#clojure – 2009-07-24

2009-07-24

00:58	mebaran151	what's the best way to convert an integer to a byte array in bigendian notation?
01:42	replaca_	anyone here no something about ring and jetty?
01:43	replaca_	I'm trying to understand how to parse posted form data
01:43	replaca_	without writing my own parser for it!
01:44	arbscht	I think they are held in :form-params in the request map. is that what you mean?
01:45	replaca_	I think that's what I mean, but they don't seem to be there
01:45	replaca_	(map doesn't have that key)
01:46	replaca_	the post is x-www-form-urlencoded fwiw
01:49	replaca_	hmm, that key doesn't exist anywhere in ring. Is it a compojure thing?
01:52	replaca_	yes, I see that it is now
02:00	hiredman	I think ring doesn't parse form stuff for you, it operates below that "level"
02:01	PatrickC	Hello.
02:08	PatrickC	Hello
02:42	Lau_of_DK	Top of the morning gents
03:10	Fossi	hi
04:00	AWizzArd	~seen kotarak
04:00	clojurebot	kotarak was last seen quiting IRC, 587 minutes ago
04:19	Lau_of_DK	~get kotarak
04:19	clojurebot	I don't understand.
04:25	AWizzArd	Maybe you can write a patch Lau, so that the bot can run that function?
04:40	Lau_of_DK	Np - It does require some cooperation on the part of the one who is being fetched though - I'm thinking phone app or something similar
04:52	Fossi	"emacs" question: is there an easy way to find out what file a method is in?
04:53	Lau_of_DK	Well... :) M-! grep -nr "foo" *, would do it
04:54	Lau_of_DK	Are you on Linux?
04:54	Lau_of_DK	@ Fossi
04:54	Fossi	well, that isn't really so helpful in jars
04:54	Lau_of_DK	Oh, you mean by way of reflection ?
04:54	Fossi	it would show up, but i'd like the namespace or such
04:55	Fossi	(source foo) is great, but unfortunately doesn't give the file either
04:56	Lau_of_DK	I'm blank, never had that question before
04:56	Lau_of_DK	Why would you want the filename?
04:57	Fossi	well, more the namespace, but it's kinda the same
04:57	Fossi	say, i (use 'compojure), but really only call start, so i would like to find out, where that is defined
04:58	Fossi	and a hundred other small use-cases like that
05:00	Fossi	it's also a bit annoying there's no automatic import cleanup :)
05:01	Chousuke	,(:file ^#'vec); It should be in the metadata
05:01	clojurebot	"clojure/core.clj"
05:01	Fossi	ah, nifty
05:01	Lau_of_DK	very
05:01	Chousuke	:line, too :P
05:01	Chousuke	,^#'vec
05:01	clojurebot	{:ns #<Namespace clojure.core>, :name vec, :file "clojure/core.clj", :line 241, :arglists ([coll]), :doc "Creates a new vector containing the contents of coll."}
05:02	Lau_of_DK	Awesome tip
05:02	Lau_of_DK	Chousuke, I hereby grant you the title of Tipster !
05:03	Fossi	that make a nice addition to repl-utils imho
05:03	Chousuke	:P
05:03	Fossi	'd
05:03	Chousuke	what would? that's just metadata :P
05:12	cschreiner	...
05:13	Lau_of_DK	;;;
06:41	rottcodd	how do I checkout the par branch?
06:42	Chousuke	git checkout -b par origin/par
06:42	Chousuke	that sets up a local par branch to track the par branch on origin
06:42	Chousuke	ie. rhickey's github repo.
06:43	rottcodd	thanks
06:44	Chousuke	if you don't need a tracking branch, git checkout origin/par works too, but you won't be able to commit in it :)
08:35	djpowell`	which jsr166y.jar do I need for the par branch?
08:35	rhickey	the one in the downloads section on github
08:36	djpowell`	ah, ok
08:36	rhickey	http://cloud.github.com/downloads/richhickey/clojure/jsr166y.jar
08:39	djpowell`	what are the current experimental branches? par, chunks, and ensure?
08:40	rhickey	yes, and par has everything chunks does
08:40	djpowell	nice. do they both have mutable collections?
08:41	rhickey	yes, exactly
08:41	rhickey	mutable vectors, and that is all the chunks branch has over master at this point, the actual chunks work is in master now
08:42	djpowell	is the main effect of the chunks work to speed up things like map over vectors?
08:43	rhickey	djpowell: yes, and any other collections that can support block access, like arrays
08:44	rhickey	not yet implemented for maps/sets, but possibly useful there too
08:45	rhickey	right now only a few ops (map/reduce/filter) have chunk support, but I'm hoping for some contributions (for, doseq etc)
08:46	rhickey	while not every seq op can be parallelized, most can be chunked
09:01	Chouser	Well, I'm still doing battle with JNI, but now finally to an issue that's Clojure-related. That is, loading me .so and calling a java method works from .java code, but not from .clj code.
09:02	Chouser	I'm wondering if it's a classloader issue of some sort
09:03	rhickey	Chouser: almost certainly
09:03	Chouser	like perhaps the classloader used to call System/loadLibrary has to have a particular relationship with the classloader used to call the method
09:03	rhickey	shouldn't it?
09:04	Chouser	I tried (do (System/loadLibrary "my_lib") (my.lib/method))
09:04	Chouser	at the repl, that should be done all in one classloader, right?
09:05	Chouser	anyway, that fails: java.lang.UnsatisfiedLinkError: my.libJNI.method()I
09:06	Chouser	Should my goal be to get the root classloader to do the loadLibrary()? Is that possible?
09:07	rhickey	Chouser: I'm trying to understand the semantics of loadLibrary
09:09	rhickey	Have you tried this advice:
09:09	clojurebot	this is not a bug
09:09	rhickey	If native methods are to be used in the implementation of a class, a standard strategy is to put the native code in a library file (call it LibFile) and then to put a static initializer:
09:09	rhickey	static { System.loadLibrary("LibFile"); }
09:09	rhickey
09:09	rhickey	within the class declaration. When the class is loaded and initialized, the necessary native code implementation for the native methods will then be loaded as well.
09:09	rhickey	from: http://java.sun.com/j2se/1.5.0/docs/api/java/lang/Runtime.html#loadLibrary(java.lang.String)
09:09	Chouser	I was just about to paste that.
09:09	Chouser	so do that in a .java and then try the calls from .clj. I'm on it...
09:09	rhickey	manual consumer calls to loadLibrary seem fraught
09:10	rhickey	the native library should be an implementation detail of some class, and loading the classes should work as always
09:11	rhickey	er, work as well as always, which is not ... always
09:11	Chouser	heh
09:12	rhickey	have you set java.library.path?
09:12	Chouser	woo!
09:12	Chouser	static { } did it
09:13	sgtarr_	Clojure is awesome, I just need to play around with it for months before I would dare to use it in a commercial / enterprise setting (which is my job).
09:13	cemerick	ut oh, do we need a static init form in clojure now? :-/
09:14	Chouser	cemerick: I don't think so.
09:14	sgtarr_	rhickey: And you are turning into the new rockstar in Computer Science, no doubt. :-)
09:14	Chouser	I've already got a pile of .java generated by swig for this JNI stuff. I just need to figure out how to inject a static init into the code it's producing.
09:15	Chouser	adding it manually to the .java is what just worked. The .java file is already there and already requires compilation, so not much value in moving it to .clj
09:15	rhickey	sgtarr_: not really, I'm just a toolmaker
09:16	sgtarr_	That's what James Gosling says too.
09:47	Chouser	I assume there's no place to put .java code that will run when any class in a package is used, right?
09:48	Chouser	swig is missing a feature I need. Not really java's fault for failing to make up for it...
09:52	cemerick	Chouser: ServiceProviders will do that, although it's JDK6+ IIRC
09:55	Chouser	I'm finding javax.imageio stuff, but I assume that's not what you're talking about?
09:57	cemerick	that was the original impl of the pattern in JDK5, I think
09:58	cemerick	hrm, I think the "run when a class in a package is accessed" functionality was tied to the imageio stuff, though.
09:58	cemerick	what I was thinking about that is generally-useful is: http://java.sun.com/javase/6/docs/api/java/util/ServiceLoader.html
09:58	cemerick	No package-level automatic stuff, tho, as far as I can tell.
10:15	rzoom	as a watcher on assembla, is there any way to not get emailed so often?
