written
06:56	raek	e.g. create the file foo-project/src/util.clj instead of reopen my old bar-project/src/util.clj
06:56	raek	how can I do this? what am I missing?
06:56	Chousuke	hm.
06:57	Chousuke	it should allow that
06:57	Chousuke	just type the full path
06:58	Chousuke	and be quick enough that the search doesn't trigger :P
06:58	Chousuke	it has a timeout after you stop typing.
06:58	raek	when I finnish ".clj" it replaces the path with my other file
06:58	bsteuber1	raek: try C-f when in the minibuffer
06:59	bsteuber1	at least that works with my fuzzy completion there
06:59	raek	ah, wonderful!
07:00	raek	Chousuke, bsteuber1: thanks
07:00	bsteuber1	yeah, I was also quite happy after finding this :)
07:00	raek	being quicker-than-emacs worked, but C-f is much more convenient... :-)
07:07	Raynes	cemerick: Morning.
07:07	cemerick	morning :-)
07:09	raek	Raynes: regarding irclj: guess what... I got stuck on the error handling.
07:10	Raynes	raek: That statement was promptly accompanied by a rimshot by the house band. :p
07:10	bsteuber	when debugging gui apps from swank, closing the applications main window always kills the swank process - any idea how to prevent this?
07:11	tomoj	isn't there some flag you set for what to do on close?
07:12	tomoj	.setDefaultCloseOperation
07:13	tomoj	the default shouldn't be EXIT_ON_CLOSE, though...
07:13	raek	creating a JFrame, showing it and closing it does not do that for me
07:13	raek	the slime-swank session remains alive
07:15	bsteuber	hmm, so maybe my application does sth. special - it's quite big and chaotic (and not mine)...
07:16	tomoj	if it's swing, grep for EXIT_ON_CLOSE
07:16	tomoj	or DISPOSE_ON_CLOSE maybe
07:17	tomoj	..ON_CLOSE should do it
07:18	bsteuber	tomoj, raek: found it, there's a custom listener calling System.exit - so thank you :)
07:46	ejackson	hello cemerick
07:51	cemerick	ejackson: good afternoon :-)
07:52	ejackson	cemerick: indeed :D
07:55	cemerick	amalloy: pling?
08:41	edw`	Could someone give me an example usage of `longs'? I'm trying to create an array of unboxed numbers.
08:44	chouser	edw`: 'longs' just casts. you probably want (long-array [1 2 3])
08:45	edw`	Ah. Thanks.
08:53	Dranik	chouser, does long-array create an array for unboxed values?
08:54	chouser	,(long-array [1 2 3])
08:54	clojurebot	#<long[] [J@413be>
08:54	chouser	yep, an array of primitive longs.
08:59	Dranik	,(int-array [1 2 3])
08:59	clojurebot	#<int[] [I@1288e12>
08:59	Dranik	wow! :-)
09:07	Somelauw	I really don't like scala.
09:08	Scriptor	why not?
09:11	Somelauw	It is too complicated.
09:11	Somelauw	It's errors are unclear
09:11	Somelauw	And it is hard to read.
09:12	Somelauw	It looks ugly
09:12	LauJensen	http://michid.wordpress.com/2010/08/24/so-scala-is-too-complex/
09:13	Somelauw	And it is has too much impliciteness.
09:14	Somelauw	I will read it, but it probably won't convince me.
09:14	clgv	you do have example for what you claim? It would be interesting to see.
09:14	cemerick	LauJensen: That's a silly post. Verbosity ≠ complexity.
09:15	LauJensen	cemerick: They are very related, though I dont agree with that posts conclusion of course. Scala is waay too complicated
09:16	AWizzArd	LauJensen: but this guy from the blog was cheating a bit in the Java version.
09:16	AWizzArd	He could have made the fields public, so no getters/setters were needed.
09:16	cemerick	"Complicated for what" is the better question.
09:17	cemerick	LauJensen: But, verbosity and complexity are entirely orthogonal, if they're to mean anything at all.
09:17	AWizzArd	He could also import scala and call the same functions as in the Scala example.
09:18	AWizzArd	The Java code would still be longer, but way shorter than what he presented in his blog.
09:18	Somelauw	His java code, isn't idiomatic java, but he is just converting all scala functions to java first.
09:18	LauJensen	cemerick: Assembly code is very simple, every line does only one thing. But if I tasked you to read the source for Notepad written in Assembly, Im sure you'd think it was complex because of the shear vastness of code. So ytou see, they are related
09:19	LauJensen	http://grep.codeconsult.ch/2010/08/26/so-java-is-more-complex-than-scala-you-must-be-kidding/
09:19	Fossi	all there is to say about scala: "+: fail"
09:19	cemerick	LauJensen: No, I'd think it was verbose.
09:19	LauJensen	cemerick: You're too stubborn for reason
09:20	cemerick	LauJensen: Words mean particular things; just stating that A is related to B doesn't make it so.
09:21	cemerick	Perhaps one may not like a language because it tends to encourage/require verbosity, but that doesn't mean that the language is simple, or complex.
09:22	AWizzArd	LauJensen: yes good, this second post shows a more realistic scenario.
09:23	LauJensen	You guys remember this one?
09:23	LauJensen	life1d=: '_#'{~ (3(2=+/\) 0,],0:)^:a:
09:24	AWizzArd	Maybe one could say that complexity is introduced when there are "many cases".
09:24	LauJensen	AWizzArd: Generally speaking, you could say "complexity is introduced, shortly after cemerick is introduced" :)
09:24	Somelauw	LauJensen, Java code can be much shorter than your example suggests: http://pastebin.com/FkswZt96
09:25	Somelauw	It won't compile since it is quickly written.
09:25	LauJensen	Somelauw: You have to type slow for the compiler to work? crazy....
09:26	Somelauw	No, it won't compile since I am too lazy to debug it, but it demonstrates that java doesn't have to be written in the way your article suggests.
09:26	LauJensen	Its not my article
09:27	shortlord	can I use 'loop' without 'recur'? My function is not linear-recursive, so recur is not an option
09:27	Somelauw	You posted it.
09:27	Chousuke	shortlord: no.
09:28	Raynes	I like it when I'm looking for something that I figure wont exist only to find out that precisely what I'm looking for does exist. c.s/trim-newline for example.
09:28	Chousuke	you can call your function in the loop but it needs a recur to actually be a loop. :P
09:29	cemerick	shortlord: you can also recur to the head of your function, but I presume that's not suitable either
09:29	hoeck	Somelauw: nice, thats what I thought too
09:29	shortlord	Chousuke: so I'd have to do alle my bindings first, then create another function within the main function and call that inner function manually?
