#clojure – 2008-12-08

2008-12-08

00:19	Kerris7	finished red black tree
00:20	Kerris7	oops wrong channel
03:04	Lau_of_DK	Good morning gents
03:22	hiredman	show
03:22	hiredman	er
03:22	hiredman	clojurebot: show
03:22	clojurebot	show is http://groups.google.com/group/clojure/msg/96ed91f823305f02
03:39	AWizzArd	When I have a (def x (new javax.swing.JFrame)) object visible as a window, close it and do (def x nil), will this object then get gc'ed sooner or later? Or does Java somehow keep track of those guys and they live long?
03:41	hiredman	gui stuff is weird, because whatever windowing system also can have a ref
03:41	hiredman	(I don't know)
03:47	AWizzArd	hiredman: why don't you sleep? ;-)
03:47	arbscht	we didn't hire him to sleep!
03:47	AWizzArd	weren't we just talking about find-factors and how fast it runs on our systems?
03:48	AWizzArd	(or: "how slow it runs when it is the Clojure version" g)
03:52	hiredman	eh?
03:52	hiredman	I slept, then someone called me (the bastards), so now I am awake
03:55	AWizzArd	Okay, so like me you got only a few moments of sleep.
03:56	hiredman	it was a few hours I think
03:56	AWizzArd	4-5
03:56	hoeck	hi all, for the germans here: has anyone tried a clojure version for http://blog.fefe.de/?ts=b7c295e7 ?
03:57	hiredman	neo4j sounds cute, maybe clojurebot can use it
03:59	hiredman	clojurebot: neo4j is http://github.com/JulianMorrison/neo4j-clojure/tree/master
03:59	clojurebot	Ack. Ack.
04:29	hiredman	svn rev 1145
04:31	hiredman	svn rev 1145
04:31	clojurebot	svn rev 1145; made RT.DEFAULT_COMPARATOR use Util.compare
04:31	hiredman	clojurebot: botsnack
04:31	clojurebot	thanks; that was delicious. (nom nom nom)
06:41	AWizzArd	Is there something in Java/Clojure for variably adding data to an url? Like (make-url ip port data), for example: (make-url "localhost" 8080 {:name "David" :lang "Clojure" :age 24}) ==> "http://localhost:8080/?name=David&lang=Clojure&age=24"
06:49	AWizzArd	hmm, probably java.net.URI comes closest
06:49	AWizzArd	Hi black
06:49	blackdog	hi AWizzArd how's rife?
06:53	AWizzArd	I didn't continue yet. My next step is to write a report tonight and post it to the Rife mailing list, so that the author of Rife can reproduce my problem.
06:53	AWizzArd	It will probably need some .java code in Rife to make Clojure accessible.
06:54	blackdog	yuck :)
06:56	AWizzArd	What is this # sign in (.toString (new java.net.URI "http" "foo.com" "/hello world/" "name=blackdog"))?
06:56	AWizzArd	==> "http://foo.com/hello%20world/#name=blackdog"
06:56	blackdog	an html anchor?
06:58	AWizzArd	ok, probably right
07:04	AWizzArd	blackdog: I am not interested in these fragment parts of an URI. Unfortunately both constructors that allow me to specify a query also take a fragment string. How can I get rid of that # in the uri?
07:07	blackdog	why don't you just pass a sinlge string
07:07	blackdog	there is a constructor for that
07:08	AWizzArd	because I want to let java correctly encode my variable data. I don't know how many vars I want to put into the url. Sometimes it is just name=blackdog, but sometimes also age=29 and lang="a b c"
07:10	AWizzArd	(.create java.net.URI "http://foo.com/path?var=some spaces") ==> IllegalArgumentException
07:12	AWizzArd	But the URI class can do this correctly for me, and I can call (.toASCIIString on-it) and get the correctly encoded result, even for unicode chars in my data.
07:14	blackdog	I use org.apache.commons.httpClient.util.URIUtil.encodeAll
07:14	blackdog	there must be something similar in the jdk but i was using that anyway
08:14	gnuvince	morning
08:14	AWizzArd	Hi Vince
08:16	AWizzArd	Is there something like CLs equalp or string-equal for strings? (string-equal "clojure" "CLoJure") ==> true
08:18	duck1123	AWizzArd: have you been doing any gen-class yet?
08:20	AWizzArd	yes
08:20	duck1123	I see you said something about needing to write Java code. I was wondering if you tried just using gen-class and discarded that approach.
08:21	AWizzArd	duck1123: the web development framework Rife (for Java) can read .class or .java files. I wanted to make it work with Clojure. As a first step I hoped that .class files, compiled by Clojure would be accepted. But that wasn't the case.
08:22	duck1123	I'm still having a hell of a time getting gen-class to produce a class file that can be read by Tomcat
08:22	AWizzArd	For me it is a .class file for Jetty (+ Rife)
08:22	AWizzArd	doesn't work for me right now
08:23	AWizzArd	Although I don't know if it is Jetty or Rife that can't handle Clojures .class files.
08:23	duck1123	AWizzArd: looks like we may be getting hit by the same problem. (whatever that may be)
08:23	AWizzArd	maybe
08:24	duck1123	have you tried running your class through javap?
08:24	AWizzArd	blackdog was able to make Jetty run with .clj files, by starting Jetty from Clojure, instead of going the other way, as I do.
08:24	AWizzArd	No, never heared of javap
08:24	duck1123	That's how Compojure works
08:24	duck1123	javap is the java classfile decompiler
08:24	duck1123	you give it the name of your classfile, and it spits out the method signatures
08:25	AWizzArd	Compojure works by making use of javap or by starting Jetty/Tomcat from Clojure?
08:25	duck1123	Compojure starts jetty within clojure
08:25	duck1123	and makes use of proxy to generate the classes
08:26	AWizzArd	duck1123: can you try to make your example run in Jetty instead of Tomcat? Do you still get the same error message(s) then?
08:26	AWizzArd	I remember that I got some error that said that my HelloWorld class can't be instantiated
08:27	duck1123	I was just thinking about giving that a try. My application is not giving me an error. (which is frustrating) it's just refusing to load and gives a 404
08:28	AWizzArd	Rife tells me about some error messages at least, which I will soon send to Rifes mailing list. Maybe the author knows more about what to do.
08:28	duck1123	AWizzArd: the reason I mentioned javap is because I'm getting a method signature of "public static {};" which looks really weird.
08:29	AWizzArd	Did you ask rhickey about that?
08:30	duck1123	no, he wasn't around when I discovered it yesterday. When I see him, I'll make sure to ask
08:31	duck1123	it's odd, because all of the other signatures are of the form: public java.lang.String toString();
08:34	AWizzArd	Possible that we have the same challenge right now.
08:35	AWizzArd	I just have not much time to look into this for some time sigh
08:36	rhickey	duck1123: isn't that just a static initializer?
08:37	duck1123	rhickey: yeah, looking at the decompiled code, it looks like it interns all the vars
08:38	duck1123	I've just never seen anything like that before, and it was the only thing that stood out for me as being out of place
09:26	Chousuke	AWizzArd: you asked about string comparison earlier: java strings have the .equalsIgnoreCase method.