10:16	rzoom	I don't see an "Alerts" tab on the top of the space
10:16	djpowell	rhickey: how many cores did you have on the machine you did the 'pvmap rocks'?
10:19	Chouser	rzoom: Stream tab, upper-right corner, "change alert settings"
10:31	rzoom	Chouser: thanks! the assembla FAQ says there is an "Alerts" tab.
10:44	rhickey	djpowell: 4 cores
11:05	stuartsierra	I can't reproduce this out-of-memory error: http://groups.google.com/group/clojure/browse_thread/thread/c0e889d2a6c200ba
11:07	hiredman	maybe he has a small heap
11:07	stuartsierra	I tried it with -Xmx64m and a 1.5 GB file. No problem.
11:08	hiredman	well
11:08	hiredman	1.5gb is hardly "really big"
11:10	hiredman	and who knows what his line processing function does
11:11	stuartsierra	he said it failed on just (write-lines "out" (line-seq (reader "bigfile")))
11:12	cemerick	rhickey: I forget: have you previously warned me off the idea of having an .isForced() method on Delay?
11:14	hiredman	failed because I don't have 36 gigs free to copy to
11:14	hiredman	wait, yes I do
11:14	hiredman	java.lang.RuntimeException: java.io.IOException: Input/output error (NO_SOURCE_FILE:0)
11:14	hiredman	hmm
11:18	hiredman	java.lang.OutOfMemoryError: Java heap space (NO_SOURCE_FILE:0)
11:18	hiredman	there we go
11:19	hiredman	after writting 1g of the 36g file
11:20	stuartsierra	Now why don't I get that?
11:24	stuartsierra	This also works: (let [lines (line-seq (reader "/tmp/bigfile"))] (write-lines "/tmp/out" lines))
11:24	rzoom	very long lines?
11:24	stuartsierra	hmm, possible. That would have to be a single line larger than the Java head, though.
11:26	stuartsierra	s/head/heap/
11:31	stuartsierra	duck-streams/write-lines isn't truly lazy; maybe that's the problem
11:32	mebaran151	is there anyway to hook in test-is with a junit runner so my ide could run them: I'm using the Netbeans plugins and I just want an easy way to run my tests
11:33	stuartsierra	mebaran151: Not yet, want to write one? :)
11:33	mebaran151	eh I could take a shot at it
11:33	stuartsierra	There's a patch for JUnit-style XML output, I need to get that applied.
11:34	mebaran151	don't need the xml output: I'd just like to integrate it with my ant task
11:35	stuartsierra	Working with JUnit from Clojure is difficult because it relies so heavily on Java annotations.
11:35	stuartsierra	That's why I didn't try.
11:35	clojurebot	http://clojure.org/rationale
11:35	mebaran151	I always thought annotations were cludgy
11:35	mebaran151	it seems like a poor man's solution to make things slightly more dynamic
11:36	mebaran151	I know Scala Specs does it though, and I don't think Scala support annotations yet
11:36	cemerick	mebaran151: you don't really need junit integration to make test-is tests work nicely with ant
11:36	stuartsierra	cemerick: That's true.
11:37	mebaran151	ah, how would I set this up
11:37	cemerick	lisppaste8: url?
11:37	lisppaste8	To use the lisppaste bot, visit http://paste.lisp.org/new/clojure and enter your paste.
11:37	lisppaste8	cemerick pasted "test-is ant harness" at http://paste.lisp.org/display/84139
11:37	cemerick	^^ that's what we use. Crude, but it works.
11:38	stuartsierra	Here's how I do it: http://github.com/stuartsierra/altlaw-backend/blob/5b9edd606df944e5c58e5ad32191b27436201950/build.xml#L95
11:38	cemerick	throws an exception on test-is failure, so ant's failonerror="true" will halt the build, etc.
11:38	mebaran151	also how would I organize my tests in /test?
11:39	cemerick	mebaran151: however you want :-)
11:39	mebaran151	no specific sort of dir format mandated from on high?
11:39	cemerick	I set it up so that all clj files in src/test are munged into namespaces, and passed as args to the class/main method I pasted above
11:40	mebaran151	I see
11:40	cemerick	if a particular test ns doesn't define any tests, then the test-ns call is a no-op, so there's no harm in casting a wide net
11:40	cemerick	stuartsierra: so do you have to manually maintain a list namespaces to load in org.altlaw.load-all?
11:41	Chouser	ok! swig indeed has the features I needed.
11:41	Chouser	now every java class it generates has "static { module.init() }"
11:42	mebaran151	I see
11:42	Chouser	and the java class named "module" has "static { System.loadLibrary("my_lib"); }"
11:42	rhickey	Chouser: makes sense
11:43	cemerick	mebaran151: irc being what it is, I can only surmise that you're uneasy ;-)
11:43	Chouser	so an attempt by java/clojure to use any of the generated java classes causes the moudule's init method to be called which itself does nothing but causes the loadLibrary to happen once.
11:43	Chouser	this seems blogworthy. swig makes this much more pleasant than it otherwise would be.
11:43	mebaran151	a little uneasy
11:44	mebaran151	right now I've just been manually running my tests from the repl
11:44	mebaran151	but before I deploy, I love having a test suite I can run just to make sure everything made to the server okay
11:44	cemerick	yeah, definitely
11:45	cemerick	FWIW, the mechanism we use is roughly what ant does with junit tests now -- it crawls around your test classpath, looking for TestCases and Suites to run shrug
11:45	cemerick	There's definitely smarter ways to do it, but hardly any easier ones :-)
11:46	stuartsierra	cemerick: not with the addition of clojure.contrib.find-namespaces
11:46	stuartsierra	cemerick: that was in response to the load-all question
11:46	cemerick	oh, that's interesting. Never knew it was there. We've got an ant contraption that converts clj filenames -> namespaces
11:47	cemerick	we don't do any loading tricks, so we have a 1:1 correspondence
11:47	stuartsierra	that's basically what find-namespaces does, but it also parses the (ns...) declaration
11:47	cemerick	well, clj code is way better than an ant contraption :-P
11:47	stuartsierra	I originally had org.altlaw.load-all when I was doing a lot of gen-class stuff, I can probably eliminate in now.
11:48	cemerick	the only advantage using ant has is that there's an easy mechanism to prevent recompilation of files that haven't changed since the last build.
11:50	stuartsierra	Yes, that's harder. I just recompile all the .clj files when any of them have changed.
11:51	cemerick	once our incremental build process crossed the 1-minute mark (because of clj recompilation), I implemented that. Now it's 12s :-)
11:52	stuartsierra	nice
11:57	mebaran151	I really wish the json.read wasn't so strict: uneval in mozilla actually creates non-spec json
12:00	hiredman	mebaran151: you might checkout clojure-json
12:01	mebaran151	how to spec is it? I was reading the source code and it looked equally inflexible (I think it demanded quotes around string keys too)
12:02	hiredman	it could be even less flexible, dunno
12:05	hiredman	~def write-lines
12:07	mebaran151	also is there a quick clojure way to turn an Integer into its big endian byte encoding and back?
12:10	hiredman	,(Integer/toBinaryString 3)
12:10	clojurebot	"11"
12:10	angerman	does anyone have experineces with clojure on terracotta?
12:11	drewr	someone was making progress with that but I haven't seen any updates
12:11	angerman	I found a blog that supposidly said it had it running
12:11	angerman	and someone on the ML said he deployed something like that in a mecial hospital
12:12	replaca_	angerman: I would go to the clojure google group and do a search. I remember there was some discussion of terracotta support a few months back
12:12	angerman	replaca_: yep, as I said :) ... I guess i sould switch vom MailingList to Group :)
12:13	replaca_	Wow! github has been super-slow the last couple of days
12:13	replaca_	angerman: someone was working on generalized support and rich was helping him IIRC
12:14	replaca_	angerman: but I haven't heard anything about it for a while
12:14	replaca_	angerman: if you're looking for distributed execution you might look at swarmiji too
12:14	stuartsierra	Ok, I put the JUnit XML output in a patch on Assembla. I think I did it right.
12:15	angerman	replaca_: I'm basically looking for a solution to kinda split all the parralliyable stuff across the systems we have here
12:15	cemerick	angerman, replaca_: pjstadig (Paul Stadig) did the initial work, with Rich's help: http://paul.stadig.name/2009/03/clojure-terracotta-update.html
12:16	angerman	I think there are like 12 nearly idle opertron systems + a few Mainframes for the sun ray system that's employed on the campus
12:16	rhickey	angerman: also see: http://github.com/amitrathore/swarmiji/tree/master
12:16	hiredman	there is always hadoop
12:16	angerman	cemerick: ok, i found that blog too. so that's the defacto piece to use :)
12:17	stuartsierra	hadoop isn't aimed at compute-intensive apps, although it may still work
12:18	cemerick	ooh, swarmiji looks interesting
12:18	angerman	hmm swarmiji looks nice
12:18	cemerick	hadoop is so friggin' complicated...