09:29	Chousuke	shortlord: you can call the main function too?
09:29	Chousuke	shortlord: ordinary recursion works normally, it just won't be optimised
09:29	shortlord	but I don't want the bindings to be overwritten
09:30	Chousuke	I'm not sure what you need to do :/
09:30	Chousuke	But using loop without recur sounds weird. :P
09:30	shortlord	I am doing some dereferencing in the first entry, which should not happen in the first recursion
09:30	LauJensen	shortlord: it sounds like you're on an imperative path
09:30	Chousuke	what point is there in looping if the loop variables are constant?
09:31	raek	you could use a helper function
09:31	shortlord	the loop variable is not constant
09:31	raek	...which would call itself
09:31	Chousuke	maybe you should come up with some example code illustrating what you need.
09:32	Chousuke	then it's easier to tell you if there's a better way.
09:33	shortlord	ok
09:33	shortlord	where can I find a good pastebin with clojure highlighting?
09:34	Chousuke	gist is good
09:34	cemerick	gist.github.com is good
09:34	Chousuke	:P
09:35	shortlord	https://gist.github.com/774476
09:36	shortlord	the last line is where the recursion should eventually occur, but the important thing is the 'let'
09:36	shortlord	I only want to do that binding once
09:36	shortlord	and not every time I use that function recursively
09:37	shortlord	'loop' sounds like the perfect match, but I can't use 'recur' since the function won't be linear recursive
09:37	shortlord	so I guess the only way to do is would be a helper function, right?
09:37	Chousuke	using an inner helper function sounds like it would work. Close it over the let binding
09:38	shortlord	Chousuke: ok, thx
09:41	fliebel	morning
09:41	fliebel	Hm, seems I'm back to fliebel again. Whatever.
09:45	ejackson	fliebel: you been dabbling in schizophrenia or MPD ?
09:45	fliebel	ejackson: No, just trying to / nick pepijndevos
09:46	cemerick	fliebel: it takes some doing to change default nicks.
09:46	cemerick	or, it took me a few days to figure out the magic combination between NickServ and my irc client.
09:47	fliebel	cemerick: Meh, probably not worth the effort and trouble.
09:57	Raynes	fliebel: What is 'fliebel'?
09:58	Raynes	What does the name mean, I mean.
09:58	fliebel	Raynes: Nothing, well, uuuhm. goes up to translate.google.com
09:58	Raynes	sexpbot can do that.
09:59	Raynes	$trans es en hola
09:59	sexpbot	hello
09:59	fliebel	$trans nl en fliebel
09:59	sexpbot	fliebel
10:01	fliebel	It's a bit of a made-up word, but what I usually mean with it, is 'something', usually something random, or a small extension, like a candle pit.
10:02	Raynes	Interesting.
10:02	fliebel	really?
10:02	Raynes	Americans don't make up words.
10:03	fliebel	Raynes: But you do have slang I believe? That's the same thing.
10:03	Raynes	Well, yeah.
10:05	markskilbeck	Americans don't make up words? "Color", "Humor", "Weapons of Mass Destruction"...
10:08	fliebel	markskilbeck: You're saying American is slang? ;)
10:08	gtrak	$trans en ru hello
10:08	sexpbot	привет
10:09	cemerick	Raynes: I daresay we're produce new words more efficiently than we should ;-)
10:10	Raynes	cemerick: Indeed. Slang completely slipped my mind.
10:11	cemerick	"Proper English" is slang until it's "proper". :-)
10:12	Raynes	Not isn't always the creation of new words as much as they are repurposing of old words. For example, "ice" is slang for crystallized methamphetamine.
10:17	fliebel	Python is hard, mainly due to misplaced braces and missing commas.
10:18	cemerick	__all__ _the under_scoring ended up bothering me after a while.
10:18	clgv	and messed up tabs ;)
10:19	fliebel	__these__ methods are magic though :)
10:19	cemerick	yeah, but you end up writing and using them a fair bit to do certain things
10:19	clgv	but there is a great library in python for algebra and other fields of computational mathematics :)
10:20	clgv	I did work with it 2 years ago while writing my thesis...
10:20	fliebel	Hm, I was looking for a specific property of Python lists, which they don't have. I thought I could do some_list[300] = 3 and it would do that on a list 5 items.
10:22	fogus`	Anyone have a favorite way to map a function to every other value in a seq (without using take-nth)? i.e. (map-eo - [1 2 3 4]) ==> [1 -2 3 -4]
10:23	fliebel	fogus`: Why not take-nth?
10:23	opqdonut	the simplest is just to write out map-eo using loop
10:23	fogus`	fliebel: Because that's what I already use and was hoping for a better way
10:24	fliebel	fogus`: ##(map (comp - last) (partition 2 [1 2 3 4 5 6]))
10:24	sexpbot	⟹ (-2 -4 -6)
10:24	mrBliss	not so pretty: ##(mapcat (fn [[a b]] [a (- b)]) (partition 2 [1 2 3 4]))
10:24	sexpbot	⟹ (1 -2 3 -4)
10:24	AWizzArd	If you want to do it index based then an index would be required. But if you have one then for can be used for that.
10:25	AWizzArd	for can combine map+filter.
10:25	opqdonut	,(map #(%1 %2) (cycle [inc identity]) [1 1 1 1 1])
10:25	clojurebot	(2 1 2 1 2)
10:25	opqdonut	I kinda like that
10:25	opqdonut	is there a better name for #(%1 %2) ?
10:27	AWizzArd	,(map-indexed #(if (even? %1) %2 (- %2)) [10 20 30 40])
10:27	clojurebot	(10 -20 30 -40)
10:27	AWizzArd	this could also work for you fogus`
10:27	fogus`	BTW, map-eo shouls also work like this: (map-eo - [1 2 3 4] [1 2 3 4]) ==> (1 0 3 0)
10:27	opqdonut	errrr
10:27	opqdonut	should it choose the first or the second?
10:27	opqdonut	consider (map-eo - [1 2 3 4] [5 5 5 5])
10:29	fogus`	It should work just like map, but only on e every other elem
10:29	opqdonut	...
10:29	AWizzArd	fogus`: what would happen for (map-eo - [10 20 30 40] [1 2 3 4]) ?
10:29	fogus`	(10 18 30 36)
10:30	AWizzArd	so, for even indexes it should just return the object from its first argument-sequence?