09:30	AWizzArd	thx Chou
09:33	Chousuke	It's worth looking through the java docs for types such as maps, strings and numbers. the java api provides a lot of functionality that isn't mentioned in the clojure docs.
09:34	duck1123	AWizzArd: Jetty is giving me an error. (which is actually a positive thing) Looks like I have an IllegalArgumentException
09:35	AWizzArd	You are getting closer!
09:36	AWizzArd	Chous: yes thanks, I will look at this more often. I was not used to the javadocs so far, but instead I used http://www.lispworks.com/documentation/HyperSpec/Front/index.htm :-)
09:37	AWizzArd	How can I remove the first occurence of a char in a string efficiently?
09:38	Chousuke	strings are immutable, so that would most likely involve copying the string in any case.
09:38	AWizzArd	right
09:39	AWizzArd	I just would like to do it in one run, and not traverse it several times
09:39	duck1123	(apply str (rest (seq "Hello"))) perhaps?
09:40	walters	AWizzArd: i wouldn't worry about those kinds of things unless it's an actual bottleneck in your app
09:40	Chousuke	if you want to remove just the first character the substring method is probably the best.
09:42	Chousuke	It seems StringBuffer is what you want.
09:42	Chousuke	it has a delete method.
09:43	blackdog	ii think stringBuilder is meant to be faster than StringBuffer
09:44	Chousuke	StringBuilder isn't thread-safe though.
09:45	blackdog	slurp and str use stringbuilder
09:45	Chousuke	well, they don't use the instance across multiple threads I suppose
09:46	blackdog	depends on AWizzArd usage I guess
09:46	Chousuke	but other than the thread-safety difference, sbuilder and sbuffer are identical, so pick whichever works
10:00	rhickey	don't use StringBuffer anymore
10:03	duck1123	interesting... I have a Hello World in jetty
10:05	Chouser	duck1123: still no love fromtomcat?
10:05	duck1123	I don't know yet. I need to move my code back home and try
10:06	Chouser	oh, ok.
10:08	duck1123	but it works on jetty 6.1.14 on Ubuntu... That's farther than I've gotten before
10:09	AWizzArd	Why no StringBuffers? Will they be deprecated?
10:10	rhickey	StringBuilder is proof that StringBuffer is wrong
10:14	duck1123	Chouser: still no love from Tomcat :(
10:14	rhickey	duck1123: what happens?
10:15	duck1123	rhickey: I'm getting a 404 when I try to load the app
10:16	duck1123	No errors that I can see. I can load static content, and manager shows the app as being loaded
10:17	rhickey	duck1123: that could be anything
10:17	rhickey	is this a simple hello world or a more involved thing?
10:17	AWizzArd	What is the function called that makes (? :a :b :c 10 20 30) ==> {:a 10 :b 20 :c 30}?
10:18	AWizzArd	i have used it yesterday but can't remember its name
10:19	Chouser	(doc zipmap)
10:19	clojurebot	Returns a map with the keys mapped to the corresponding vals.; arglists ([keys vals])
10:20	AWizzArd	thanks Chous
10:20	lisppaste8	duck1123 annotated #71768 with "latest attempt at getting Tomcat to work" at http://paste.lisp.org/display/71768#2
10:39	RSchulz	Is there something special about function names that begin with a hyphen? Or is it just a naming convention of some sort?
10:40	duck1123	that's how genclass identifies the functions
10:40	duck1123	the default prefix is "-"
10:40	RSchulz	What is the difference between those with and those without the hyphen?
10:43	duck1123	I think gen-class only looks for implementations of the for (str prefix methodname) other functions are just normal functions
10:43	rhickey	RSchulz: when you gen some class Foo, with a method bar, and say the implementing ns is my.baz, then the Foo stub will look for my.baz/-bar at runtime when bar is called. If you define it, it will get called. You can still define anything else you want in my.baz. You can pick any prefix you want with :prefix
10:44	RSchulz	So the prefixed functions become public methods in the generated class? Is that it?
10:45	RSchulz	It establishes the connection between Clojure functions and generated class methods?
10:45	duck1123	if it's an override of the base class, or specified in :methods
10:45	RSchulz	The first time I saw it, I thought it signified static methods, 'cause it was a "main" that had it.
10:45	duck1123	check out http://clojure.org/api#toc247
10:46	RSchulz	And then I guessed it meant private (as in (defn- ...))
10:46	rhickey	http://clojure.org/api#gen-class works too
10:46	RSchulz	I did read that, but I didn't grok it.
10:46	rhickey	RSchulz: well, there is a hint of that interpretation - the prefix is used to prevent collisions
10:47	RSchulz	What happened to the distinction between generating and writing a class file? Didn't there use to be two differen gen-...-class functions?
10:48	rhickey	but nothing you write in the implementation shapes the stub, the gen-class drives everything
10:48	rhickey	RSchulz: gen-and-load-class engendered bad behavior and false expectations
10:48	cemerick	rhickey: I noticed that the Reflector requires that all fields/methods it invokes are public. That caught me by surprise when I was trying to invoke a method on an instance of a package-private class from a clojure impl of a genclass in the same package. It'd be nice if the invocation context were threaded through the reflector usage.
10:50	rhickey	cemerick: this is a reflection limitation - there is no threading of context through reflection
10:51	cemerick	rhickey: you mean just of the clojure impl, right? I seem to vaguely remember using reflection to forcefully invoke a private method on an object many years ago...
10:52	rhickey	cemerick: you can't with default privileges call anything other than public members via reflection
10:55	cemerick	ah, I forgot about ReflectPermission
10:56	drewolson	hey all, i'm trying to read an rss feed from my internal cruisecontrol box, any idea why i'm getting nothing from this: http://gist.github.com/33497
10:56	drewolson	it just hangs
10:56	drewolson	and i can curl the url just fine
10:57	cemerick	rhickey: isn't that only relevant if the jvm was configured to use a security manager?
10:59	danlarkin	drewolson: works for me with http://www.reddit.com/.rss
10:59	drewolson	danlarkin: hrm, strange...maybe java is being funky with respect to the proxy
10:59	drewolson	although it's an internal site
10:59	drewolson	and i don't have my http_proxy env var set
11:00	duck1123	authentication, maybe?
11:00	drewolson	duck1123: naw, i can hit it from ruby just fine
11:00	drewolson	using open-uri
11:03	AWizzArd	Is there something similar to doseq, but which stops as soon one of the sequences is traversed (like map). (doseq [x (range 5) y (repeat 0)] (println x y)) <-- this should print 5 lines
11:05	Chouser	AWizzArd: you could add ':while x', though I guess that's not exactly the same thing.
11:05	rhickey	cemerick: I'm really not interested in going there
11:06	Chouser	oh, wait, that doesn't even make sense.
11:06	AWizzArd	Chouser: I think this won't help, as x will begin to cycle from the beginning.
11:06	Chouser	y is an inner loop -- goes forever.
11:06	AWizzArd	yes
11:07	Chouser	so what do you want again? When should y stop?
