#clojure – 2009-12-21

2009-12-21

00:15	Blackfoot	what's the best way to get a library like http-client installed with OSX after `port install clojure +rlwrap`
00:16	Blackfoot	i tried installing the jar from clojars, but get an error: java.lang.NoClassDefFoundError: clojure/lang/RT (NO_SOURCE_FILE:0)
01:15	pdk`	can doto be applied to a class and have static methods of that classed used therein
01:16	pdk`	or just use ..
01:31	bytecoder	if it's just a single call you can use it like ,(Math/sqrt 4)
01:31	bytecoder	,(Math/sqrt 4)
01:31	clojurebot	2.0
02:44	johnmn3	by chance, anyone know how to pull the text to the left or right of a caret in a JTextArea lazily?
03:10	johnmn3	how do you type hint a char[] and an int?
03:25	LauJensen	Morning team
03:32	johnmn3	an Unmatched delimiter: ) is kicking my butt!
03:33	johnmn3	I honestly don't see it.
03:33	vegai	a better text editor might help ;)
03:33	ska2342	johnmn3: that's probably from the smilies in your code ;-)
03:33	johnmn3	gah
03:33	ska2342	johnmn3: can we look at the code?
03:34	johnmn3	sure, one sec
03:35	lisppaste8	johnmn3 pasted "get-expr unmatched paren?" at http://paste.lisp.org/display/92408
03:35	johnmn3	forgive the ugliness
03:36	johnmn3	Exception in thread "main" java.lang.Exception: Unmatched delimiter: ) (ed.clj:69)
03:37	ska2342	johnmn3: I think the line .addKeyListener input is missing it's closing paren
03:38	ska2342	johnmn3: according to your indentation it should be closed on the line with .requestFocus input ...
03:39	johnmn3	ska2342: if you hover over the whole form, don't you see the colored parens matched up?
03:40	vegai	johnmn3: do you ever close the [out starting on l66
03:40	johnmn3	ah there it is.
03:41	johnmn3	must have bumped a key or something.. was working fine before
03:41	johnmn3	the unmatched ")" had me looking for a close parens
03:42	johnmn3	ok, it compiles now. thanks
03:56	edbond	how to get nth element from lazySeq?
04:03	somnium	,(nth (range 1000) 42)
04:03	clojurebot	42
05:33	fliebel	Good morning! At least for me it is… Is there a simple way to walk a directory tree in Clojure? I know there is duck-streams, but as far as I can tell that does not include walking. Even looking around in java.io does not give me a clear way to do it.
05:33	cark	,(doc file-seq)
05:33	clojurebot	"([dir]); A tree seq on java.io.Files"
05:36	fliebel	wow, thanks :)
05:36	cark	=)
05:36	cark	carefull printing this sequence ... might take a while !
05:36	fliebel	I understand :)
05:37	fliebel	hmm, it seems I have to create a java.io.file?
05:37	cark	that's easy : (take 10 (file-seq (java.io.File. "c:\\")))
05:39	fliebel	user=> (file-seq (java.io.File. "~/downloads/"))
05:39	fliebel	(#<File ~/downloads>)
05:39	cark	i don't know if ~works in java land
05:40	fliebel	no, it does not, it seems :D
05:40	cark	there are some java properties that do the same i think
05:41	cark	,(System/getProperty "user.dir")
05:41	clojurebot	java.security.AccessControlException: access denied (java.util.PropertyPermission user.dir read)
05:41	somnium	'(doc file-str)
05:41	somnium	,(doc file-str)
05:41	clojurebot	"clojure.contrib.duck-streams/file-str;[[& args]]; Concatenates args as strings and returns a java.io.File. Replaces all / and \\ with File/separatorChar. Replaces ~ at the start of the path with the user.home system property."
05:41	cark	ah there you go !
05:42	fliebel	cool
05:42	fliebel	I don't need that, but ~ was the most lazy way to get to a dir with very few files.
05:42	cark	though on windows it goves me the windows home directory instead of the one i defined using the HOME variable
05:42	cark	ahwell
05:43	somnium	hmm, it calls System/getProperty
05:44	cark	well it's the right thing to do anyways
05:44	somnium	maybe we can blame it on java :P
05:44	cark	booo java !