10:31	opqdonut	,(let [app (fn [f & rest] (apply f rest)) fst (fn [f & rest] f)] (map app (cycle [fst -]) [10 20 30 40] [1 2 3 4]))
10:31	clojurebot	(10 18 30 36)
10:33	fogus`	AWizzArd: right. First seq is special in that way
10:35	raek	fogus`: why include n/2 dummy values in the second seq?
10:36	raek	(map-eo - [10 20 30 40] [:dummy 2 :dummy :4])
10:37	fogus`	raek: ?
10:37	AWizzArd	fogus`: I would probably suggest to go with map-indexed if you have one argument seq and probably map when you have several and just do the checking yourself
10:38	AWizzArd	,(map-indexed #(if (even? %1) %2 (- %2)) [10 20 30 40])
10:38	clojurebot	(10 -20 30 -40)
10:38	AWizzArd	,(map #(if (even? %1) %2 (- %2 %3)) (range) [10 20 30 40] [1 2 3 4])
10:38	clojurebot	(10 18 30 36)
10:39	fogus`	AWizzard: oooo
10:40	opqdonut	(my solution is general!)
10:40	opqdonut	(yes, it's also somewhat horrible)
10:41	mrBliss	,(letfn [(map-eo [f & colls]
10:41	mrBliss	(apply map #(%1 %&) (cycle [first #(apply f %)])
10:41	mrBliss	colls))]
10:41	mrBliss	(map-eo - [10 20 30 40] [1 2 3 4]))
10:41	clojurebot	EOF while reading
10:42	mrBliss	I shouldn't copy paste it from a repl ##(letfn [(map-eo [f & colls] (apply map #(%1 %&) (cycle [first #(apply f %)]) colls))] (map-eo - [10 20 30 40] [1 2 3 4]))
10:42	sexpbot	⟹ (10 18 30 36)
10:42	AWizzArd	fogus`: although, if you require it often then maybe something like (def map-eo (fn [f & args] (apply map #(if (even? %1) %2 (apply f %2 %&)) (range) args))) would be handy
10:44	fogus`	These are pretty nice... but now how to make map-eo-nth :-O ;-)
10:45	AWizzArd	That is left for the interested reader.
10:46	ejackson	exercise for the student...
10:46	fogus`	LOL
10:46	ejackson	or, if it was really written by an academic: the extension to map-eo-nth is trivial.
10:46	fogus`	Thanks all. It was fun
10:47	AWizzArd	fogus`: plus you would need to specify if F should be applied to every nth element, or if every nth element is the one to which F will not be applied.
10:48	AWizzArd	(map-eo-nth 3 - [10 20 30 40 50 60 70 80 90]) ==> (-10 20 30 -40 50 60 -70 80 90) ?
10:48	AWizzArd	Or (10 -20 -30 40 -50 -60 70 -80 -90)?
10:48	fogus`	AWizzArd: Which makes me very happy that Clojure has avoided the proliferation of kwargs
10:49	fogus`	i.e. (map - [10 20 30 40 50 60 70 80 90] :from 0 :every 4)
10:50	AWizzArd	But what would be the result?
10:51	AWizzArd	From your map-eo I would interpret it as: F is not applied to every n-th element, but to all others.
10:51	AWizzArd	So (map-eo-nth 3 - [10 20 30 40 50 60 70 80 90]) should result in ==> (10 -20 -30 40 -50 -60 70 -80 -90)
10:53	fogus`	AWizzArd: No idea... I was just playing
10:54	Guest66680	does anybody know if clojureql can do "select ... where column in(v1,v2,v3...)"?
11:06	Guest66680	answer to own question: it can be done by using clojureql.predicates/in directly; clojureql.core/where doesn't seem to support it
11:30	zanes	I have some confusion about using Clojure with Emacs and swank. I just finished setting everything up and when I evaluate an expression with C-x e the output goes to the minibuffer rather than the REPL buffer. Is that normal?
11:30	mrBliss	zanes: perfectly normal
11:31	mrBliss	(although you probably mean C-x C-e :)
11:31	LauJensen	Guest66680: You're on a dated version, 1.0.0 does support (where (in ....))
11:33	fliebel	Funny, this morning I thought sorting was a strange thing. Sorting as I knew it involved pretty much comparing every element to every other element, while the way I sort a pile of cards is much simpler. So I wrote a Clojure function to do that, and then found out it's called pigeonhole sort. Anyway, here's my implementation: https://gist.github.com/774664
11:36	Berengal_	fliebel: Sorting can be pretty interesting
11:36	fliebel	Sad thing is that it is still slower than the built in sort.
11:38	Berengal_	What does the built-in sort use?
11:38	fliebel	No idea...
11:39	Scriptor	I wonder if it's optimized depending on what the list is like
11:39	Berengal_	Probably
11:39	fliebel	Berengal, Scriptor: http://download.oracle.com/javase/6/docs/api/java/util/Arrays.html#sort(byte[])
11:40	Berengal	Is that what clojure uses?
11:40	Scriptor	ah, quicksort
11:40	fliebel	Berengal: Yes
11:40	Scriptor	https://github.com/clojure/clojure/blob/b578c69d7480f621841ebcafdfa98e33fcb765f6/src/clj/clojure/core.clj#L2329
11:41	Berengal	fliebel: How does that work when clojure structures are trees?
11:41	Scriptor	there's the exact clojure code, looks like it basically wraps Array.sort
11:41	mrBliss	"The java.util.Arrays.sort function officially uses a Dual-Pivot Quicksort, though the OpenJDK implementation recently switched from a modified mergesort to a Timsort."
11:43	Berengal	I don't think the Timsort version has been officially released yet
11:43	mrBliss	I got it from this interesting comparison: http://stringoftheseus.com/blog/2011/01/10/api-sorting-algorithms/
11:45	Berengal	Seems there are some good benefits to be had from specializing sort on certain structures
11:45	Berengal	Particularly lists and vectors
11:46	fliebel	What would be then main reason my sorting thing is a lot slower than Java sort?
11:49	Somelauw	You can't quicksort a linked link, can you?
11:49	mrBliss	Somelauw: it would be slowsort :)
11:50	Berengal	Somelauw: Depends on how you define quicksort. You can't do the clever swapping during partitioning
11:50	mrSpec	Hello! I have function thats returns list of structures. But max-key doesnt takes list as argument, but structures. Is it possible to put all elements from list as argument into this function?
11:51	Somelauw	apply
11:51	mrBliss	,(apply max-key first [[1 2] [4 3] [2 3]])
11:51	clojurebot	[4 3]
11:51	mrSpec	thanks!
12:07	ymasory_	clojurebot: (println (count (str (filter (fn [ch] true) "abc"))))
12:07	clojurebot	Counterclockwise aka ccw at http://code.google.com/p/counterclockwise/
12:08	ymasory_	why would that be 26 and not 3?