11:07	Chouser	or do you want x and y traversed in parrallel, instead of nested?
11:07	cemerick	rhickey: fair enough. I might submit some patches to make the associated error msgs more descriptive, then.
11:07	AWizzArd	yes, like (map vector (range 10) (repeat 4)), but for doing side effects
11:07	AWizzArd	(such as appending to a StringBuilder)
11:08	AWizzArd	Maybe there exists already some function to do that.
11:08	Chouser	AWizzArd: currently I think the only thing that takes seqs in parallel like that is 'map' (or I guess zipmap)
11:09	Chouser	AWizzArd: so: (doseq [[x y] (map list (range 5) (repeat 0))] ...)
11:10	AWizzArd	sure, I can work around it, just thought there could be something like (domap ...)
11:36	hiredman	clojurebot: latest?
11:36	clojurebot	latest is 1147
11:36	hiredman	svn rev 1147
11:37	clojurebot	svn rev 1147; added unchecked-remainder
11:42	AWizzArd	Oh, that is cool stuff!
11:43	AWizzArd	Chouser: do you know who added the unchecked-remainder?
11:45	duck1123	I wonder why the github mirror is falling behind. It used to be really good, but it says the latest is 1146
11:46	hiredman	AWizzArd: I think it was rhickey, he was in last night with his own factor using all the unchecked stuffs
11:51	danlarkin	interesting talk: http://www.infoq.com/presentations/click-fast-bytecodes-funny-languages
11:54	Chouser	For the record, only Rich can add anything to clojure svn, and I wouldn't expect that to change anytime soon.
11:55	Chouser	If what he adds is entirely or mainly from a patch submitted to the google group, he's careful to give credit in the checking comment. No attribution generally means it all his.
13:16	clojurebot	svn rev 1147; added unchecked-remainder
13:16	clojurebot	svn rev 1147; added unchecked-remainder
13:17	clojurebot	svn rev 1147; added unchecked-remainder
13:17	clojurebot	svn rev 1147; added unchecked-remainder
13:18	clojurebot	svn rev 1147; added unchecked-remainder
13:19	clojurebot	svn rev 1147; added unchecked-remainder
13:19	rhickey	clojurebot: stop
13:20	clojurebot	excusez-moi
13:20	clojurebot	svn rev 1147; added unchecked-remainder
13:20	abrooks	rhickey: /kick clojurebot
13:20	clojurebot	svn rev 1147; added unchecked-remainder
13:20	gnuvince	wth?
13:21	clojurebot	svn rev 1147; added unchecked-remainder
13:21	duck1123	clojurebot: klaatu barada nikto
13:21	clojurebot	I don't understand.
13:21	clojurebot	svn rev 1147; added unchecked-remainder
13:22	duck1123	clojurebot: latest?
13:22	clojurebot	latest is 1147
13:22	abrooks	rhickey: Which brings up something I've been meaning to mention in the Google group but want to know how you feel about it: Branching off a #clojure-dev for developers and quiet lurkers, having #clojure as a place for general user questions and runaway bots?
13:22	hiredman	oh
13:22	hiredman	rhickey: so sorry
13:22	rhickey	hiredman: no problem
13:23	hiredman	I was messing with it and wonder why it wasn't responding to me in my privmsg window
13:24	rhickey	abrooks: clojure-dev would be for contributors?
13:25	abrooks	rhickey: That's the general idea - contributors and people interested in quietly watching the contribution process. Other opensource projects are set up with a #foo and #foo-dev channel scheme.
13:26	rhickey	abrooks: fine by me - you want to set it up?
13:26	abrooks	rhickey: The tradeoff is that developers (contributors) sometimes withdraw to the #foo-dev channel by habit and won't pay as much attention to #foo. It's desirable that they still be in both but have a high signal-to-noise ratio in #foo-dev when the want to communicate together but still get stuff done.
13:26	abrooks	rhickey: Who set up #clojure?
13:27	gnuvince	jcowan
13:27	abrooks	rhickey: Ah. It would probably be good to make you and tomhickey (and perhaps some others) operators so you can kick/ban runaway bots and users if the need emerges.
13:28	Lau_of_DK	Evening gents
13:28	Lau_of_DK	Gentlemen, wonderful news: Emacs 23 is stable!! :)~~~
13:29	gnuvince	Lau_of_DK: post it to reddit!
13:29	abrooks	Chouser: Do you still have reservations about starting a #clojure-dev at this point?
13:32	danlarkin	abrooks: is there reason enough? Does this channel get too chatty for discussion of clojure development?
13:33	duck1123	is this channel really that busy that it needs to be split in two? I could see that more with the mailing list, but not so much here.
13:35	blackdog	i don't think it's busy enough
13:35	Lau_of_DK	I will say though, that Im having a hard time keeping up with the mailing list
13:39	jvoorhis	hello
13:39	Lau_of_DK	hello jvoorhis
13:40	jvoorhis	i've just started liking clojure, and wondered what the story is with java generics
13:41	jvoorhis	how does one go about instantiating a generic collection, for example http://java.sun.com/j2se/1.5.0/docs/api/java/util/PriorityQueue.html
13:42	drewr	jvoorhis: In Clojure you would do well to use j.u.concurrent. http://java.sun.com/j2se/1.5.0/docs/api/java/util/concurrent/LinkedBlockingQueue.html
13:42	abrooks	danlarkin, duck1123: Perhaps not. I intend to raise this question in the Google group to get wider feedback. I think it's getting chattier all the time. The real question is: Do rhickey and the contributors feel that it's a distraction.
13:43	rhickey	jvoorhis: just leave off the generic part: (java.util.PriorityQueue.)
13:43	jvoorhis	drewr, rhickey: thanks, that's helpful
13:43	rhickey	abrooks: I would appreciate it if people stayed on topic, when they do it's fine
13:44	rhickey	If we create a culture of that, we'll all be better off
13:44	drewr	Anybody watch any football yesterday?
13:44	rhickey	I just don't want to be the 'can we please stop talking about editors/vcs?' police :)
13:46	abrooks	rhickey: So it sounds like you'd want to try to keep the focus on #clojure and keep #clojure focused. More easily said than done but it's a good goal. I suspect that if enough of the regulars give gentle nudges to the others to stay on-topic we can avoid #clojure-dev for a while.
13:46	rhickey	abrooks: sounds like a plan
13:47	danlarkin	a little healthy chatter when the channel is otherwise silent is okay I'd think... helps build community :)
13:48	abrooks	drewr: STAY ON TOPIC, DARN YOU!!! ;-)
13:48	abrooks	Oh.. I said gentle didn't I...
13:48	abrooks	danlarkin: Agreed. It's a tricky blance though.
13:48	Lau_of_DK	abrooks: The only one not staying on topic for the last 10 minutes is.. can you guess? :)
13:49	abrooks	Lau_of_DK: Ack -- you beat me to it. :)
13:49	Lau_of_DK	Anyway, back to Emacs23
13:49	abrooks	Ha!
13:50	Lau_of_DK	HA!