12:19	angerman	cemerick: well I won't bother with the datastore part. I got all the data in a postgresql db already, which did a lot of the prefiltering of the data using pl/R
12:19	matting	Hi! I came accross a problem I can't solve: I need to use some java gui and therefore i need to reimplement the paint() method. So my guess is to use 'proxy'. But i need more information (here: what stuff to draw) in paint() that is supplied by the arguments.
12:19	angerman	matting: ?
12:20	replaca_	the advantage of swarmiji is that it seems pretty simple, but I haven't had a chance to actually use it yet
12:20	matting	i'd like to have a function that draws a path
12:20	stuartsierra	matting: proxies can be closures
12:20	rhickey	matting: (let [other stuff] (proxy ... other))
12:20	hiredman	matting: the "methods" in proxys are fns so they can close over stuff
12:21	stuartsierra	Ha! 3 different forms of the same answer.
12:21	matting	:)
12:22	matting	but i still don't get it. suppose i have two paths path1 and path2. i'd like to do sth like (draw-path path1) or (draw-path path2)
12:22	angerman	though before i tired that I did use a jpanel in a jframe
12:22	replaca_	hiredman: I have a new json file for clojurebot to chew on that has the pointers to clojure.contrib on github
12:22	matting	does this mean i need to create a new proxy for each call?
12:22	replaca_	hiredman: it's at http://github.com/richhickey/clojure-contrib/raw/gh-pages/api-index.json
12:23	replaca_	hiredman: I think it should be identical to the old old, format-wise
12:23	replaca_	*old old => old one
12:23	hiredman	ok
12:27	replaca_	hiredman: let me know if you have any issues
12:27	djpowell	paint is a callback that is called by java - you put the instructions on how to paint the screen in it, and java will call it when it needs to repaint the screen - which you can also trigger by telling it to repaint. so you'd have a single proxy for your window or panel or whatever that overrides paint.
12:27	hiredman	replaca_: it works
12:28	hiredman	(I have no issues)
12:29	djpowell	as to what to put in paint - you might structure your application so that another thread updates a ref with the current state of the world, in a loop, that also calls repaint on the component, and sleeps for a bit, and then your paint method renders that state. maybe.
12:30	cemerick	rhickey: any idea to what degree (if any) cinc will impact the usability of clojure data structures/APIs from the Java side? e.g. will PHM, etc., still have stable names?
12:30	stuartsierra	mebaran151: Re "uneval in mozilla actually creates non-spec json": c.c.json.read is a JSON parser, not a Javascript parser. It won't parse Javascript objects. That's a design choice; parsing real Javascript is quite a bit harder. You might as well use a real Javascript interpreter like Rhino.
12:31	mebaran151	heh, I actually am using rhino, but sometimes I want to get the values it generates back into clojure!
12:32	replaca_	hiredman: cool!
12:32	replaca_	that was easy
12:32	Chouser	cemerick: interesting question. The intefaces and abscract classes would have stable names of course.
12:32	mebaran151	I kind of wish Mozilla had included to_json function in their standard distribution
12:32	matting	but how can i make my reimplementation of paint() dependent on the state?
12:32	stuartsierra	mebaran151: ah, should be fairly straightforward to transform Rhino's objects into Clojure types, then.
12:32	stuartsierra	I do that with JRuby.
12:32	matting	i think that's what i don't understand yet
12:33	Chouser	cemerick: but if the concrete classes are defined using newnew (which I would assume) than their actual classnames would be generated I would think.
12:33	cemerick	Chouser: I'm thinking of the impact on RMI, which uses big-S Serialization, which depends on stable names
12:33	mebaran151	the included scripting JDK isn't very easy to work with
12:33	stuartsierra	ah
12:33	hiredman	replaca_: how difficult would it be to add clojure.core to that?
12:33	clojurebot	clojure is the brand
12:33	mebaran151	I did end up swapping it out for Rhino proper
12:33	hiredman	~botsnack
12:33	clojurebot	thanks; that was delicious. (nom nom nom)
12:34	mebaran151	but I wouldn't have had to if Mozilla hadn't thought in their infinite wisdom to produce adhoc JSON instead of real JSON
12:34	stuartsierra	yeah
12:34	Chouser	cemerick: hm... that brings up print-dup as well. I guess we'll either need stable constructor names (still seems unlikely to me) or print-dup will have to use the factory functions (hash-map, sorted-set-by, etc.)
12:35	cemerick	the latter seems most reasonable
12:35	replaca_	Q: does anyone know where there's a regexp pattern or something for finding links in raw text? I'm trying to take docstrings and convert the embbeded things like http://json.org to <a> tags when I render them in HTML. This seems like a very solved problem, but I can't come up with a google search for it :-)
12:35	cemerick	There's always writeObject and readObject overrides to get proper Java Serialization (presumably using print-dup)
12:36	matting	you see, if i would reimplement paint() in java paint() would also depend on the internal state of the object. i don't know how you can get the same thing done with clojure.
12:36	replaca_	hiredman: shouldn't be hard at all. Would it be OK if it were in a separate file?
12:36	cemerick	But that puts serialized clojure objects out of the reach of the various tools that expect "normal" serializations.
12:36	hiredman	replaca_: that would be fine
12:36	stuartsierra	matting: your proxy function needs to close over a variable representing the state. If it's mutable, then it should be one of Clojure's mutable types, probably a Ref.
12:36	replaca_	cool, i'll put it on the todo list
12:37	rottcodd	with 2 packages p1 and p2, is it possible to have p2 export p1/foo as p2/foo?
12:37	hiredman	replaca_: clojurebot uses #"[A-Za-z]+://[^ ^/]+\.[^ ^/]+[^ ]+" to snag urls out of irc
12:37	djpowell	or, you could just make the variable a global ref I guess
12:38	replaca_	hiredman: works pretty well?
12:39	hiredman	seems to
12:39	replaca_	hiredman: I worry about links at the end of a sentence
12:39	hiredman	what do you mean?
12:39	matting	stuartsierra: can you point me to some clojure source where i can see how this is done?
12:40	replaca_	"For more information, see http://json.org/overview.html."
12:40	replaca_	want to get the middle period, but not trailing
12:40	hiredman	,(re-find #"[A-Za-z]+://[^ ^/]+\.[^ ^/]+[^ ]+" "For more information, see http://json.org/overview.html.")
12:40	clojurebot	"http://json.org/overview.html."
12:40	hiredman	hmmm
12:41	replaca_	that tends to be pretty common in docs
12:41	djpowell	matting: in http://clojure.googlegroups.com/web/ants.clj the state is just held in global vars that hold refs (eg world), they are manipulated by lots of threads, and the paint method calls the render function, which shows how to call methods on the Graphics object in order to draw stuff
12:41	opqdonut	,(re-find #"[A-Za-z]+://[^ ^/]+\.[^ ^/]+[^ ]+" "For more information, see http://json.org/overview.html.")
12:41	clojurebot	"http://json.org/overview.html."
12:41	opqdonut	gah, sorry
12:41	opqdonut	,(re-find #"[A-Za-z]+://[^ ^/]+\.[^ ^/]+[^. ]+" "For more information, see http://json.org/overview.html.")
12:41	clojurebot	"http://json.org/overview"
12:42	Chouser	irc-log-to-html.clj uses #"(?:https?://\|www\.)(?:<[^>]>\|[^<>\s])(?=(?:>\|<\|[.()\[\]])*(?:\s\|$))"
12:42	clojurebot	see logs
12:42	stuartsierra	matting: not a great example, but here's one: http://tinyurl.com/lwqbhq
12:42	matting	thanks
12:42	hiredman	,(re-find #"(?:https?://\|www\.)(?:<[^>]>\|[^<>\s])(?=(?:>\|<\|[.()\[\]])*(?:\s\|$))" "For more information, see http://json.org/overview.html.")
12:42	clojurebot	"http://json.org/overview.html."
12:43	hiredman	:/
12:43	replaca_	Chouser: now there's a regex!