12:08	mrBliss	,(str (filter (fn [ch] true) "abc"))
12:08	clojurebot	"clojure.lang.LazySeq@1ecc1"
12:09	mrBliss	,(println (count (apply str (filter (fn [ch] true) "abc"))))
12:09	clojurebot	3
12:10	ymasory_	i still don't understand
12:11	ejackson	,(str (filter (fn [ch] true) "abc"))
12:11	clojurebot	"clojure.lang.LazySeq@1ecc1"
12:11	ejackson	,(count (str (filter (fn [ch] true) "abc")))
12:11	clojurebot	26
12:12	mrBliss	ymasory_: str on a seq serializes it, i.e. "clojure.lang.LazySeq@1ecc1", you have to 'apply' str. ##(doc str)
12:12	sexpbot	⟹ "([] [x] [x & ys]); With no args, returns the empty string. With one arg x, returns x.toString(). (str nil) returns the empty string. With more than one arg, returns the concatenation of the str values of the args."
12:12	ejackson	you need to use apply like mrBliss says
12:12	raek	or use pr-str
12:12	ymasory_	i see thank you
12:12	raek	,(pr-str (filter (fn [ch] true) "abc"))
12:12	clojurebot	"(\\a \\b \\c)"
12:20	S11001001	,(String. (into-array Character/TYPE (filter (constantly true) "abc")))
12:20	clojurebot	"abc"
12:59	mrSpec	Can't I do: (subvec (sort ...) 0 5) ? What should I use instead of subvec?
13:00	qbg	(take 5 (sort ...))
13:01	mrSpec	o! thanks :)
13:01	hiredman	sort doesn't return vector, subvec works on vectors, why would you expect that to work?
13:04	mrSpec	yes, so I was looking for some other function which would work.
13:20	Berengal	Is there a way I could have a function drop into a repl and let me manually return a value?
13:24	opqdonut	(eval (read)) ?
13:25	opqdonut	for example, user=> (defn f [] (+ 1 (eval (read)))) (f) ENTER (+ 1 1) ENTER evaluates to 3
13:25	raek	read does not work very well if you try to use it in the emacs repl
13:26	opqdonut	ok, sure
13:26	Berengal	Nor is it a proper repl, where you can muck around as much as you like
13:27	opqdonut	well you can spawn a repl using some mechanism and then use refs or something for communication
13:27	raek	opqdonut: you could also use swing (consider the ask part): https://gist.github.com/480608
13:27	opqdonut	heh
13:27	raek	...in case that program has access to an X display
13:28	Berengal	raek: A safe assumption, considering it's mostly useful for development
13:28	opqdonut	but that's not a real repl either
13:28	Berengal	No, indeed not
13:29	raek	how would a real repl respond?
13:29	Berengal	(exit return-value)
13:29	Berengal	or something
13:29	raek	(def query)
13:29	opqdonut	Berengal probably wants to be able to evaluate a bunch of exprs and see the results and then say (the-final-answer-is foo)
13:30	Berengal	Yes.
13:31	raek	(let [p (promise)] (later-var-root #'query (constantly p)) (deref p)) ;; receiver end
13:31	raek	(deliver query 123) ;; sender end
13:31	Berengal	That would probably work just fine
13:32	raek	a bit ugly, since it only allows one pending query
13:32	Berengal	That can be hidden with some queuing and plumbing functions
13:32	Berengal	And it won't be a problem if there's only one thread anyway
13:33	Berengal	(It will be a problem if it's the same thread as the repl thread)
13:33	raek	you could perhaps generate a unique number for each, print a message to the user containing that, and let him answer with (answer <id> <value>) or something
13:33	raek	ok, just generalizing... :-)
13:35	Berengal	Well, implementing some queuing and a decent interface, such as switching from one query to another, automatically binding query-locals and whatnot should only be a couple of lines of code anyway
13:38	Berengal	Now I only need to figure out how to instantiate an arbitrary class with arbitrary args
13:39	raek	Berengal: you need java reflection to do that at run-time
13:40	Berengal	raek: Nothing like (apply .newInstance class args)?
13:41	Berengal	Oh, neat, java.lang.reflect.Constructor takes an Object[] :)
13:42	raek	no, since 'new'/dot-at-the-end is a special form
13:43	raek	if compile-time is fine, you could always make a macro that expands into a 'new' form
13:43	amalloy	Berengal: clojure.lang.Reflector does the things you want, yes?
13:43	Berengal	amalloy: Hadn't looked at that. Thanks for the tip
13:43	Berengal	raek: I know, but I was hoping there was an equivalent alternative
13:46	amalloy	,(clojure.lang.Reflector/invokeConstructor java.util.Date (into-array [10 10 10]))
13:46	clojurebot	#<Date Thu Nov 10 00:00:00 PST 1910>
13:46	amalloy	Berengal: tada!
13:46	Berengal	amalloy: nice :)
13:49	signalseeker	Can someone refer me to an uptodate post on setting up clojure with emacs/slime? Thanks
13:49	technomancy	signalseeker: if the swank-clojure readme doesn't cover it please mention what's missing/confusing on the mailing list.
13:51	signalseeker	technomancy: I was looking for a starting point because there are tons of blog posts about it. Let me give it a shot
13:51	raek	(jochu's repo is first on google, but technomancy's one is the up to date one)
13:53	raek	technomancy: are your changes always pulled into jochu's repo?
13:54	technomancy	raek: no; I need to push a deprecation notice to jochu's readme.
13:59	Berengal	Has anyone tried the clojure maven plugin on windows?
14:01	Berengal	I tried it today, but couldn't get it to start swank. I found a couple of issues on github regarding errors in command-line handling, but they were both solved
14:54	mrSpec	one more newbie question... what is the name of funtion remove-if in clojure? I've used it but forgot its name :(
14:55	hiredman	you mean remove?
14:57	mrSpec	sorry, not remove remove-if-not
14:57	ossareh	my love for -> / ->> grows daily
14:58	kotarak	(->> "Strange things happen" println (when fiasco))
14:58	mrSpec	I have list: '({:a 1 :b 2} {:a 3 :b 4}) and I'd like to get elements where :a is 1
14:59	raek	mrSpec: filter
14:59	mrSpec	thanks!
15:00	ossareh	mrSpec: filter, map, apply and reduce are really really helpful functions - definitely where I'd start a book where I writing one.