13:50	Lau_of_DK	:)
13:51	AWizzArd	Lau_of_DK: did you find out more about Clojure+JavaFX? Do you think it's possible to write FX apps with exclusively using Clojure as lang?
13:52	Lau_of_DK	AWizzArd: Im not deep enough into it to confirm one way or the other, but 1) I _think_ its possible, 2) its worth pursuing, so I will most likely spend some time on it, once I wrap up another project
13:52	AWizzArd	would be cool :-)
13:53	Lau_of_DK	Yes, would be awesome :)
13:54	gnuvince	Emacs 23 was officially released?
13:54	gnuvince	link?
13:56	whidden	how do you make a "reference" to a java enum/constant from within clojuer, e.g. something to java.util.concurrent.TimeUnit.MILLISECONDS?
13:56	Lau_of_DK	gnuvince: not official, just stable
13:56	Lau_of_DK	cvs -z3 -d:pserver:anonymous@cvs.savannah.gnu.org:/sources/emacs co -r emacs-unicode-2 emacs
13:56	Lau_of_DK	cvs update -A
13:56	gnuvince	Lau_of_DK: what's the difference?
13:57	gnuvince	stable is like RC or something?
13:57	RSchulz	whidden: java.util.concurrent.TimeUnit/MILLISECONDS
13:57	Lau_of_DK	gnuvince: I'll explain later when Rich isnt looking, go check it out :)
13:57	RSchulz	Assuming MILLISECONDS is static.
13:57	whidden	RSchulz... ah forgot the '/' thanks.
13:58	stuarthalloway	whidden: you could also (import-static java.util.concurrent.TimeUnit MILLISECONDS)
13:59	RSchulz	Where is import-static defined?
13:59	stuarthalloway	contrib
13:59	stuarthalloway	it will be in the book (macros chapter) in the next beta
13:59	RSchulz	Sounds like it ought to be in core. Why isn't it?
14:00	stuarthalloway	whoa man, book prose written, don't go "improving" things :-)
14:00	RSchulz	http://www.infoq.com/presentations/click-fast-bytecodes-funny-languages is much more interesting than the slides alone!
14:00	rhickey	stuarthalloway: who me?
14:01	stuarthalloway	rhickey: no, you should improve as much as you like! I was giving RSchulz a hard time
14:01	RSchulz	Stu, if you wanted to write a book that could be "correct" and "complete", you should write abount ancient Rome.
14:02	duck1123	ok, no more bug fixes for stuarthalloway's sake
14:02	stuarthalloway	duck1123: Nooooo!
14:02	stuarthalloway	I have a test suite now, break what you want, I can take it
14:08	whidden	stuarthalloway, right now your chapter on macros is a little thin :) and if Mr. Hickey keeps making late Sunday night addtions to the core you might have a difficult time finishing your book :)
14:08	stuarthalloway	whidden: you will have a macro chapter by tomorrow, mehopes
14:09	RSchulz	stuarthalloway: Are you saying beta4 will be out tomorrow??
14:09	stuarthalloway	I am going to ship the prose to the prags today
14:10	whidden	is there a way to load all of contrib in one fell swoop?
14:10	stuarthalloway	they usually turn it around in hours
14:10	RSchulz	Great. How long does it take them to get it to the gerbils?
14:10	RSchulz	Hours! Then they'll send out notice to us pre-purchasers? Fabulous.
14:10	stuarthalloway	I'd be done already except I let Chouser look at it first. :-)
14:11	RSchulz	Something you do no doubt at your own peril.
14:12	whidden	yeah.. i realize that. but REPL emacs connection is only inferior-lisp and not swank so name completion doesn't work so well :(
14:14	RSchulz	rhickey: Why isn't import-static in core? Is it somehow inappropriate?
14:15	rhickey	RSchulz: I haven't looked at it much
14:15	RSchulz	Does anyone know why import-static presents static methods as macros? It hardly seems intuitive.
14:18	AWizzArd	rhickey: thanks for reacting so fast and providing an unchecked-remainder thumbs-up
14:18	rhickey	AWizzArd: np
14:18	danlarkin	RSchulz: you'll have to ask Stuart Sierra I suppose
14:18	RSchulz	rhickey: Did you give any thought to a TRUNCATE counterpart for Clojure?
14:18	rhickey	RSchulz: I imagine perf - wrapper fns would hide types
14:19	rhickey	RSchulz: yes, I used to have that
14:19	rhickey	got lost during numbers reworking a while back
14:19	RSchulz	Are you considering reinstating it?
14:20	Lau_of_DK	AWizzArd: Did you ever figure out why that function you did was super slow ?
14:20	RSchulz	Lau_of_DK: You mean the one Rich made super fast...?
14:20	Lau_of_DK	Could be
14:21	AWizzArd	Lau_of_DK: I made some changes. Two type hints #^Long and an unchecked-divide and an unchecked-remainder.
14:21	rhickey	RSchulz: sure - patch welcome
14:21	Lau_of_DK	AWizzArd: And I mathced Java in terms of speed?
14:21	RSchulz	I started looking at it, but I'm even more relucant to fiddle with the Java. ... Why, I don't know. At least I know that language...
14:21	RSchulz	Do patches require a CA?
14:22	Lau_of_DK	RSchulz: considering that Rich is American, I imagine even posting to the googlegroup will eventually require signing on some type of legal document :)
14:22	AWizzArd	These four changes made the runtime going from 1360 msecs runtime down to now 124 msecs
14:22	Lau_of_DK	wow
14:22	RSchulz	Are you suggesting Americans are excessively legalistic?
14:23	Lau_of_DK	Have you pasted it somewhere AWizzArd, with the annotations?
14:23	Lau_of_DK	RSchulz: We're a little beyond suggestions :)
14:23	AWizzArd	Lau_of_DK: well, the Java version still runs at 29 msecs, but now the Clojure version is fast enough. It's a good result now.
14:23	rhickey	RSchulz: yes please
14:23	RSchulz	Does anyone know if Java 7 is going to use escape analysis to effect stack-local allocations?
14:24	Lau_of_DK	AWizzArd: Why such overhead? In % its still masive
14:24	RSchulz	Cliff Click's talk (is that really his name?) was pretty interesting, especially the part about how much allocation Clojure demands.
14:25	Lau_of_DK	RSchulz: You got a link to that talk ?
14:25	lisppaste8	AWizzArd annotated #71780 with "Update on find-factor" at http://paste.lisp.org/display/71780#1
14:25	RSchulz	http://www.infoq.com/presentations/click-fast-bytecodes-funny-languages
14:25	RSchulz	danlarkin posted it earlier today. Up 'til then, I'd only seen the slides.
14:25	AWizzArd	Lau, there you can see the changes
14:25	Lau_of_DK	Thank you Mr. Schulz
14:25	Lau_of_DK	Thanks alot AWizzArd
14:25	RSchulz	Eh? My father isn't here.
14:26	duck1123	I hate that you can't download the infoq videos
14:26	RSchulz	But thank _you_ for spelling it correctly.
14:26	RSchulz	Yeah. In that case, the slides are sychronized in a separate Flash animation on the same page, which is kinda' nice.