05:45	fliebel	Why is Clojure built on Java? It gives awesome power, but also awesome ugliness…
05:45	somnium	it will go other places eventually
05:45	cark	i think clojure wants to be practical, and the jvm is very much so
05:45	somnium	Ive been wondering about PLT Scheme lately, it has immutable lists now and compiles to C
05:46	cark	yep plt scheme is nice
05:46	fliebel	I think having Clojure compiled to C and running on Parrot would be 2 cool things to do :)
05:46	cark	i'd love to have serialisable continuations in clojure =P
05:46	somnium	and its GCed, so the path should be shorter than C#/Objective-C
05:46	somnium	er
05:47	somnium	C++ I meant
05:47	somnium	a msg on the mailing list about clojure in C++
05:48	somnium	fliebel: is parrot finished?
05:48	fliebel	somnium: no idea… they are at verion 1.9 I believe.
05:50	fliebel	But I'm not sure if it's taken very seriously by developers, there are a lot of hobby project and dead/unfinished languages...
06:38	johnmn3	if anyone is using chrome, trying going here and tell me if you get popups: http://gauss.ececs.uc.edu/Users/Franco/Week6/H7/InputStreamTest.java.html
06:38	johnmn3	for me, chrome is actually running the .java files
06:38	johnmn3	and they actually work.
06:43	somnium	johnmn3: they come up here, nothing in them
06:44	somnium	johnmn3: whats the expected behavior?
06:44	johnmn3	right, if you write into the input stream, it comes out in the output stream.
06:44	johnmn3	boxes
06:44	johnmn3	I'm just surprised chrome compiles them and runs them.
06:44	johnmn3	could be a security issue.
06:45	somnium	ah, its not an applet
06:45	somnium	thats a definite bug report
06:46	johnmn3	InputStreamTest extends applet, but nothing says "do you want to run this applet"
06:46	johnmn3	you running chromium on linux? might me linux specific
06:47	johnmn3	/s/me/be
06:49	LauJensen	Has anyone tried working with cookies+contrib/http-agent ?
06:53	somnium	yeah linux, still beta I guess
06:54	somnium	the js console is awesome though
06:57	johnmn3	that's what I hear.. better than firebug..
06:57	somnium	johnmn3: it even has built in autocompletion
06:58	johnmn3	I like how when searching a page for a term, the right scrolling slide area shows highlights for hits, like a code editor
07:00	somnium	Im bet they have some amazing tools in-house
07:02	johnmn3	yea, have you heard about that emacs in firefox thing?
07:02	johnmn3	ymacs or yemacs or something..
07:03	johnmn3	clojure on top of JS, using ymacs :)
07:03	somnium	yes! that is really impressive
07:03	johnmn3	(would be cool, I mean)
07:03	somnium	totally possible
07:04	somnium	the hardest part of clojure -> js is mutability
07:05	somnium	but no threads, so maybe a js 'dialect' would be more appropriate
07:05	somnium	that or hide all data in a closure
07:05	LauJensen	You guys haven't seen ClojureScript ?
07:06	somnium	LauJensen: of course
07:07	somnium	its almost a translation from clojure to js
07:07	somnium	does it run aganst master?
07:07	somnium	Ive browsed the source quite a bit, but havent actually fired it up
07:09	johnmn3	never tried it.
07:10	LauJensen	I think its bit-rotted
07:10	LauJensen	Chris has talked about picking it back up, after clojure-in-clojure is done
07:12	somnium	Im doing a special-form only compiler, something like a mini-scheme with clojure semantics
07:13	somnium	implementing the entire clojure runtime and its data-structures is a task better left to smart people like Chris :)
07:16	somnium	\me wonders if mini-scheme might be a redundant statement
07:40	LauJensen	somnium: Whats the purpose ?
07:45	somnium	LauJensen: of my yet-another-js-generator?
07:45	LauJensen	mini-scheme
07:46	somnium	clojure has lots of xml generators, why not multiple js ones?
07:46	cark	the task is harder
07:46	somnium	its for my own use in web app stuff
07:46	LauJensen	Ok, fair enough
07:47	somnium	var(kind of), let, loop, recur, do, fn*, if
07:47	somnium	thats my mini scheme in essence, + an approximation of deftype
07:48	cark	and it compiles to js ?