15:01	dakrone	,(filter #(= (:a %) 1) '({:a 1 :b 2} {:a 3 :b 4}))
15:01	clojurebot	({:a 1, :b 2})
15:21	danwerner	hi there
15:22	danwerner	i'm trying to find a var in clojure.core: letfn*
15:22	kotarak	letfn* is a special form, IIRC
15:24	danwerner	kotarak: that's interesting. do you know why it wasn't implemented as a macro?
15:24	LauJensen	hey danwerner
15:24	kotarak	danwerner: no. Probably because you can't get the mutual recursiveness from Clojure's other constructs.
15:25	danwerner	kotarak: yes, makes sense. thanks meikel
15:26	danwerner	LauJensen: hey lau. hope you got some calm & peace during the christmas celebrations
15:27	LauJensen	danwerner: I sure did - Especially since I dont celebrate christmas :) Have you put Clojure to good use at the work place yet?
15:31	danwerner	LauJensen: I might have found just the opportunity. we'll have to build a web scraping cron job thingy that doesn't depend on the rest of the code base, and isn't mission-critical, so I'll try to convince my superior to give Clojure a shot
15:31	LauJensen	Great, that sounds like a perfect fit
15:32	danwerner	LauJensen: Yes, especially with enlive -- I'm glad you convinced me to give it another try
15:33	LauJensen	Yea scraping with Enlive is really a treat
15:35	danwerner	is there a general word about how clojure-in-clojure is coming along?
15:35	LauJensen	danwerner: I dare not even ask anymore
15:43	fliebel	danwerner: Someone else has done a unofficial CinC.
15:44	danwerner	mhh, i wonder whether it would make sense to incrementally convert the java part of the implementation to deftype/def{interface,protocol}, a naive almost 1:1 port of the current structure that could then be refined
15:45	danwerner	fliebel: nice. do you have an URL or something?
15:45	fliebel	I'm searching
15:46	fliebel	ah! https://github.com/jarpiain/cljc
15:47	SergeyD	Slightly off-topic question: what does it mean "ANN:" in the beginning of a Clojure email subject?
15:47	danwerner	SergeyD: ANN = announcement
15:47	fliebel	danwerner: If you use it, please tell me how it goes, or blog about it or something.
15:47	SergeyD	danwerner: thanks :)
15:51	goodmike_mh	Does anyone know why I would be unable to require clojure.string? I'm in a swank session I started from lein swank, in a project that lists [org.clojure/clojure "1.2.0"] as a dependency.
15:52	goodmike_mh	but I get "Could not locate clojure/string__init.class or clojure/string.clj on classpath"
15:52	goodmike_mh	I swear I've been here before.
15:53	ossareh	goodmike_mh: could you throw your ns statement into a gist or pastebin or something?
15:54	ossareh	goodmike_mh: string is definitely available to you - likely it is either a syntax issue or a conflict in bringing it in (I think replace occurs in both string and core)
15:54	goodmike_mh	If I just try '(require '[clojure.string :as str])' at the REPL, it fails
15:54	goodmike_mh	(require '[clojure.string :as str])
15:55	goodmike_mh	I can pastie the project.clj
15:55	Licenser	goodmike_mh: be careful with that, str is a function :)
15:55	Licenser	cal it cs that is what I usually do
15:56	ossareh	goodmike_mh: the problem isn't going to be in project.clj - string is part of c.core
15:56	goodmike_mh	right
15:56	Licenser	anyway I need to go to bed :) take care and good luck with fingind that strings of yours
15:57	brehaut	Licenser: i didnt think requiring a namespace with an alias overrides a var with the same name?
15:57	brehaut	Licenser: or do you just mean its bad style?
15:57	ossareh	goodmike_mh: fwiw that form evals for me - so likely to be something else.
15:57	goodmike_mh	@Licenser It's common to see clojure.string required as str
15:58	ossareh	what ns are you in when you do that? is it a new ns?
15:58	goodmike_mh	E.g. the project I'm working in is fogus's Marginalia
15:58	goodmike_mh	user
15:59	danwerner	ossareh: did you try something like lein clean; lein deps to rebuild your dependencies? i seem to recall that accidentally having different clojure versions mixed in your libs can cause problems
15:59	ossareh	goodmike_mh: for you ^
15:59	goodmike_mh	danwerner: I'll try, thanks.
15:59	danwerner	oops ;)
15:59	ossareh	=)
16:02	goodmike_mh	danwerner: ran lein clean then lein deps then lein swank ...
16:02	goodmike_mh	In the terminal window I see Could not locate clojure/string__init.class or clojure/string.clj on classpath
16:03	danwerner	fliebel: thanks for the link to cljc. i'm looking for a way to write clojure on top of javascript, but cljc seems out of my league (complexity wise). especially not porting it to js :)
16:03	technomancy	goodmike_mh: what's clojure-version say?
16:03	fliebel	danwerner: There is already a basic project that does some Clojure to JS.
16:03	brehaut	goodmike_mh: eval clojure-version in your repl
16:03	danwerner	fliebel: i'm looking at https://github.com/kriyative/clojurejs right now
16:04	goodmike_mh	technomancy: {:interim true, :major 1, :minor 2, :incremental 0, :qualifier "master"}
16:04	fliebel	If you can get that to a state where it compiles Cljc, you're there, I think.
16:05	Berengal	So a promise looks very much like an I-var. Is there any discussion around adding M-vars too?
16:05	goodmike_mh	Does leiningen create your own sandbox of dependencies per project? Or could it be getting confused because of other versions of Clojure or other libraries?
16:05	danwerner	fliebel: it doesn't have persistant data structures, though, and really only translates a clojure-like language to javascript. (which in itself is pretty cool already.)
16:05	technomancy	goodmike_mh: it does, but clojure-version confirms you have the right one
16:06	goodmike_mh	I blame fogus for everything.
16:07	amalloy	mrSpec: i was at lunch, but it looks like you got only a vague answer re filter
16:08	amalloy	&(filter (comp #{1} :a) [{:a 1 :b 2} {:a 3 :b 4}])
16:08	sexpbot	⟹ ({:a 1, :b 2})
16:08	amalloy	mrSpec: ^ is the easiest and most idiomatic way to do what you were trying to do
16:10	goodmike_mh	So, here's what I did. Would someone like to run through these steps to reproduce (or not?)
16:10	goodmike_mh	1. git clone https://github.com/fogus/marginalia.git
16:10	danwerner	goodmike_mh: is the error message caused by "lein swank", which then refuses to run?
16:10	goodmike_mh	2. cd marginalia
16:10	goodmike_mh	3. lein deps
16:10	goodmike_mh	4. lein swank
16:11	technomancy	3 is actually unnecessary now
16:11	goodmike_mh	danwerner: error comes from swank
16:11	goodmike_mh	technomancy: really? lein swank will pull in deps?