12:43	hiredman	,(re-find #"[A-Za-z]+://[^ ^/]+\.[^ ^/]+[^ ]+" (.replaceAll "For more information, see http://json.org/overview.html." "\.$" ""))
12:43	clojurebot	Unsupported escape character: \.
12:43	djpowell	stuartsierra: re your jmx server - do you think it is worth using OpenMBeans instead of straight DynamicBeans?
12:43	hiredman	,(re-find #"[A-Za-z]+://[^ ^/]+\.[^ ^/]+[^ ]+" (.replaceAll "For more information, see http://json.org/overview.html." "\\.$" ""))
12:43	clojurebot	"http://json.org/overview.html"
12:44	stuartsierra	djpowell: not mine
12:44	djpowell	oh - oops
12:44	cemerick	replaca_: java regexes also support forward non-capturing assertions, which are generally cleaner than the alternatives, IMO
12:44	stuartsierra	djpowell: Stuart Halloway, I think
12:44	djpowell	ah
12:44	replaca_	cemerick: yeah, that might be part of the sol'n
12:44	Chouser	oh, you know what -- my regex is meant to operate on html. that's why it's so crazy
12:45	replaca_	hiredman: $ loses because this doesn't necessarily happen at the end of the source text
12:45	Chouser	that is, it's skipping over tags and entities embedded in the url.
12:45	replaca_	Chouser: yeah, makes sense
12:46	replaca_	when I was doing the autodoc on the google code wiki, it did this for me (its one nice feature), but in the new autodoc, I've got to do it myself
12:46	hiredman	,(re-find #"[A-Za-z]+://[^ ^/]+\.[^ ^/]+[^ ]+" (.replaceAll "For more information, see http://json.org/overview.html. foo bar" "(\\.$\|\\.\\s)" " "))
12:46	clojurebot	"http://json.org/overview.html"
12:47	replaca_	hiredman: which finds it at some cost to my source text
12:47	replaca_	:)
12:48	replaca_	the idea is to be able to transform "For more information, see http://json.org/overview.html. foo bar"
12:48	Chouser	,(map second (re-seq #"(\w+://.?)[.>](?: \|$)" "For more information, see http://json.org/overview.html. foo bar"))
12:48	clojurebot	("http://json.org/overview.html")
12:49	replaca_	into "For more information, see <a href="http://json.org/overview.html">http://json.org/overview.html</a>. foo bar"
12:50	replaca_	Chouser: nice!
12:50	replaca_	thanks guys, I think I have a good starting point now.
12:53	Chouser	(str2/replace "For more information, see http://json.org/overview.html. foo bar" #"(\w+://.?)([.>](?: \|$))" (fn [[_ url etc]] (str "<a href='" url "'>" url "</a>" etc)))
12:53	Chouser	,(require '[clojure.contrib.str-utils2 :as str2])
12:53	clojurebot	java.io.FileNotFoundException: Could not locate clojure/contrib/str_utils2__init.class or clojure/contrib/str_utils2.clj on classpath:
12:55	replaca_	Chouser: beautiful!
12:55	replaca_	why think when you have #clojure?! :-)
12:58	beutdeuce	how does cons work?
13:01	hiredman	(fn [a b] (fn [c] (c a b)))
13:02	stuartsierra	beutdeuce: (cons x s) adds the object x to the front of the sequence s
13:02	beutdeuce	ah, like the Haskell : operator?
13:03	stuartsierra	more or less, yes
13:03	stuartsierra	,(cons 1 (cons 2 (cons 3 nil)))
13:03	clojurebot	(1 2 3)
13:03	beutdeuce	thnx
13:04	hiredman	,(let [_cons (fn [a b] (fn [c] (c a b))) car (fn [a b] a) cdr (fn [a b] b)] (car (_cons 1 nil)))
13:04	clojurebot	java.lang.IllegalArgumentException: Wrong number of args passed to: sandbox$eval--3535$car
13:05	hiredman	bar
13:05	hiredman	,(let [_cons (fn [a b] (fn [c] (c a b))) car (fn [c] (c (fn [a b] a)) cdr (fn [c] (c (fn [a b] b))] (car (_cons 1 nil)))
13:05	clojurebot	Unmatched delimiter: ]
13:05	hiredman	bah
13:05	hiredman	forget it
13:06	beutdeuce	:P
13:06	stuartsierra	hiredman: what are you trying?
13:06	hiredman	,(let [_cons (fn [a b] (fn [c] (c a b))) car (fn [c] (c (fn [a b] a))) cdr (fn [c] (c (fn [a b] b)))] (car (_cons 1 nil)))
13:06	clojurebot	1
13:06	hiredman	,(let [_cons (fn [a b] (fn [c] (c a b))) car (fn [c] (c (fn [a b] a))) cdr (fn [c] (c (fn [a b] b)))] (cdr (_cons 1 nil)))
13:06	clojurebot	nil
13:06	Raynes	O_o
13:07	hiredman	cons cells as lambdas
13:07	hiredman	or lambdas as cons cells
13:08	stuartsierra	ah
13:13	hiredman	,(let [_cons (λ [a b] (λ [c] (c a b))) car (λ [c] (c (λ [a b] a))) cdr (λ [c] (c (λ [a b] b)))] (car (_cons 1 nil)))
13:13	clojurebot	1
13:15	beutdeuce	,(filter (fn [x] mod x 2 == 0) (range 1 100 1))
13:15	clojurebot	(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)
13:16	beutdeuce	hm
13:16	stuartsierra	You're not writing in lisp!
13:16	replaca_	,(filter (fn [x] mod x 2 == 0) (range 1 100 1))
13:16	clojurebot	(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)
13:17	stuartsierra	,(filter (fn [x] (= 0 (mod x 2)) (range 1 100 1))
13:17	replaca_	,(filter (fn [x] (= (mod x 2) 0) (range 1 100 1))
13:17	clojurebot	EOF while reading
13:17	clojurebot	EOF while reading
13:17	stuartsierra	,(filter (fn [x] (= (mod x 2) 0) (range 1 100 1)))
13:17	clojurebot	java.lang.IllegalArgumentException: Wrong number of args passed to: core$filter
13:17	replaca_	,(filter (fn [x] (= (mod x 2) 0)) (range 1 100 1))
13:17	clojurebot	(2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 60 62 64 66 68 70 72 74 76 78 80 82 84 86 88 90 92 94 96 98)
13:17	hiredman	…
13:17	stuartsierra	that's it
13:17	beutdeuce	right, i'm forgetting about the prefix evaluation
13:17	mebaran151	,(doc import)
13:17	clojurebot	"([& import-symbols-or-lists]); import-list => (package-symbol class-name-symbols*) For each name in class-name-symbols, adds a mapping from name to the class named by package.name to the current namespace. Use :import in the ns macro in preference to calling this directly."
13:18	mebaran151	cool so that's how the clojurebot works
13:18	mebaran151	I'd never actually thought about it before
13:18	durka42	,(filter even? (range 1 100))
13:18	clojurebot	(2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44 46 48 50 52 54 56 58 60 62 64 66 68 70 72 74 76 78 80 82 84 86 88 90 92 94 96 98)
13:18	hiredman	can't you just tell range to step by 2?
13:18	hiredman	(doc range)
13:19	clojurebot	"([end] [start end] [start end step]); Returns a lazy seq of nums from start (inclusive) to end (exclusive), by step, where start defaults to 0 and step to 1."
13:19	durka42	that would work too
13:19	hiredman	I guess not
13:19	durka42	,(range 1 10 2)
13:19	clojurebot	(1 3 5 7 9)
13:19	hiredman	,(take 10 (iterate (partial + 2)))
13:19	clojurebot	java.lang.IllegalArgumentException: Wrong number of args passed to: core$iterate
13:19	hiredman	,(take 10 (iterate (partial + 2) 0))
13:19	clojurebot	(0 2 4 6 8 10 12 14 16 18)
13:19	durka42	TIMTOWTDI
13:20	beutdeuce	Can clojure do lazy evaluation of infinite lists in list comprehensions?