07:48	somnium	my main hope is to bring the macro system to bear on ius
07:48	somnium	yeah, though its hardly been tested
07:49	somnium	defype needs some tricks with closures to simulate immutability
07:49	cemerick	writing another lisp is a self-justifying exercise :-)
07:49	somnium	js is a poor man's scheme anyway :P
07:49	cark	it's a superset of scheme, give it some credit !
07:49	cark	on some respect at least
07:50	somnium	:)
07:50	cark	the object system
07:50	cark	pretty neat
07:50	somnium	I quite like js actually, but I like clojure more
07:50	cark	the syntax is horrible
07:50	cark	there is no way to have a sane indentation
07:51	LauJensen	cemerick: really? :) I've always found it pointless when people implemented other lisps in Clojure - Like Qi for instance. Nobody asks Mercedes to implement a Fiat on their technology.
07:51	cark	LauJensen : what if you want some features you don't have in clojure
07:52	cark	LauJensen: looking at your blogpost about arc challenge
07:52	LauJensen	See ClojureQL for an example
07:52	cark	you don't have continuations in clojure, so your solution is very different
07:52	LauJensen	Thats right
07:52	cark	also you might want a restricted scripting language for your app
07:53	cemerick	LauJensen: programming languages are modes of expression, as well as tools, and the search space in both cases is infinite. Once you stop looking for better options in either category, you're doing yourself a huge disservice.
07:53	cark	there are plenty good reasons to do that kind of stuff
07:53	somnium	one nice thing about js is you get interface injection for free
07:54	somnium	and functions carry their source around with them
07:57	somnium	my motivation is practical though, macros for dojo/mootools, map-reduce fns for couch or mongo, in clojure
07:58	cark	well js has good support for literals, in some respect that's quite the same as macros
07:58	cemerick	somnium: there is a clojure view server for couchdb out there, FYI
07:59	cemerick	somnium: http://github.com/tashafa/clutch/
08:17	somnium	cemerick: im not very familiar with couchdb, are views defined in json or javascript? (ie. can you execute arbitrary js in a view?)
08:17	cemerick	somnium: views are defined by code stored in JSON
08:17	cemerick	by default, that code is js
08:18	cemerick	erlang is now supported, though I'm not familiar with the mechanism there
08:18	cemerick	clojure views are clojure code, stored in json, and executed by a clojure view server running next to couchdb
08:23	somnium	cemerick: is the range of fns that the clojure view server can store/execute limited at all?
08:23	cemerick	nope
08:23	cemerick	the code for any view is just a string, so it's entirely up to the view server to determine the semantics of that
08:24	cemerick	obviously a concern if you're running an "open" couch, but hardly anyone does
08:28	G0SUB	I have an abstract class which is later implemented in some other place. I am facing some problem with type hints. If I put the type of the abs. class, it does not work.
08:28	G0SUB	what is the way to deal with this?
08:28	somnium	cemerick: can the view server translate any clojure fn to js I mean?
08:29	cemerick	somnium: translate a clojure fn to js?
08:29	cemerick	no, surely no
08:29	cemerick	not
08:29	cemerick	what would the point of that be?
08:29	G0SUB	somnium: this is wrt MongoDB. I need to put a type hint while using the .findOne method which is specified in DBCollection, but is implemented in DBAPILayer.
08:30	somnium	I mostly use mongodb, you can write store js to map-reduce over a collection, aggregate and gen a new one
08:30	somnium	like a stored procedure in sql
08:30	somnium	writing those in clojure is one of my motivations
08:32	somnium	G0SUB: #^DBCollection still gens a warning?
08:32	G0SUB	somnium: no, but then, the function returns nil.
08:32	G0SUB	somnium: I remove the type hint, and it works.
08:36	chouser	that's pretty odd.
08:36	somnium	G0SUB: in congo it appears #^DBCollection is working
08:36	G0SUB	somnium: line number? I will check.
08:37	somnium	71
08:37	somnium	er
08:37	somnium	wait thats in new branch
08:38	G0SUB	somnium: I see that. what about findOne ?