14:26	duck1123	for some reason my box here at work won't play sound in flash. I have to wait till I get home to watch any videos like that
14:30	RSchulz	You know, even the latest version of find-factors still produces the wrong value for my numeric-pad mash: 43587438757
14:30	RSchulz	user=> (find-factors 43587438757)
14:30	RSchulz	[21517]
14:30	RSchulz	According to Gnu "factor:"
14:30	RSchulz	factor 43587438757
14:30	RSchulz	43587438757: 21517 2025721
14:31	AWizzArd	RSchulz: maybe it is not doing what you expect, but anyway, this was the function that was posted in c.l.l
14:31	RSchulz	Yes, but since it clearly has a bug, that should be addressed, don't you think?
14:31	AWizzArd	I just mimiced that, without wanting to correct it, as I wanted a comparison between those. So by definition, that program is behaving correctly.
14:32	RSchulz	So your contention is that it's just misnamed...
14:32	AWizzArd	one could maybe rename it to: find-some-factors :-)
14:39	abrooks	find-factors-<=-sqrt-of-n
14:43	AWizzArd	rhickey: let's say a company has a codebase in CL. Now we imagine they want to slowly begin to rewrite their app in Clojure. Both sides would have to communicate with each other during the transition phase.
14:43	AWizzArd	How should they do this? Sockets? Trying Kawa or ABCL?
14:44	danlarkin	AWizzArd: seems like that isn't a decision for rich to make... wouldn't it depend on what the app is doing, etc?
14:44	AWizzArd	What approach would be the one about which you would think first? I ask you because you were there, you even developed jfli and so on.
14:44	rhickey	AWizzArd: if it runs under ABCL, they could at least share a VM. Kawa isn't CL, so just another port
14:45	rhickey	depends totally on the granularity of the conversations
14:46	AWizzArd	Would be cool if both sides could use code developed by the other. So, sitting inside the same vm sounds nice.
14:49	dmiles_afk	i am very interested in finding ways of using ABCL and clojure together
14:49	Lau_of_DK	rhickey: Are you putting any focus on the bytecode emission issues that Mr. Click raises in his talk in infoq ?
14:50	rhickey	Lau_of_DK: what issues?
14:50	AWizzArd	And jfli could maybe also be helpful? Parts of the app that were rewritten in Clojure could be called from the Lisp side with it, and the Lisp code could work as it did before.
14:50	Lau_of_DK	rhickey: If you didnt see the talk you should, he explains it very well, but it seems that Clojure bytecode emission has some odd holes in it, compared to going straight to inline
14:50	rhickey	AWizzArd: there was a jFli port to ABCL for Java interop
14:51	rhickey	Lau_of_DK: I saw the talk and spent a lot of time talking to Cliff over 3 days
14:51	Lau_of_DK	rhickey: Ok, did you learn anything interesting?
14:51	rhickey	Lau_of_DK: yes
14:51	Lau_of_DK	hehe
14:51	Lau_of_DK	...that youre willing to share?
14:51	rhickey	seriously, trying to work here
14:52	Lau_of_DK	Ok
14:52	dmiles_afk	what the benefits of clozure and ABCL working together?
14:54	dmiles_afk	well the benefits to cojure it seems to use java w/o the boxing of primitives it looks like.. where ABCL is still boxing everything
14:55	dmiles_afk	i like Kawa because it attempts to do more than just Lisp.. but yet it doenst make it all the way to common lisp the way abcl does
14:56	dmiles_afk	ABCL has enough jclass/jstatic/jcall etc it could use cojure pretty well.. and clojure is smart enough to use user defined types (the ones that are from ABCL)
14:58	dmiles_afk	clojure seems like some routines on Conses since in clojure they are immuatble they probably faster operations?
14:59	AWizzArd	Well, I should have a look at both then, abcl and jfli
15:02	dmiles_afk	yeah i'd start out by making sure their code ran in ABCL first
15:03	dmiles_afk	it probly will, but still you can undo any tangles that come up from abcl if any
15:03	Lau_of_DK	AWizzArd: Can you explain unchecked-remained for me please?
15:03	AWizzArd	Lau_of_DK: it is modulo. 10%3 = 1
15:03	AWizzArd	(rem 10 3) ==> 1
15:03	AWizzArd	(unchecked-remainder 10 3) ==> 1
15:05	Lau_of_DK	How does their implementation differ?
15:06	AWizzArd	the unchecked-remainder simply uses the % from Java
15:08	Lau_of_DK	hmm, odd, I would have expeced rem to do that ?
15:08	AWizzArd	compare it: (let [#^Long x 9999999999999999999999999999999999999999999999999, #^Long y 17] (unchecked-remainder x y))
15:08	AWizzArd	==> -7
15:09	AWizzArd	but try rem instead, and it will give you the correct result
15:09	AWizzArd	no overflow of cpu registers and so on
15:09	AWizzArd	rem does not simply call Javas % operator, because that one is broken and only works for a fraction of possible cases
15:10	AWizzArd	where the number of possible cases increases with the amount of ram you have available
15:10	Lau_of_DK	oh, thats surprising
15:10	AWizzArd	well, a long consumes about 8 bytes of ram, something like that. Only very few numbers can be represented with 64 bit
15:11	AWizzArd	But if you have gigabytes of ram you could encode much bigger numbers, for example a 400 megabyte number. Javas modulo does not always work for those.
15:12	Lau_of_DK	Sloppy coders
15:12	AWizzArd	genau
15:12	Lau_of_DK	?
15:12	AWizzArd	well, they just cared about efficiency
15:13	Lau_of_DK	hehe
15:13	AWizzArd	it's efficient when you do it only for cases for which the cpu gives you direct support
15:14	AWizzArd	although I have to admit that in my programs not often numbers greater than a few billions are needed
15:14	AWizzArd	so for lot's of cases this unchecked-remainder guy comes in handy
15:14	Lau_of_DK	Thats true - Its just not typically seen in public libraries that they knowingly flunk in certain areas.
15:15	Chousuke	flunk?
15:16	Lau_of_DK	flunk, as in produce incorrect results
15:16	Chousuke	oh, well % not working for arbitrarily large numbers is quite normal.
15:17	Chousuke	after all, on very few platforms you have "true" arbitrary precision integers.
15:17	Lau_of_DK	k
15:18	Chousuke	some techniques could be used to automatically decide if native number types can be used, and fall back to bigints when the native precision isn't enough though, but the JVM doesn't offer that as far as I know :/
15:22	Lau_of_DK	Its surprising Chousuke, that in 2008 we're still struggling with lowlevel prolems like that
15:23	Lau_of_DK	I mean, in the 80's people were anticipating flying cars in the new millenium, and here we still have work out when to use ints and bigints :)
15:24	cemerick	when implementing a genclass method that has a boolean return type, it looks like one needs to wrap values with (boolean ...) in order to avoid a classcastexception. It seems that the genclass machinery should automatically convert logical truth values to booleans on their way out of clojure-land.