13:21	mebaran151	it should be able to
13:21	hiredman	infinte sequences
13:21	beutdeuce	right, they are referred to as sequences
13:21	hiredman	sequences are lazy and can be infinite
13:21	mebaran151	,(take 10 (repeatedly 5))
13:21	clojurebot	java.lang.RuntimeException: java.lang.ClassCastException: java.lang.Integer cannot be cast to clojure.lang.IFn
13:21	hiredman	beutdeuce: lists and sequences are not the same thing
13:22	hiredman	sequences are a list like interface over most collection types
13:22	mebaran151	,(take 10 (repeat 5))
13:22	clojurebot	(5 5 5 5 5 5 5 5 5 5)
13:22	hiredman	,(class (repeate 5))
13:22	clojurebot	java.lang.Exception: Unable to resolve symbol: repeate in this context
13:22	hiredman	,(class (repeat 5))
13:22	clojurebot	clojure.lang.LazySeq
13:23	hiredman	,(count (repeat 5))
13:24	clojurebot	Execution Timed Out
13:27	mebaran151	is there anyway to sandbox Clojure btw?
13:28	mebaran151	such as turning off its java interop and such
13:28	stuartsierra	not in Clojure itself, but Java sandbox tools will work
13:28	mebaran151	ah you mean the security domain and such
13:28	hiredman	,(.getContents (java.net.URL. (java.io.File. "/etc/passwd")))
13:28	clojurebot	java.lang.ClassCastException: java.io.File cannot be cast to java.lang.String
13:29	hiredman	,(.canRead (java.io.File. "/etc/passwd"))
13:29	clojurebot	java.security.AccessControlException: access denied (java.io.FilePermission /etc/passwd read)
13:31	beutdeuce	http://pastie.org/557925 => runs forever, though i specified that an empty list returns 0
13:31	hiredman	beutdeuce: no
13:31	hiredman	[] means no args
13:31	beutdeuce	o
13:31	hiredman	clojure does not have pattern matching, it has destructered binding which does some similar things
13:32	beutdeuce	how would that work?
13:32	mebaran151	beutdeuce, you need to apply your mylen
13:32	mebaran151	right now you just continually pass in a length
13:33	beutdeuce	(doc apply)
13:33	clojurebot	"([f args* argseq]); Applies fn f to the argument list formed by prepending args to argseq."
13:33	mebaran151	pass in a list I mean
13:33	hiredman	uh
13:33	hiredman	mebaran151: I doubt that is what he wants
13:34	hiredman	(defn length ([lst] (length lst 0)) ([lst count] (if (seq lst) (recur (rest lst) (inc count)) count)))
13:37	hiredman	,(reduce (fn [a b] (inc a)) '(1 2 3))
13:37	clojurebot	3
13:37	hiredman	,(reduce (fn [a b] (inc a)) '())
13:37	clojurebot	java.lang.IllegalArgumentException: Wrong number of args passed to: sandbox$eval--3744$fn
13:37	hiredman	:/
13:38	hiredman	oh
13:38	hiredman	duh
13:38	hiredman	,(reduce (fn [a b] (inc a)) 0 '())
13:38	clojurebot	0
13:38	hiredman	,(reduce (fn [a b] (inc a)) 0 '(1 2 3))
13:38	clojurebot	3
13:39	beutdeuce	so reduce is haskell's fold ?
13:39	hiredman	but strict
13:39	beutdeuce	ah
13:39	hiredman	no lazy reduce in clojure
13:39	clojurebot	lazy is hard
13:41	hiredman	everything in clojure is strict, except lazy sequences
13:42	hiredman	some operations are sequences are lazy and some are not, generally if it returns another sequence (map, filter, etc) it is lazy
13:43	beutdeuce	k
14:38	rzoom	is there a standard "default" case for cond?
14:38	Chouser	yes, any true expression will do. :else is used by convention
14:38	hiredman	,(cond (= 1 0) :foo :else :bar)
14:38	clojurebot	:bar
14:39	rzoom	nice!
14:39	rzoom	thanks
14:39	hiredman	YEAH!
14:39	hiredman	TO THE MAX!
14:41	danlarkin	aw you guys... having all this fun
14:41	rzoom	is dmiles_afk ever here?
14:42	hiredman	~seen dmiles
14:42	clojurebot	dmiles was last seen quiting IRC, 759 minutes ago
14:44	dmiles_afk	i am here. clojure is in my tyodo pipeline for what i am coding
14:45	dmiles_afk	which is a secondlife client for clojure
14:45	dmiles_afk	but right now still using DotLisp.. but will soon enough use ClojureCLR
14:46	dmiles_afk	DotLisp was what rhickey wrote right before Clojure.. actually has a few nice extras that the JVM didnt offer
14:53	mebaran151	what does it have that the JVM can't offer?
16:42	wlr	Mention of real world Clojure ~3/4 thru article: http://www.expresscomputeronline.com/20090727/technology01.shtml
17:27	angerman	whee I just wrote my first macro
17:27	angerman	allows me to write (query "Item" where (= :name nil)
17:29	Raynes	Congratulations.
17:33	Anniepoo	Just a note to rich to say thanks for such a cool language
17:33	Anniepoo	8cD
17:35	angerman	though there is one thing I don't like so much. Descruturing in macros does not work :/
17:36	angerman	i couldn't do this: (defmacro [kind _ (op a b)] ...)
17:39	Anniepoo	I'm not getting the 'why is this cool?' of destructuring
17:40	angerman	Anniepoo: the cool part about destructuring is that it's short
17:41	Anniepoo	hmm....
17:41	angerman	for an extreme example take Prolog. If you couldn't do destructuring in the head it would be very painful
17:41	duck1123	it saves you a few let statements, and having to think up new names
17:41	Anniepoo	don't know Prolog
17:42	Anniepoo	ok, duck, but you're saying it's not really theoretically powerful, it's a 'typing' level convenience
17:42	hiredman	angerman: uh, destructured binding in clojure does not use parens
17:43	angerman	hiredman: that's what I guessing towards ...
17:43	angerman	well ... bedtime
17:43	hiredman	ugh
17:43	hiredman	~destructuring
17:43	clojurebot	destructuring is http://clojure.org/special_forms#let
17:44	Anniepoo	I can see that clojure will be an insanely great language to write graphic programs in
17:45	Anniepoo	cause they usually bog down in the details of breaking down and packaging up small data structures
17:46	Anniepoo	I've had an idea for graphics and never had a good language to express it in.
17:46	Anniepoo	most bugs in graphics code turn out to be using the wrong coordinate system
17:46	Anniepoo	so a way to 'enforce type' on coord systems would be insanely cool
17:47	duck1123	Clojure gives you loads of tools to make your code as close to the actual idea as possible
17:47	duck1123	destructuring is just one of those tools
17:49	Anniepoo	yah, I want to have some (move-widget-to [pt] and have it cope with a point in any format and coordinate system without a bunch of code in move-widget
17:50	Anniepoo	ideally, so you just say (cast-coord widget-frame 'Point pt) and you get a java.awt.Point in the widget's frame of ref
17:50	duck1123	sounds like a job for multimethods
17:51	Anniepoo	definitely
17:51	duck1123	there was a blog post where the guy was making a raytracer in clojure, you might want to check that out,
17:52	Anniepoo	ah, interestiing
17:52	duck1123	http://www.fatvat.co.uk/2009/01/ray-tracing-in-clojure.html
17:53	Anniepoo	nice, thanks
17:53	Anniepoo	that acutally kind of answers my question about destructuring
17:54	hiredman	http://tu.be/graphics/ <-- this a more recent attempt by someone
17:56	Anniepoo	nice
17:56	hiredman	he went by fsm and was in here a few days ago looking for performance tips for array operations if I recall
17:58	Anniepoo	In a curly brace language it's good style not to 'pass the world' around. Does destructuring make that advice more or less apt?
17:59	Anniepoo	right now I'm making a swing GUI editor for bunch of widgets
17:59	hiredman	destructuring has nothing to do with that
17:59	hiredman	it jsut makes binding forms shorter
18:00	Anniepoo	example, right now I'm writing the call in paintComponent that will acutally draw the widgets
18:00	Anniepoo	(draw-laid-out-widgets (.state this) g)
18:00	hiredman	(let [a (first lst) b (second lst) c (rest (rest lst)] ) becomes (let [[a b & c] lst] )
18:01	Anniepoo	Ah!!!!!
18:01	Anniepoo	8cD
18:01	Anniepoo	Destructuring, it's not just for defn any more
18:01	Anniepoo	sorry, I didn't get it before, somebody mentioned let
18:02	Anniepoo	now I see
18:02	Anniepoo	like the first raytracer posted, the guy has a bunch of (:x pt) (:y pt) stuff
18:04	duck1123	that one of the great uses, you can pull those values out into their own variables to make the rest of the code shorter
18:04	Anniepoo	he could have (let [[x y z] pt] ... x y ...
18:04	Anniepoo	got it! thanks!