16:11	technomancy	goodmike_mh: you only need to call deps manually when you change project.clj
16:12	technomancy	even that should be unnecessary in lein 2
16:12	goodmike_mh	but if you've just cloned a project repo, you would need to lein deps to start, right?
16:13	technomancy	no
16:13	goodmike_mh	wow
16:13	danwerner	goodmike_mh: followed your steps, got "Connection opened on local port 4005\n#<ServerSocket ...>
16:13	technomancy	goodmike_mh: just cloned marginalia; works for me
16:13	danwerner	goodmike_mh: i used lein 1.3.1 though
16:14	goodmike_mh	Oh well. The gods are against me.
16:14	technomancy	goodmike_mh: is it any different in lein repl vs swank?
16:14	goodmike_mh	good question...
16:14	goodmike_mh	no difference. same error message, although REPL starts.
16:16	goodmike_mh	danwerner: Leiningen 1.3.1 for me too.
16:16	danwerner	goodmike_mh: do you perhaps have any stale swank processes open that use the wrong clojure version?
16:16	technomancy	goodmike_mh: no reason not to run the latest; try "lein upgrade"
16:17	goodmike_mh	danwerner: killing my terminal windows and emacs...
16:17	goodmike_mh	running 'lein upgrade'
16:18	goodmike_mh	Now running Leiningen 1.4.2 on Java 1.6.0_22 Java HotSpot(TM) 64-Bit Server VM
16:18	danwerner	technomancy: lein 2 is not yet "stable", is it?
16:18	mrSpec	amalloy: Ok, thanks! :)
16:18	goodmike_mh	lein repl gives me the same error
16:18	technomancy	danwerner: it has maybe two commits on that branch so far
16:19	technomancy	fogus`: you want to set :eval-in-leiningen true in marginalia's project.clj if you want tests etc. to run in the same process as lein
16:19	danwerner	technomancy: lein swank gives "That's not a task" if i don't do "lein deps" before
16:20	technomancy	danwerner: oh right; that would only work if you had swank installed for your user
16:20	goodmike_mh	I just tried lein repl in a new project made from lein new, and no error message.
16:20	technomancy	but in general lein will know to run deps when it's needed
16:21	danwerner	goodmike_mh: "ls lib/" for me says "clojure-1.2.0.jar clojure-contrib-1.2.0.jar dev/ hiccup-0.3.0.jar markdownj-1.0.2b4.jar". is it the same for you?
16:21	goodmike_mh	in Marginalia?
16:22	goodmike_mh	clojure-1.2.0.jar dev markdownj-1.0.2b4.jar
16:22	goodmike_mh	clojure-contrib-1.2.0.jar hiccup-0.3.0.jar
16:22	goodmike_mh	yuck
16:22	danwerner	goodmike_mh: yes. "ls lib/dev/" is: "clojure-1.2.0.jar hiccup-0.3.0.jar jsch-0.1.42.jar lein-clojars-0.5.0-20091119.123056-6.jar markdownj-1.0.2b4.jar clojure-1.2.0.jar hiccup-0.3.0.jar jsch-0.1.42.jar lein-clojars-0.5.0-20091119.123056-6.jar markdownj-1.0.2b4.jar"
16:22	goodmike_mh	clojure-1.2.0.jar dev markdownj-1.0.2b4.jar clojure-contrib-1.2.0.jar hiccup-0.3.0.jar
16:23	goodmike_mh	danwerner: looks the same, yes
16:25	danwerner	mystifying
16:27	fogus`	technomancy: Is the lack of :eval-in-leiningen causing issues?
16:27	technomancy	fogus`: it's complicating things since there are two processes where there should be one.
16:28	technomancy	it's unlikely it's causing issues, but it could be masking the causes of issues like the above
16:28	fogus`	technomancy: sold
16:28	fliebel	fogus`: How far along is JoC?
16:29	fogus`	fliebel: done
16:29	fliebel	fogus`: "Softbound print: February 2011 (est.)"
16:29	cow_	can someone tell me what is the right function to compare to java byte arrays? (byte[])
16:29	fogus`	fleibel: Sounds about right to me
16:30	cow_	s/to/two/
16:30	sexpbot	<cow_> can someone tell me what is the right function two compare two java byte arrays? (byte[])
16:31	cow_	i have two separate arrays with the same contents and (=) returns false.
16:31	fogus`	technomancy: this is what you mean correct? https://github.com/fogus/marginalia/commit/f9f212ffbfa5e8f9d1556eeca240a20803d4aa80
16:31	fliebel	cow_: Uh? .equals?
16:31	technomancy	fogus`: quite
16:31	fogus`	Thank you sir
16:31	technomancy	fogus`: also, for plugins the convention is to put a one-liner in the namespace docstring and the longer version in the defn's docstring
16:31	technomancy	makes the lein help output prettier
16:32	technomancy	np
16:32	fogus`	OK. I will take care of those other changes tonight.
16:32	cow_	,(.equals (.getBytes "abc") (.getBytes "abc"))
16:32	clojurebot	false
16:32	amalloy	&(apply = (repeatedly 2 #(into-array Byte/TYPE (range 10))))
16:32	sexpbot	java.lang.IllegalArgumentException: argument type mismatch
16:33	technomancy	fogus`: minor point, but I thought I'd mention it.
16:33	cow_	(= (.getBytes "abc") (.getBytes "abc"))
16:33	cow_	,(= (.getBytes "abc") (.getBytes "abc"))
16:33	clojurebot	false
16:33	amalloy	cow_: java arrays don't implement equals by value
16:33	amalloy	try seq-ing them first
16:33	cow_	amalloy: right... i'm trying to discover if there is a function
16:33	cow_	ahhh... ok
16:33	amalloy	&(= (seq (.getBytes "abc")) (seq (.getBytes "abc")))
16:33	sexpbot	⟹ true
16:34	amalloy	cow_: kinda a hack imo, but it will work
16:34	cow_	amalloy: thanks. i forgot all about that
16:34	goodmike_mh	I can require other core libraries, e.g. (require '[clojure.walk :as walk]) is fine.
16:35	goodmike_mh	what's special about string?
16:35	danwerner	goodmike_mh: IIRC it got promoted to core in 1.2. what happens if you require clojure.contrib.str-utils2?
16:36	goodmike_mh	danwerner: It works fine
16:36	goodmike_mh	(require '[clojure.contrib.str-utils2 :as foo])
16:36	goodmike_mh	no errors
16:37	danwerner	goodmike_mh: okay, works for me as well. it's deprecated but hasn't been removed yet
16:39	goodmike_mh	Thanks, danwerner and technomancy, for the help. I'll try coming back to the problem in a bit.