15:27	Chousuke	Lau_of_DK: computer hardware doesn't get magically smarter. :)
15:27	Chousuke	Lau_of_DK: software tricks can help, but those come at the expense of efficiency.
15:29	Lau_of_DK	Well aint that just a fact :)
15:40	kib2	Hi. Is there any equivalent to CL's atom ?
15:41	Chouser	kib2: (not (coll? x)) is pretty close
15:41	gnuvince	(not (coll? x))
15:41	gnuvince	oops
15:41	gnuvince	too slow
15:41	kib2	thanks Chouser and gnuvince :)
15:53	Lau_of_DK	Hellooo kota :)
15:53	kotarak	Zdravej, Lau.
16:10	RSchulz	Hey, Meikel, I'd watch out for Lau. He has disclosed Denmark's imperial designs... You guys are just about first in line!
16:10	Lau_of_DK	haha
16:11	Lau_of_DK	Our mission was simple - Try to make germany collapse on itself by imposing ugly hair-do's and an unmistakeable Arnold Schwartzenegger accent on the population
16:11	Lau_of_DK	So far its an immense success
16:15	danlarkin	ours was to creep angela merkel out with unsolicited back rubs
16:16	kotarak	oh dear. Maybe I should think about emmigrating...
16:18	kotarak	When she became chancellor a newspaper had a picture of her with the label "Miss Germany". A colleague from Auburn Hills took it with him to shock his colleagues.
16:20	Lau_of_DK	kotarak: Is it true, that since she became the new leader of your Empire, that you start a cold war every month around 28/29/30.th ?
16:20	kotarak	?? Not that I am aware of. She actually does ... well ... nothing.
16:21	Lau_of_DK	k
16:21	danlarkin	lol @ Miss Germany
16:22	kotarak	danlarkin: It's rather depressing....
16:22	Lau_of_DK	/names #clojure-dev
16:22	Lau_of_DK	ERC> abrooks
16:22	Lau_of_DK	Hmm.. maybe we should talk about Clojure :)
16:23	danlarkin	haha yeah rich is itching to tell us to shut up
16:23	kotarak	o.O What's clojure-dev?
16:23	Lau_of_DK	Kota, abrooks gave a 40 minute lecture today on why #clojure should be pure code :)
16:23	Lau_of_DK	It was good too
16:23	Lau_of_DK	(maybe it was 20)
16:24	Lau_of_DK	around 19:00
16:26	Lau_of_DK	kotarak: when youre done, have a look at this http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.11.4048
16:27	kotarak	Lau_of_DK: yes. I know this.
16:28	whidden	is there a sqrt function for bigints/bignumbers in clojure?
16:39	Chouser	whidden: (Math/sqrt x) works, I believe.
16:52	stuarthalloway	how hard would it be to make the :reload option to use/require recursive?
17:00	Chouser	to recurse into dependencies?
17:01	Chouser	stuarthalloway: as easy as saying :reload-all perhaps?
17:02	stuarthalloway	sound of head-slapping
17:02	Chouser	I didn't know it was there either, if that saves your head any pain.
17:03	stuarthalloway	in the best spirit of Clojure, I am now a lazy reader, and was never forced to read past the "," in the doc for require
17:03	rhickey	stuarthalloway: me too, I was going to suggest it should be called something different - maybe :reload-all
17:03	rhickey	:)
17:04	stuarthalloway	more people will care about that flag once people are writing unit tests
17:04	stuarthalloway	and debugging them in the REPL
17:11	Lau_of_DK	Can somebody remind me, why did we want this confusion of (ns foo.bar.baz) and elsewhere (in-ns 'foo.far.baz) ?
17:14	danlarkin	because ns is macro that calls in-ns
17:17	rhickey	ns sets up a namespace declaratively, in-ns just puts you in one
17:18	Lau_of_DK	Sure - But why not just either skip the ' as a whole, or use it in every instance. To other way around seems confusing
17:19	Lau_of_DK	s/To/The
17:31	stuarthalloway	and in similar vein, find-doc could take an unquoted symbol
17:32	Lau_of_DK	true - Like cgrandes javadoc
17:42	Lau_of_DK	Ok, lets say I have a regex that finds ints, and I run it on "a b c 1 2 3", is there a function which will return [1 2 3] to me ?
17:43	RSchulz	Isn't that just an application of (filter ...)?
17:43	AWizzArd	I think only for digits
17:43	RSchulz	Oh, you mean starting from the string?
17:43	Lau_of_DK	I need regex
17:43	AWizzArd	Lau_of_DK: yes, it is possible. No, I don't remember how :-)
17:44	Chouser	re-seq?
17:44	Chouser	sigh
17:44	AWizzArd	I need regexps 2-3 times per year, and every time I basically have to learn from scratch how to use them: http://www.regular-expressions.info/
17:44	Chouser	'reseq'?
17:44	Chouser	're-seq'?
17:44	RSchulz	I thought there was an re-split, but it doesn't turn when I (find-doc "re-")
17:44	AWizzArd	but Lau, one easy thing you could do, although not really efficient: split the string on each position that is not a digit char
17:44	Lau_of_DK	Thats the one Chouser
17:44	Lau_of_DK	Thanks alot! :)
17:45	Chouser	(map #(Integer. %) (re-seq #"\d+" "a b c 1 2 3"))
17:45	Lau_of_DK	AWizzArd: Like I said, I need regex :)
17:45	RSchulz	But (re-seq ...) should work.
17:45	Lau_of_DK	Its in the bag
17:45	AWizzArd	Lau: that's what I mean
17:45	Lau_of_DK	Thats alot guys
17:45	Chouser	RSchulz: re-split is in clojure.contrib.str-utils I think
17:46	RSchulz	Got regex? "Mastering Regular Expressions" is pretty good, and comprehensive. It's in its 3rd ed. now. I have the 2nd (I passed on my copy of the 1st ed. to a friend who always finds REs troublesome.)
17:46	RSchulz	REs are like FP: A mind-set thing.
17:46	Lau_of_DK	I'll look into it Mr. Schulz - I need it
17:47	AWizzArd	Chouser: yes, and I used it to clean a string from unwanted chars
17:47	RSchulz	Chouser: Yup. There it is. In contrib.str-utils, as you said.
17:47	AWizzArd	just split on everything that is not in ^[0-9a-z��] or something like that
17:47	AWizzArd	and then one can (filter #(not (empty? %)) result)
17:48	RSchulz	AWizzArd: For that approach, you'd be much better off inverting the character class you want, rather than trying to enumerate ever non-numeric Unicode character.
17:48	Chouser	#(seq %) is shorter and clearer than #(not (empty? %))
17:49	AWizzArd	yes
17:50	AWizzArd	Chouser: or simply seq instead of #(seq %)?
17:50	Chouser	AWizzArd: oh, of course.
17:50	Chouser	http://clojure-log.n01se.net/date/2008-09-05.html#13:21b
17:51	Lau_of_DK	pictures (re-seq #"(?<=<(img.)>).(?=</img>)" raw-page)]
17:51	Lau_of_DK	Guys, how would you write this, so that it actually works? :)
17:51	stuarthalloway	with an xml parser :-)
17:51	AWizzArd	i c
17:51	Chouser	ding ding ding stuarthalloway wins
17:51	Lau_of_DK	hmm
17:52	Lau_of_DK	Would xml-> work straight on a html ?