18:04	Anniepoo	Y'all are wonderful
18:05	duck1123	actually (let [{x :x y :y z :z} pt] ... I believe
18:05	Anniepoo	oops
18:06	duck1123	check out the let section of the special forms page
18:06	Anniepoo	my own code has a bunch of positional args
18:06	duck1123	it shows all the things you can do
18:07	Anniepoo	yah, reading now
18:07	replaca_	you can try (let [{:keys [x y z]} pt] ... ) and get the same effect
18:07	duck1123	(let [{:keys [x y z]} pt] works too
18:07	Anniepoo	that's nice, and shorter
18:08	Anniepoo	I'm shocked at the power of all this
18:08	Anniepoo	I'm making a swing GUI - not a promising candidate for shortening, one would think
18:08	duck1123	it's never worth it to be shocked by the power of clojure, you'll wear yourself out.
18:10	duck1123	I've never played with swing, but isn't there some contrib libraries for doing gui stuff? Have you checked those out?
18:10	Anniepoo	I started writing this in java, thinking I didn't know Clojure well enough to use it on a real project
18:10	codyK	any examples of dealing with circular references?
18:11	hiredman	,(doc declare)
18:11	clojurebot	"([& names]); defs the supplied var names with no bindings, useful for making forward declarations."
18:11	codyK	deref on a object that contains circular references seems to unroll the whole thing & blow the stack
18:11	Anniepoo	sure, if you print it
18:11	codyK	even in code
18:12	duck1123	"don't do that then" :)
18:12	codyK	hahaha
18:12	codyK	yeah, so, in languages i'm familiar with, if youre say writing a serializer, its not too big a deal to deal with circular references
18:12	codyK	but clojures printer doesnt seem to deal with them
18:13	codyK	and searching for examples or discussion of it isn't very helpful
18:13	duck1123	how do the other languages deal with them? just elide the recursion?
18:13	Anniepoo	you can't deal with it in general
18:14	Anniepoo	it requires solving the halting problem
18:14	codyK	eh?
18:14	codyK	just keep a list of addresses you've seen
18:14	codyK	check it before you deref a pointer
18:14	codyK	only deref one level down at a time
18:14	hiredman	uh
18:15	hiredman	sure sure, pointers
18:15	hiredman	…
18:15	hiredman	of course the jvm being designed to stop you from dealing with raw pointers sort of throws a monkey wrench in that
18:16	Anniepoo	or mark and sweep
18:16	codyK	the language clearly has reference types
18:16	hiredman	sure
18:16	Anniepoo	cody, you don't even need to do that
18:16	Anniepoo	add a bit to each object
18:16	hiredman	but it doesn't let you touch pointers
18:17	Anniepoo	reset them all, if you have some global list of stuff
18:17	Anniepoo	then traverse and print but say 'oh, I visited'
18:17	codyK	the issue i'
18:17	codyK	m having is that deref seems to be all or nothing
18:18	Anniepoo	it's reasonable to blow up on a circ ref for something general like print
18:18	codyK	i dont think its reasonable
18:18	codyK	common lisp handles it for instance
18:18	Anniepoo	thre's a way to set the max depth of printing
18:18	Anniepoo	~println
18:18	clojurebot	,(println "clojure bot is the best!")
18:18	hiredman	,(let [a (ref 10)] (deref (ref a)))
18:18	clojurebot	#<Ref@94f797: 10>
18:18	hiredman	^- not all or nothing
18:19	durka42	,print-depth
18:19	clojurebot	java.lang.Exception: Unable to resolve symbol: print-depth in this context
18:19	codyK	ok, maybe my code is bad
18:19	hiredman	it only derefs one layer
18:19	durka42	it was something like that...
18:19	durka42	,print-level
18:19	clojurebot	30
18:19	hiredman	codyK: anyway, IDerefable things are not printable
18:20	hiredman	and not readable
18:21	codyK	the things they reference are / could be
18:21	hiredman	sure
18:21	hiredman	so deref before you print
18:26	codyK	;; (def a (atom [0]))
18:26	codyK	;; (def b (atom [1]))
18:26	codyK	;; (compare-and-set! a @a (conj @a b))
18:26	codyK	;; (compare-and-set! b @b (conj @b a))
18:27	Chouser	you can do the same thing with java collections and their .toString() method
18:28	Chouser	but that just crashes instead of respecting print-level
18:29	hiredman	codyK: so?
18:29	codyK	I would expect to be able to get some unique identifier for b without unrolling it
18:30	hiredman	unique how?
18:30	codyK	well, i'm currently trying to add it to a set
18:30	codyK	so that I can compare it for set membership
18:30	hiredman	across all possible jvms which might run clojure and read it in?
18:30	codyK	no
18:31	hiredman	a unique identifier for a mutable thing like an atom or a ref is not worth much
18:31	codyK	but the act of comparing it for set membership blows the stack
18:31	codyK	(afaict anyway)
18:31	hiredman	b is an atom
18:31	hiredman	atoms are designed to mutate
18:32	codyK	again, if i have a pointer, the address of that pointer doesnt change
18:32	blbrown_win	Chouser, are you the owner of clojure and me
18:32	hiredman	you don't have a pointer
18:32	codyK	even though the contents of that address might change
18:32	hiredman	this is not C or C++
18:32	hiredman	you do not have a pointer
18:33	codyK	look, i'm not having a language argument
18:33	codyK	i promise you
18:33	codyK	s/pointer/reference/
18:33	codyK	clearly there is something consistent about b, because i can refer to it with the same name
18:33	Chouser	blbrown_win: no
18:33	codyK	so whats the correct operator to get that name?
18:35	Chouser	codyK: by default printing an IDeref also derefs and prints that value. But = doesn't do that
18:35	hiredman	codyK: how are you adding b to a set?
18:35	blbrown_win	Chouser, oh, I just followed you on twitter,
18:35	Chouser	= on reference types should be just 'identical?'
18:35	codyK	My code doesnt even get to the point of adding it to a set
18:35	hiredman	(conj #{} b) will print the results to the repl
18:35	blbrown_win	ahh, he is cgrand
18:35	hiredman	codyK: so add some prns and figure out where it actually stops
18:36	codyK	if my set is seen, bound to an empty set, doing (seen some-circ-ref)
18:36	codyK	alone
18:36	codyK	seems to blow the stack
18:36	codyK	i'm willing to bet its bad code on my part
18:37	codyK	just curious about examples
18:37	hiredman	all works fine for me
18:38	hiredman	user=> (#{} b)
18:38	hiredman	nil
18:38	hiredman	user=>
18:38	codyK	yeah i just tried it
18:41	codyK	(and ((conj #{} b) b) 'thing)
18:41	codyK	seems to work ok
18:41	codyK	so like I said, bad code
18:41	akhudek	Hi, I'm fairly new to clojure and failing at figuring out the right way to do this. I need to store a map of positions to counts that is updated roughly at random. There are typically around a million positions. In other languages I'd implement this as a vector of ints where the position in the vector is the 'key'. But looking at clojure vectors, perhaps this is not the right way to go about it? A big hash map doesn't seem like a good i
18:44	hiredman	,(do (vec (range 1000000)) nil)
18:44	clojurebot	nil
18:44	hiredman	what is the problem with clojure vectors?
18:44	akhudek	I can't figure out a way to update a single position.
18:44	hiredman	,(assoc [1 2 3] 0 5)
18:44	clojurebot	[5 2 3]
18:45	Anniepoo	oooh
18:45	akhudek	ah, thanks, I didn't see how assoc worked on vector given the docs
18:46	hiredman	,(doc assoc)
18:46	clojurebot	"([map key val] [map key val & kvs]); assoc[iate]. When applied to a map, returns a new map of the same (hashed/sorted) type, that contains the mapping of key(s) to val(s). When applied to a vector, returns a new vector that contains val at index. Note - index must be <= (count vector)."
18:46	Anniepoo	ok, what's going the other way, how do I get the 593rd element?