16:41	dakrone	,(.compareTo (java.nio.ByteBuffer/wrap (.getBytes "abc")) (java.nio.ByteBuffer/wrap (.getBytes "abc")))
16:41	clojurebot	0
16:41	dakrone	,(.compareTo (java.nio.ByteBuffer/wrap (.getBytes "abc")) (java.nio.ByteBuffer/wrap (.getBytes "nth")))
16:41	clojurebot	-1
16:41	dakrone	cow_: ^^
16:42	amalloy	dakrone: goodness, that seems like much heavier machinery than necessary
16:43	dakrone	amalloy: may be :)
16:54	TakeV	I have a list of structs, and each struct has a list of hash maps. How do I filter the top list so that I get just a list containing structs with a certain component in their hash maps?
16:55	goodmike_mh	OK, so when I unpack the clojure JAR file in my project's lib directory, there is no string.clj in there
16:55	amalloy	TakeV: this sounds like what mrSpec was doing an hour or two ago
16:56	danwerner	goodmike_mh: wow
16:56	goodmike_mh	yeah
16:56	goodmike_mh	I must be special.
16:58	danwerner	goodmike_mh: do you have string$*.class ?
16:58	goodmike_mh	nope
16:59	TakeV	amalloy: Hmm, I don't see it.
16:59	amalloy	TakeV: do you really mean each struct has a list of hash-maps (like foo is {[{:a 1} {:b 2}]}), or do you mean something more normal like the values in your struct are hash-maps?
16:59	danwerner	goodmike_mh: i'd rid my ~/.m2/repository/org/clojure directory from all evil by deleting it
16:59	goodmike_mh	hmm. decisive action.
16:59	danwerner	goodmike_mh: then lein clean; lein deps would be forced to acquire new jars
16:59	TakeV	amalloy: Sorry, each struct has a list of hash-maps.
17:00	amalloy	TakeV: and that one list is always under the same key, like :maps or something?
17:00	danwerner	is there any way to import functions marked as :private true in another namespace?
17:01	danwerner	'use and 'require both complain
17:01	TakeV	amalloy: Yes. :components
17:01	qbg	danwerner: Sounds like a bad idea
17:02	TakeV	amalloy: I want to split off a list from the top level by what components the bottom levels has, update them, then remerge them back into the main list.
17:02	danwerner	qbg: i'd like to use clojure.core/assert-args without copying it
17:03	goodmike_mh	danwerner: nuking the maven repo file seems to have worked.
17:03	qbg	Its probably a bad idea to use something that is not part of a documented interface
17:03	qbg	But if you really need to, you can go through the var
17:03	danwerner	goodmike_mh: cool. no missing clojure.string anymore?
17:03	goodmike_mh	It can't hide from me anymore!
17:04	amalloy	(filter (comp #(some #{:a} (mapcat keys %)) :components) all-the-structs) maybe?
17:04	qbg	Though I'm not sure how well that will work considering it is a macro
17:04	qbg	Probably won't
17:06	amalloy	TakeV: ^
17:06	qbg	You could do (def ^{:macro true} assert-args @#'clojure.core/assert-args)
17:06	TakeV	amalloy: Sorry, trying to figure out what that does.
17:08	amalloy	heh, me too
17:08	TakeV	amalloy: What does "#{:a}" do?
17:08	amalloy	&(class #{})
17:08	sexpbot	⟹ clojure.lang.PersistentHashSet
17:08	danwerner	qbg: nice hack. i've seen somebody else use assert-args before, and he copy&pasted it as well. might be worthwhile to add it to the public interface since it's so useful
17:08	amalloy	&(map #{:a} [:a :b])
17:08	sexpbot	⟹ (:a nil)
17:09	amalloy	TakeV: constructs a set containing the single keyword :a. when called as functions, sets return something truthy iff the set contains the function argument
17:10	TakeV	amalloy: Oh, right. Sorry, I was parsing that as a fn macro.
17:11	S11001001	&(#{false} false)
17:11	sexpbot	⟹ false
17:11	S11001001	false
17:11	amalloy	S11001001: yeah, i know. very sad
17:11	amalloy	so if you want to include false, you can put a (complement nil?) at the front of that comp
17:11	S11001001	better to remember that it returns the thing, then
17:12	S11001001	or just use (partial contains? #{...})
17:14	tomoj	amalloy: would you rather sets returned booleans?
17:15	amalloy	tomoj: i'm not sure. it seems rare that i use the self-returning feature, and common that i use it in a filter, but that could be sampling bias
17:16	tomoj	seemed rare to me too
17:17	amalloy	same thing with (some), really - i never want (f x) back, i always want x
17:17	amalloy	can call f my damn self if i want :P
17:20	Berengal	raek: That drop-into-repl thing? Works like a charm
17:26	Berengal	Are there any persistent dequeues around?
17:30	amalloy	Berengal: chouser's finger trees would work, but i'm not sure how stable they are
17:30	amalloy	$google chouser github finger tree
17:30	sexpbot	First out of 30 results is: Chouser/talk-finger-tree - GitHub
17:31	sexpbot	https://github.com/Chouser/talk-finger-tree
17:31	amalloy	Berengal: it's actually at https://github.com/Chouser/finger-tree - not sure why google finds the wrong one
17:32	qbg	Maybe the talk is more interesting than the code
17:32	amalloy	qbg: chris is an interesting talker
17:32	Berengal	They aren't already in core/contrib?
17:32	Berengal	Ooh, talk, where?
17:32	amalloy	Berengal i don't think they are, no
17:33	qbg	I think they the finger trees might move to contrib after 1.3
17:36	Berengal	I can't really depend on something that isn't automagically downloaded...
17:36	arohner	is there a preduce anywhere?
17:36	Berengal	Do vectors support easy adding at the front?
17:37	tonyl	list do
17:38	Berengal	But lists don't support easy adding at the end
17:38	amalloy	Berengal: it's probably in clojars
17:39	amalloy	well, i guess not
17:39	Berengal	amalloy: I couldn't find it there
17:39	S11001001	well "easy", if it just does (vec (cons elt vec)) then what is it doing really?
17:40	S11001001	hmm, there's a binding problem there
17:40	amalloy	Berengal: you can download it and put it there yourself, if you want to make it available for auto-download
17:40	Berengal	amalloy: What, even though I'm just some random person on the internet?
17:41	amalloy	Berengal: indeed!