17:52	stuarthalloway	on good html :-)
17:52	Lau_of_DK	but I love xml-> :)
17:52	AWizzArd	Lau: html can do things that are not possible with xml or s-expressions
17:52	hiredman	what is xml-> ?
17:53	pjb3	Just saw this on twitter: http://www.turbojugend-cossebaude.de/clojure/clojure-logo.svg
17:53	RSchulz	Lau: What are you trying to do with that?
17:53	RSchulz	(I love an RE challenge.)
17:53	Lau_of_DK	Mr. Schulz - I need to extract links and pictures specifically, in order to crawl a website and make local copies
17:54	RSchulz	pjb3: I see a Java logo when I view that URL.
17:54	pjb3	RSchulz: look carefully, you'll see the lambda
17:54	RSchulz	pjb3: I was wondering about that. It's very similar to the origina, no? Probably a copyright infringement.
17:55	RSchulz	Lau: You want the attribute value, right?
17:55	Lau_of_DK	Yes sir
17:55	pjb3	RSchulz: yeah, probably, it's cool anyway :)
17:55	RSchulz	I've not totally figured out the escaping requirements for the look-ahead and look-behind sub-patterns.
17:56	RSchulz	I think that may be what you're bumping into.
17:56	Lau_of_DK	RSchulz: Theyre actually easy, but Clojure does it a little differently than others
17:56	RSchulz	Could you give me a sample that should be matched?
17:57	RSchulz	What's different about Clojure? It just passes everything on to java.util.Pattern, right?
17:57	RSchulz	(or java.util.regex.Pattern, as the case may be)
17:57	Lau_of_DK	I dunno
17:57	Lau_of_DK	Try this (re-seq #"(?<=<(\w+)>).*(?=<\/\1>)" some-html)
17:57	RSchulz	It's just that with #"..." you don't have to go through the rigamarole of extra escaping.
17:57	Lau_of_DK	You'll get a regex error, but it would work fine in C# or Perl
17:57	RSchulz	I'm looking for the some-html to use to test.
17:58	Lau_of_DK	(defn get-data
17:58	Lau_of_DK	[url]
17:58	Lau_of_DK	(let [s (.openStream (java.net.URL. url))]
17:58	Lau_of_DK	(apply str
17:58	Lau_of_DK	(map char (take-while pos? (repeatedly #(.read s)))))))
17:58	Lau_of_DK
17:58	Lau_of_DK
17:58	Lau_of_DK	(def some-html (get-data "http://www.google.com"))
17:58	RSchulz	Unmatched delimiter.
17:59	RSchulz	The very first close-paren is being seen as ending the zero-width look-behind, yes?
17:59	RSchulz	That leaves the second open paren unmatched.
17:59	RSchulz	That's what I meant by escaping nested patterns in the look-behind notation.
18:00	RSchulz	Also, an unbounded look-behind appears to be a no-no.
18:00	Lau_of_DK	I can read the error also, but the point is, this works in C#
18:01	RSchulz	So what? It has its own RE syntax, no doubt.
18:01	RSchulz	Besides, what are you trying to capture in the look-behind?
18:01	sarcher\|work	how do you run a clojure program so that it exits when it's one?
18:01	sarcher\|work	i'm a newbie
18:01	sarcher\|work	when it's done
18:01	sarcher\|work	sorry
18:01	Lau_of_DK	In this particular, anything between <htmltag> ... </htmltag>
18:02	RSchulz	sarcher\|work: You can just call (System/exit status-value)
18:02	sarcher\|work	thanks
18:02	Lau_of_DK	RSchulz: How do I fix it ?
18:02	sarcher\|work	i knew it would be something easy, i just didn't see an (exit) or (close)
18:02	RSchulz	Lau_of_DK: How generic does it need to be? What can be assumed about what appears inside the element?
18:03	RSchulz	The tag itself is fixed, right? Say, <img ...>...</img> ?
18:03	Lau_of_DK	Specifically I need <img src="**"> and <a href="**">
18:03	RSchulz	OK. And to be clear, you want the value of the src= or href= attributes, not the content of the element, right?
18:03	hiredman	uh
18:04	Lau_of_DK	Exactly right
18:04	Lau_of_DK	(I know the last example was the opposite, but the current task is the most important)
18:04	RSchulz	OK. Give me a minute. In the mean time, consider that the only context you care about is the <img on the left and the very next > on the right.
18:12	RSchulz	Lau_of_DK: OK. Here goes:
18:13	Lau_of_DK	oops
18:13	RSchulz	(def img-src-RE #"<img\b[^>]+src=['\"]([^'\">]+)['\"]")
18:13	RSchulz	(def some-html "this is stuff with an <img foo='bar' src=\"http://foo.bar/image.img\"> in the middle")
18:13	RSchulz	user=> (re-find img-src-RE some-html)
18:13	RSchulz	["<img foo='bar' src=\"http://foo.bar/image.img\"" "http://foo.bar/image.img"]
18:13	duck1123	parsing html with regex is almost always a bad idea
18:14	RSchulz	(I don't know where that < came from. It's a literal < in the RE I pasted.
18:14	duck1123	which is a shame, because it's so handy most of the time
18:14	RSchulz	duck1123: I agree, but snippet extraction of this very limited sort is manageable.
18:14	kib2	Is there any html lib like the PCL one here : http://www.gigamonkeys.com/book/practical-an-html-generation-library-the-interpreter.html ?
18:15	RSchulz	Someone recently posted something like that on the Google groups / list.
18:15	duck1123	there are tons of good java libs you could use to tidy up bad html, and the xml lib to take care of the well-formed ones
18:17	RSchulz	duck1123: You convince Lau to change his approach.
18:17	kib2	RSchulz: any hint about when it has been posted ?
18:17	Lau_of_DK	RSchulz: thats amazing
18:17	RSchulz	I'm just a RE hired gun!
18:17	RSchulz	kib2: Let me check. BRB.
18:17	Lau_of_DK	Im dead tired and its past midnight here, so I hope I can return tomorrow for a walkthrough
18:17	kib2	RSchulz: thanks
18:18	RSchulz	Subject: Clojure template library; Initial posting date: 2008-12-03@10:04
18:18	RSchulz	(that's 10:04 PST)
18:19	Lau_of_DK	Mr. Schulz, have a good night sir, and thank you for your help, youve been instrumental
18:19	kib2	RSchulz: thanks a lot, I'm looking for it
18:20	AWizzArd	Will clojure compile to the same bytecode for (filter seq x) and (filter #(seq %) x)?
18:20	duck1123	I'm sure everyone's seen it, but it seems relevant: http://xkcd.com/208/
18:21	RSchulz	I wonder if I can guess which that one is before I look (I've read them all...)
18:22	RSchulz	Wrong. I was thinking it was the one where God admits he hacked the universe toget in Perl instead of using Lisp...