18:47	Chouser	nth
18:47	Anniepoo	,nth
18:47	clojurebot	#<core$nth__4198 clojure.core$nth__4198@13e38a7>
18:47	hiredman	,((vec (range 1000000)) 593)
18:47	clojurebot	593
18:47	akhudek	I should read more thoroughly, I saw map in there and didn't read the last part. :/
18:47	hiredman	,(nth (vec (range 1000000)) 593)
18:47	clojurebot	593
18:47	hiredman	vectors are functions of their indices
18:48	akhudek	right, so I guess they are a type of map
18:48	hiredman	not really
18:49	hiredman	but they implement a few of the same interfaces
18:49	Chouser	maps and vectors are both associative -- can get and set elements based on a key
18:49	hiredman	,(ancestors (calls []))
18:49	clojurebot	java.lang.Exception: Unable to resolve symbol: calls in this context
18:49	hiredman	,(ancestors (class []))
18:49	clojurebot	#{clojure.lang.IPersistentStack clojure.lang.Seqable java.util.List clojure.lang.Reversible clojure.lang.Sequential clojure.lang.AFn clojure.lang.Obj clojure.lang.Counted java.util.concurrent.Callable clojure.lang.APersistentVector java.io.Serializable java.util.Collection clojure.lang.IObj clojure.lang.IPersistentCollection clojure.lang.IMeta clojure.lang.IPersistentVector clojure.lang.IFn java.lang.Iterable java.lang.Ob
18:49	Anniepoo	,([1 3 15 4] 2)
18:49	clojurebot	15
18:50	Chouser	,(instance? clojure.lang.Associative [])
18:50	clojurebot	true
18:50	Chouser	,(instance? clojure.lang.Associative {})
18:50	clojurebot	true
18:51	akhudek	There are so many clever things in clojure. It really is a joy to use.
19:09	blbrown_win	Anyone know where the widefinder input is. The input log file .I am running widefinder in clojure.
19:17	hiredman	http://svn.effbot.org/public/stuff/sandbox/wide-finder/getdata.py is apparently a python script to download the logs
19:22	blbrown_win	so you can't just dow the download
19:22	blbrown_win	hiredman, did you google that, I couldn't find it apparently
19:22	hiredman	it looks like you can
19:23	hiredman	just copy the url out of the python script
19:26	blbrown_win	isn't kind of a crappy benchmark
19:28	hiredman	have you read the rationale behind it?
19:29	blbrown_win	there seem to be volumes out there? it also seems to be centered around ruby
19:29	blbrown_win	but not really, I guess just a random test for processing data. How do other languages compare
19:29	hiredman	http://blogs.zdnet.com/Murphy/?p=1030
19:31	hiredman	http://www.hanselman.com/blog/TheWeeklySourceCode9WideFinderEdition.aspx some C# widefinders
19:33	hiredman	man
19:33	hiredman	I wish I had a T2
19:37	krumholt__	is there an api for quote?
19:37	hiredman	uh
19:38	hiredman	like ' quote?
19:38	krumholt__	yes
19:38	hiredman	what's there to say about it?
19:38	hiredman	it supresses evaluation
19:38	krumholt__	for example why does (quote 123) evals to 123 a number
19:39	hiredman	because the reader readers 123 as a number
19:40	hiredman	and emits the list (quote 123) which is a two element list, the first element is the symbol quote and the second is the number 123
19:40	hiredman	then when (quote 123) is evaluated, 123, a number, is the result
19:41	JAS415	in a macro, how should I handle java classes... for example i have a code that looks like `(. ~URLEncoder (encode ~string "UTF-8")) should I be importing the URLEncoder in the macro code and escaping it with ~ or should I be importing the URLEncoder in the code that the macro goes with?
19:41	krumholt__	hiredman, seems strange that (= 123 '123) => true
19:42	hiredman	'123 is not a symbol
19:42	hiredman	(class '123)
19:42	hiredman	,(class '123)
19:42	clojurebot	java.lang.Integer
19:42	JAS415	is because numbers evaluate to themselves
19:42	hiredman	' does not make symbols
19:42	hiredman	' quotes the following form, which is often a symbol
19:42	JAS415	,(= true 'true)
19:42	clojurebot	true
19:42	JAS415	(= false 'false)
19:43	JAS415	,(= false 'false)
19:43	clojurebot	true
19:43	krumholt__	seems wrong to me :)
19:43	JAS415	think of quote as a way of delaying evaluation
19:44	hiredman	krumholt__: so when the reader comes across the string "(quote 123)" it goes "oh look, a paren, that means I am reading a list!"
19:44	hiredman	so then it gets into list reading mode
19:45	hiredman	it reads in the string "quote" says "well this is a symbol" and the string "123" and says oh, this is a number
19:45	hiredman	so out of the reader you get a list, (quote 123) where quote is a Symbol and 123 is an Integer
19:46	hiredman	then the list is evaluated
19:46	krumholt__	ok i get it thanks
19:46	hiredman	the evaluater looks at the list and says "this is a list, so it could be function application, but quote is a special form"
19:46	hiredman	really?
19:47	hiredman	I was having fun anthropomorphizing the repl
19:47	hiredman	man
19:47	clojurebot	☝(^_^)☝
19:47	krumholt__	:)
19:47	hiredman	I could get some puppets
19:47	cemerick	hiredman: been a long week? :-)
19:47	hiredman	and do a video
19:47	hiredman	cemerick: come to think of it, it has
19:48	cemerick	yeah. Been a long year so far, too.
19:48	hiredman	I don't even think _why has done puppets
19:49	cemerick	I don't get the appeal of his stuff at all.
19:49	JAS415	oh man
19:49	cemerick	though, I don't generally enjoy instructional material anyway
19:49	JAS415	puppets to explain clojure would be awesome
19:50	hiredman	someone did some kind of alagator toy lisp thing
19:51	hiredman	http://visual-languages.blogspot.com/2009/07/alligator-eggs-revisited.html
19:53	krumholt__	i have another question. how can i estimate the cost of a recur? in the documentation it says it will only use constant space. whats the difference to tail call optimization?
19:53	hiredman	tco is more general
19:54	hiredman	recur only recurs to the closest loop or fn
19:54	Chouser	there is no tail call optimization in clojure
19:54	hiredman	tco would allow you to call any function
19:55	hiredman	from tail call position, and not build up stack frames
19:55	krumholt__	ok thanks
20:21	blbrown_win	http://paste.lisp.org/display/84158 anyone see the error here. I got the code from a blog (cgrand-rec, hehe). I can't find 'val' created either
20:24	lisppaste8	Anniepoo pasted "compile error" at http://paste.lisp.org/display/84159
20:25	Anniepoo	me, meanwhile, I've just reached the end of my wits trying to figure out why the compiler is complaining
20:26	JAS415	that's weird
20:26	Anniepoo	what's weird?
20:27	JAS415	if you use fn instead of # does it work?
20:27	Anniepoo	for the # in draw-laid-out-widgets?
20:28	durka42	filter's fn only gets one argument
20:28	JAS415	ah yeah
20:28	JAS415	that is it
20:28	Anniepoo	ah!
20:28	Anniepoo	thanks
20:28	durka42	you can destructure it however
20:28	Anniepoo	yes, meant [[_ lvl _]]
20:28	durka42	precisely
20:28	Anniepoo	thanks
20:31	Chouser	fwiw I make that mistake frequently
20:31	krumholt__	what does _ do?
20:32	Chouser	Anniepoo: that's yours
20:35	Anniepoo	sorry?
20:35	Anniepoo	what's mine?
20:36	Chouser	krumholt__'s question is yours to answer. time to step up! :-)
20:36	Anniepoo	underscore is idiomatic for an argument that you will ignore
20:36	Anniepoo	yes, you're right
20:37	Anniepoo	krumholt, if you need a function that takes 3 args, because that's what the caller expects, but
20:37	krumholt__	ah ok so i could write (foo nil a nil) or (foo _ a _) ?
20:38	Anniepoo	suppose you have a bunch of places that hang onto a function and call it
20:39	Anniepoo	like, I'm making a thing where each widget is constrained to others, so each widget has a reference to it's constraint function
20:39	krumholt__	ok i get it i can only use _ in function definiton
20:39	Anniepoo	it's not actually part of the language, it's just an idiom
20:40	Anniepoo	(defn foo [x x x] ... something....) is valid Clojure, just not very useful
20:40	Anniepoo	(foo 1 2 3)
20:40	Anniepoo	means x is 3 inside
20:41	Anniepoo	_ is, by convention, what you use to mean 'I'm not using this'
20:41	Anniepoo	but you're right, it normally only appears in a defn or destructuring in a let
20:41	krumholt__	oh ok so _ is just a normal symbol. it has no special syntax. you just use it so a human reader knows you will ignore the parameters
20:41	Anniepoo	yes
20:41	krumholt__	ok thanks
20:42	Anniepoo	ok, now, looking upwards in the Clojure food chain
20:42	lisppaste8	Anniepoo annotated #84159 "vec" at http://paste.lisp.org/display/84159#1
20:43	Anniepoo	259 is the last line of draw-laid-out-widgets
20:43	Anniepoo	I assume I'm somehow cdr'ing off the end of the list
20:43	Anniepoo	but don't understand how
20:45	ataggart	it might make things clearer to use a struct instead of relying on position
20:46	Chouser	Anniepoo: is there more to that exception?