17:41	Berengal	Quite the free-for-all :)
17:42	amalloy	Berengal: just don't make it look like an official finger-tree release
17:42	amalloy	put it under berengal/finger-tree instead of (chouser\|finger-tree)/finger-tree
17:42	Berengal	Ah, ok
17:43	Berengal	Eh... Cons doesn't implement IPersistentStack?
17:44	Berengal	Guess that means it's time for bed
17:45	tomoj	S11001001: seems if you call vec at all you've already lost
17:49	amalloy	tomoj: ##(nth (iterate vec (range 100)) 10000) doesn't cost too much :)
17:49	sexpbot	⟹ [0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99]
17:52	tomoj	that's a very short vector
17:52	amalloy	tomoj: but i suspect (vec [...]) is a no-op
17:52	tomoj	(vec (conj v x)) can be hundreds of times slower than just (conj v x)
17:52	tomoj	for a long vector
17:52	tomoj	I would've thought so too
17:52	tomoj	maybe I screwed my test up
17:53	tomoj	,(let [v (vec (repeat 1000000 0))] (time (when (vec (conj v 0)))) (time (when (conj v 0))))
17:53	amalloy	$source vec
17:53	sexpbot	vec is http://is.gd/kAbaj
17:53	clojurebot	Execution Timed Out
17:53	qbg	,(class (vec [1 2 3]))
17:53	clojurebot	clojure.lang.PersistentVector
17:54	qbg	,(class (vec '(1 2 3)))
17:54	clojurebot	clojure.lang.PersistentVector
17:54	danwerner	is there a manual for sexpbot/clojurebot somewhere? =)
17:54	amalloy	tomoj: wow, you're right
17:54	amalloy	$help
17:54	sexpbot	You're going to need to tell me what you want help with.
17:54	amalloy	bah
17:54	tomoj	,(apply identical? ((juxt identity vec) [1 2 3]))
17:54	clojurebot	false
17:55	ossareh	http://github.com/Rayns/sexpbot/wiki/Plugin-quick-start-guide
17:55	amalloy	danwerner: usage manual or dev manual?
17:55	danwerner	amalloy: usage, e.g. what's the difference between & and ##
17:56	amalloy	danwerner: & only works at the beginning of a string, whereas ##(inc 1)-style requests are allowed ##(inc 2) anywhere
17:56	sexpbot	(inc 1) ⟹ 2
17:56	sexpbot	(inc 2) ⟹ 3
17:58	danwerner	amalloy: thanks
17:58	amalloy	danwerner: np. i added the ## feature, and i'm more or less the only person who uses it :)
17:59	danwerner	amalloy: well, ##(identity "not anymore") ;-)
17:59	sexpbot	⟹ "not anymore"
18:00	ossareh	#("image we only spoke through sexpbot")
18:00	ossareh	fail.
18:00	amalloy	ossareh: it's so hard to hit the same key twice in a row, eh?
18:01	ossareh	hah, that or I really don't know the difference between 3 and 7
18:01	danwerner	##"any form"
18:02	amalloy	danwerner: actually ## only triggers on ##(
18:02	danwerner	##(str "ah" "not" "any" "form")
18:02	sexpbot	⟹ "ahnotanyform"
18:02	amalloy	it's on my todo list
18:02	amalloy	for sometime in 2011
18:06	danwerner	amalloy: have you looked at the clojurebot source II lately? it now uses conduit for its data shoveling
18:08	amalloy	danwerner: not lately, but i remember seeing the conduit thing
18:10	danwerner	amalloy: most of the people i've asked who used monads in clojure at times, eventually discontinued to do so. have you used monads in clojure yet?
18:11	amalloy	no. read about them a few times and sorta understand what the idea is, but i haven't internalized them enough to notice when i ought to use them
18:33	raek	with all the on-by-default goodies of emacs-starter-kit, I am king of the world.
18:36	danwerner	raek: does that include proper paredit support for clojure? i remember i had to patch something up manually
18:37	raek	yes, it includes fixes for { }
18:37	raek	at least that was what I had to patch
18:38	arohner	is there a preduce implemented anywhere?
18:39	amalloy	arohner: preduce can't work unless your function is associative, so i don't think there's a built-in
18:39	amalloy	it might even have to be commutative? i'm not sure
18:40	arohner	amalloy: yes, it has to be commutative. That's pretty standard. Google's map reduce has no problems running massively parallel
18:40	arohner	rhickey has a couple of side projects that contain preduce, but both require jsr166, which I can't find anywhere
18:48	ossareh	arohner: is jsr166 the fork/join stuff?
18:48	arohner	ossareh: yeah
18:50	ossareh	http://gee.cs.oswego.edu/cgi-bin/viewcvs.cgi/jsr166/src/jsr166y/
18:50	ossareh	arohner: ^
18:51	arohner	ossareh: thanks. I've seen that, still trying to figure out how to make clojure happy with it :-(
19:00	danwerner	arohner: i've never tried it, but it should be sufficient to put http://clojure.googlegroups.com/web/jsr166y.jar on your classpath
19:02	arohner	danwerner: aha, thanks
19:02	arohner	danwerner: that's a different version from the 'official' one
19:06	danwerner	arohner: in what way different?
19:06	arohner	danwerner: the ParallelArray.class is in a different namespace/package
19:07	arohner	apparently the up to date official version has it in a different location than it was when rhickey built that one
19:07	danwerner	i see
19:07	arohner	and clojure.parallel imports the old location
19:08	danwerner	arohner: have fun then, and a good night
19:09	arohner	danwerner: I will! thanks for the tip
19:26	Zane19	hi
20:32	a_strange_guy	/msg NickServ identify w3r5z7
22:00	currentB	so if I want a program to continually print lines while it's running, what's the best way to avoid the nil that's being returned to constantly be shown in the repl?
22:01	amalloy	currentB: don't evaluate that line at the repl top-level?
22:01	currentB	k
22:01	amalloy	&(do (println 10) "not nil")
22:01	sexpbot	⟹ 10 "not nil"
22:01	amalloy	&(do (println 10))
22:01	sexpbot	⟹ 10 nil
22:13	amalloy	chouser: is https://github.com/Chouser/finger-tree the state of the art on finger trees, or has development moved somewhere else?
22:16	dnolen	amalloy: I thought fingers-tree were here, https://github.com/clojure/data.finger-tree
22:17	amalloy	dnolen: aha, thanks
23:46	gtech	if you evaluate a buffer in emacs of a series of expressions is it normal that only the last output is printed?
23:46	gtech	last expression*
23:53	gtech	ah using, pr doesn't newline so it overwrites the output