18:22	RSchulz	duck1123: And I'm flattered at the comparison!
18:23	RSchulz	http://xkcd.com/224/
18:25	AWizzArd	never saw those before
18:27	duck1123	he made a shirt out of the regex one
18:31	dmiles_afk	RSchulz, are you from Teknowlege?
18:31	RSchulz	Howdy, Doug.
18:31	dmiles_afk	Hey :)
18:31	RSchulz	Teknowledge was long in the past, though.
18:32	dmiles_afk	clojure doing you well?
18:33	RSchulz	My brain has yet to warp itself to fit. Even Lisp has been quite a while. But FP really stretches you.
18:33	RSchulz	But I like (Clojure, that is)
18:33	dmiles_afk	i am working on the java port of Cyc (cycorp sanctioned) and seeing what bit of clojure helps
18:34	RSchulz	Interesting. This is based on the openCyc KB?
18:34	dmiles_afk	well its based on ResearchCyc, but the OpenCyc works from the same code when the port source is protected enough to be released as OpenCyc
18:35	dmiles_afk	the OpenCyc/ResearchCyc/FullCyc are the same codebase.. SubLisp translated to C or Java.. then compiled to make the binary
18:35	RSchulz	That could work.
18:36	dmiles_afk	i was researching a SubLisp -> Clojure port .. still waiting on something 2 weeks late from the "vendor" cycorp
18:37	dmiles_afk	its be the same amount of work for anyone deciding if they could indomatically translate Common Lisp to Cojure for the same effect
18:38	dmiles_afk	I keep waiting for anyone in here to undertake something like this :) LISP2CLOJ
18:38	SimonAdameit	Hi
18:39	SimonAdameit	How can I define a local function?
18:39	SimonAdameit	let does not allow recursion
18:40	dmiles_afk	RSchulz, i still have to say your implmentation of a indexed KB using java int[]s still beats everything i've seen
18:40	AWizzArd	SimonAdameit: with fn or loop
18:40	AWizzArd	anonymous functions can call themselves in Clojure
18:40	RSchulz	Did I do that?
18:40	SimonAdameit	how?
18:41	AWizzArd	with (recur ...)
18:41	AWizzArd	so, no Y-Combinator needed ;-)
18:41	SimonAdameit	cool, thanks :)
18:41	AWizzArd	Just think about recur as if it were the name of the function.
18:41	dmiles_afk	RSchulz, well the Teknowleedge Sigma code arround SKBs .. i figured was your authorship
18:42	dmiles_afk	RSchulz, up until three months ago someone exposed the Tek Web CVS for annonymous browsing for 6 years of public unfetered access
18:42	SimonAdameit	AWizzArd: "Can only recur from tail position"
18:45	SimonAdameit	Ok, got it, fn takes a "name" argument. Though it is somewhat unexpected at first doing it this way (when you're comming from lisp)
18:53	AWizzArd	SimonAdameit: you don't need the name argument
18:53	AWizzArd	and yes, you should use recur only in a tail position, so that it can optimize the overhead away
18:54	SimonAdameit	How do I do non tail recursion without the name?
18:54	AWizzArd	well, for that you need a name
18:55	AWizzArd	I just don't see why you would want to do it ;-)
18:56	SimonAdameit	Ahh, come on, I'll paste to have some fun :)
19:00	lisppaste8	SimonAdameit pasted "cfg parser" at http://paste.lisp.org/display/71828
19:02	hiredman	SimonAdameit: (fn somename [args] body)
19:02	hiredman	you can use (somename x) in the body
19:02	hiredman	I normally use "this"
19:03	SimonAdameit	hiredman: thanks, I already discovered that one :)
19:17	Chouser	hiredman: earlier versions of clojure bound 'thisfn' for you.
19:29	hiredman	I remember
19:35	powr-toc	Is there a function in clojure that returns a seq where only the first item to match a predicate is removed?
19:40	blackdog	powr-toc, (doc some)
21:39	danlarkin	can someone give an example of fnseq? Not sure why (fnseq 4 #(vector 45)) throws a ClassCastException
21:41	Chouser	a vector is not an ISeq
21:42	Chouser	(fnseq 4 #(list 45))
21:45	danlarkin	ah perhaps the documentation should be updated to reflect the need for an ISeq?
21:49	Chouser	The exception described the need pretty emphatically. :-)
21:55	jvoorhis	i noticed some similarities between the doc strings in src and the website -�is there a clojure documentation tool?
22:02	joha1	I'm testing compojure and want to add a function from the REPL, but I get an unresolved error when calling defservlet. Probably just a namespace problem, but I'm new to clojure and can't figure out what to do. Help?
22:02	Chouser	you have a .clj file that's working?
22:03	joha1	I'm just testing from the repl script supplied in compojure
22:03	Chouser	jvoorhis: there's 'doc' and 'find-doc' -- I don't know what's used to convert that to a web page.
22:04	jvoorhis	thanks
22:04	Chouser	(doc doc)
22:04	clojurebot	Prints documentation for a var or special form given its name; arglists ([name])
22:04	jvoorhis	begs the question, is the website code checked into the main repo?
22:05	Chouser	jvoorhis: it's a wiki of some sort
22:05	rhickey	wikispaces
22:05	jvoorhis	cool, thanks
22:07	jvoorhis	looking forward to writing some more clojure code :) so far, i've been able to send midi messages to ableton live from the repl
22:07	rhickey	jvoorhis: neat
22:07	jvoorhis	with little effort: http://gist.github.com/32973
22:09	jvoorhis	it's cool to see others are doing multimedia stuff with clojure, too
22:09	jvoorhis	someday, i expect to see a processing demo
22:10	Chouser	jvoorhis: like this? http://github.com/rosado/clj-processing/tree/master
22:11	jvoorhis	Chouser: that's it
22:15	jvoorhis	Chouser: the example code is nice
22:16	Chouser	Great. I've not done anything with it -- just saw it go by on the group.
22:17	jvoorhis	oh, i see
22:17	Chouser	joha1: sorry, I'm apparently just as lost as you are.
22:18	joha1	np
22:18	jvoorhis	i should probably subscribe, and poke around in clojure-contrib while i'm at it
22:21	Chouser	joha1: shot in the dark: (use 'compojure.http)
22:22	joha1	cool, it worked. I'm a real n00b at this :) thanks
22:25	Chouser	np
23:02	danlarkin	http://nothings.org/computer/iterate.html -- it's funny, the title is "Iteration is hard" but it should really be "State is hard"
23:19	Chouser	that's a great link for when anybody complains about how "hard" it is to "just set a variable" in clojure.
23:19	Chouser	You want mutable state? Really?
23:24	danlarkin	sadly it's too verbose for the type of person complaining about that :-/
23:26	Chouser	hm.
23:29	danlarkin	hopefully that didn't come off as insulting
23:38	Chouser	no, you probably have a point. And stated sufficiently broadly that we can each justify our own exclusion from that group.
23:42	RSchulz	Along the lines of of the "talking Barbie" shtick, remember the line uttered by the talking Malibu Stacie doll: "Thinking give you wrinkles!"