20:47	Chouser	That doesn't look like the root cause trace
20:47	Anniepoo	I'll paste the whole thing
20:47	Chouser	I wonder if somewhere you have 'vec' when you want 'vector'
20:48	ataggart	btw where are these "widgets" created?
20:48	ataggart	I infer a widget is some vector or list
20:49	ataggart	but I never see it get created in that code
20:49	Anniepoo	no, this is a fragment
20:49	ataggart	so we may be missing some crucial info
20:49	Anniepoo	yes, but I can't paste the entire project in
20:50	ataggart	if I had to guess, the code we aren't seeing is being indirectly realized intoa vec during the destructuring, but it can't so it blows up
20:50	Anniepoo	Chouser is right, there's other stuff highter in the stack trace
20:51	lisppaste8	Anniepoo annotated #84159 "more of the code" at http://paste.lisp.org/display/84159#2
20:52	Anniepoo	240 is this let in draw-a-laid-out-widget
20:52	Anniepoo	(let [[x y w h] (as-screen state r)]
20:52	Chouser	vec only takes 1 arg
20:52	Chouser	,(vec (list 1 2 3 4))
20:52	clojurebot	[1 2 3 4]
20:52	ataggart	(vector, not vec
20:52	Chouser	if you want a vector of 4 things, use (vector 1 2 3 4) or just [1 2 3 4]
20:52	ataggart	ya, what chouuser said
20:53	Anniepoo	ah, ok
20:53	Anniepoo	the vec in as-screen
20:53	ataggart	(defstruct screen :x :y :z :w :h) imo
20:53	Anniepoo	incidentally, it's obvious there's a more idiomatic way to write that lanie
20:54	Anniepoo	but I don't know what it is
20:55	ataggart	(map #(* scale %) [ x y z])
20:55	Anniepoo	thanks!
20:55	ataggart	except with the right names hehe
20:55	Anniepoo	doh!
20:55	Chouser	and then you do want vec to get a vector
20:55	ataggart	no need
20:55	ataggart	unless you need to manipulate it
20:56	ataggart	leave it as a lazy seq
20:56	ataggart	if that grouping is important, it might be better to make use a struct instead of a vector
20:57	Anniepoo	yes, I understand
20:58	ataggart	though the input looks to be some kind of list, so it might be irrelevant
20:58	Anniepoo	I have a bunch of 'widgets' - rectangles essentially
20:58	Anniepoo	they're constrained to each other - I'm effectively making a layout manager
20:58	ataggart	(defn scale [factor values] (map #(* scale %) values))
20:58	ataggart	erm
20:59	ataggart	(defn scale [factor values] (map #(* factor %) values))
20:59	ataggart	then it'll take anything
21:00	Anniepoo	there's a root widget, it has 'children' and 'slaves', so there's an n-ary tree that has two kinds of parent-chi
21:00	Anniepoo	child relationship
21:00	ataggart	ya I just mean you can avoid the "x y w h" stuff
21:01	Anniepoo	oh, I see
21:01	Anniepoo	actually, there's a stylistic reason to leave it in this instance
21:01	ataggart	k
21:02	Anniepoo	in this case we've got a lot of different small structs with coordinates running around
21:02	Anniepoo	[x y w h] is pretty clear compared to widget-location
21:03	ataggart	you can have both
21:03	ataggart	e.g., have one call the other
21:03	ataggart	but, yeah it's a style thing now
21:05	Anniepoo	I write lots of graphics stuff so I see this issue with representing points and rects all the time
21:06	Anniepoo	it's the main reason I'm excited by Clojure, I can see a cure for all that
21:06	krumholt__	is clojure compiling to java or directly to bytecode?
21:06	Anniepoo	bytecode
21:07	ataggart	isn't there an easier way to do something like (reduce #(let [[k v] %2] (assoc %1 k (* 2 v))) {} {:a 1 :b 2})
21:08	ataggart	esenatially applying a function to the values of a map
21:10	krumholt__	(map foo (vals{:a 1 :b 2}))
21:10	krumholt__	?
21:10	ataggart	but end up with a new, altered map
21:10	ataggart	not a seqence of the values
21:11	ataggart	,(reduce #(let [[k v] %2] (assoc %1 k (* 2 v))) {} {:a 1 :b 2})
21:11	clojurebot	{:b 4, :a 2}
21:11	cschreiner	(trying really hard to understand agents)
21:12	ataggart	it's easy, they're just like theads, except totally different ;)
21:12	cschreiner	right..
21:13	cschreiner	they are just values?
21:13	ataggart	they hold a value
21:13	ataggart	then you send it an action to update that value, and that action will be run in a thread pool
21:13	cschreiner	and when you want to change the holded value, you send off a function?
21:14	ataggart	send or send-off, yeah
21:14	cschreiner	thats it?
21:14	ataggart	yup
21:14	ataggart	well, sort of
21:14	cschreiner	jebuz
21:14	ataggart	the "neat" part is how they work within a transaction
21:15	cschreiner	how they operate with the STM
21:15	ataggart	in fact the one agent I've ever used, the value never changed since it was used to generate output
21:15	ataggart	~agents
21:15	clojurebot	Huh?
21:15	ataggart	http://clojure.org/agents
21:15	ataggart	grep for transactin
21:15	ataggart	except spelled correctly
21:16	cschreiner_	,agents
21:16	clojurebot	java.lang.Exception: Unable to resolve symbol: agents in this context
21:16	cschreiner_	,agent
21:16	clojurebot	#<core$agent__4304 clojure.core$agent__4304@171a8f6>
21:17	cschreiner_	to look at an agent as a generator is odd
21:17	ataggart	yep, who does that?
21:17	cschreiner_	you do
21:18	ataggart	sorry, "generator" means something specific to me
21:18	cschreiner_	ok
21:18	ataggart	since you can't have side effects in a transaction (such as outputting logging) you do it via an agent
21:18	cschreiner_	so its like a monda?
21:18	cschreiner_	monad?
21:18	ataggart	or rather, any side effects will reoccur
21:18	ataggart	it's an agent
21:18	ataggart	think of it as an agent, and it'll be easier
21:19	cschreiner_	i can speak to the agent, I can also receive something
21:19	cschreiner_	the agent is independent from me, that is, it runs in another thread?
21:20	ataggart	I suggest reading the link I posted above
21:20	cschreiner_	I have
21:20	hiredman	http://www.amazon.com/Multi-agent-systems-introduction-distributed-intelligence/dp/0201360489 found this today for $2 at goodwill
21:20	Anniepoo	ok,what if it's the side effect that has to succeed for the transaction to succeed?
21:21	ataggart	ha! where do you live that suck would appear in goodwill
21:21	ataggart	*such
21:21	hiredman	Anniepoo: no
21:21	cschreiner_	ah, great find!
21:21	hiredman	seattle
21:21	ataggart	ah ok then
21:21	ataggart	Anniepoo: I misspoke earlier
21:22	ataggart	it's not that you can't have sideeffects, it's that if the transaction needs to be rerun, those side effects will occur again
21:22	ataggart	e.g. logging may get printed multiple times
21:22	Anniepoo	I found a palm pilot at goodwill a few weeks ago that had been 'experimented' on - they'd messed with the IR
21:22	hiredman	but sends to an agent are held unto the transaction completes
21:22	Anniepoo	ok, so suppose I have a machine that punches a hole in a paper tape and advances
21:23	hiredman	Anniepoo: ah, a late model turing machine
21:23	cschreiner_	ah, i was thinking of an agent in the spy-movie setting, now, when put into a manager-for-someone, it makes more sense...
21:23	ataggart	lol
21:23	Anniepoo	and some call that initiates the hole punching
21:24	Anniepoo	8cD that's a side effect for sure
21:24	Chouser	Anniepoo: yeah. don't do that in a transaction.
21:25	ataggart	if you can't roll it back...
21:25	Anniepoo	not without sticky tape!
21:26	Anniepoo	(initiate-range-safety)
21:40	hiredman	(pl (λx.x 1))
21:40	hiredman	,(pl (λx.x 1))
21:40	clojurebot	1