08:38	somnium	G0SUB: in the java driver find-one calls find and returns the first item
08:39	somnium	so its the same method essentially
08:40	G0SUB	somnium: in my case, find works well with the type hint. only find-one is failing.
08:42	somnium	G0SUB: can you paste the code? the javadocs would seem to indicate #^DBCollection is the right type
08:43	G0SUB	somnium: just a sec.
08:56	lisppaste8	G0SUB pasted "Type hint issue" at http://paste.lisp.org/display/92419
08:56	G0SUB	somnium: ^^^
08:57	ksm	How can I do some sort of list unpacking so I can send it as parameters to a function? such as: (def parameters ["value1" "value"]) (defn function [var1 var2] ...) (function (unpack parameters))
08:57	ksm	In python I would do *parameters
08:57	chouser	ksm: apply
08:57	ksm	thanks
08:58	chouser	(apply function parameters)
08:58	ksm	Neat, thanks
08:58	chouser	G0SUB: are there javadocs for that DBCollection class online?
08:58	somnium	G0SUB: what do you get if you print the type of the collection in both? #<MyCollection> ?
08:59	G0SUB	chouser: http://api.mongodb.org/java/1.1/com/mongodb/DBCollection.html
08:59	G0SUB	somnium: yeah, DBAPILayer$MyCollection
09:00	G0SUB	hmm, odd. it's working now.
09:00	somnium	does #^DBAPILayer$MyCollection work on *find-one?
09:00	G0SUB	don't know what's wrong
09:00	G0SUB	somnium: will check.
09:00	chouser	there are two .findOne methods that take a single arg
09:01	G0SUB	chouser: how do I give it a proper type hint?
09:01	chouser	What happens if you hint the arg as well?
09:02	somnium	you can put #^DBObject on query
09:02	chouser	#^DBObject (ClojureDBObject. query)
09:02	chouser	Doesn't seem like it should be necessary, but might be interesting to try.
09:02	G0SUB	hmm
09:04	G0SUB	chouser, somnium: working now.
09:05	chouser	huh.
09:06	chouser	The compiler ought to be able to see that (ClojureDBObject. foo) returns a DBObject. I wonder why the hint is necessary.
09:07	chouser	G0SUB: What version of Clojure?
09:07	G0SUB	"1.1.0-alpha-SNAPSHOT"
09:08	somnium	I wonder if it sends as Object rather than interface
09:08	somnium	though, that should throw an exception somewhere
09:09	G0SUB	somnium: another help, I call .toClojure on the return value of an update operation. that generates a reflection warning. what type hint should I give it?
09:09	chouser	I think it's calling the other .findOne method.
09:09	somnium	G0SUB: Im clueless, but might be worth running some unit tests with different hints to see if its repeatable
09:10	chouser	G0SUB: those reflection warnings aren't dangerous
09:10	G0SUB	somnium: OK.
09:10	somnium	for .toClojure hint as #^ClojureDBObject
09:11	chouser	G0SUB: I wouldn't bother adding type hints unless you need them for some specific reason.
09:11	G0SUB	somnium: well, I tried that hint, didn't work.
09:11	G0SUB	chouser: mainly for perf.
09:12	G0SUB	chouser: that code is very heavily used...
09:12	chouser	ok
09:13	rhickey	where is ClojureDBObject defined?
09:14	chouser	http://tinyurl.com/yd978tf/src/main/java/somnium/congomongo/ClojureDBObject.java
09:14	somnium	G0SUB: the new branch uses protocols, it should be faster if youre doing a lot of fetch-ones
09:15	G0SUB	somnium: cool. I anyway don't use congomongo. I just took the ClojureDBObject idea from it. everything else is custom :)
09:15	chouser	rhickey: BasicDBObject: http://api.mongodb.org/java/1.1/com/mongodb/BasicDBObject.html
09:15	somnium	ah, no worries then :)
09:16	chouser	G0SUB: oh! your ClojureDBObject is not the one I just linked to?
09:16	somnium	huh, they changed it to a linked hash-map
09:16	G0SUB	chouser: it is. I just use that file.
09:17	G0SUB	somnium: changed what?
09:19	somnium	G0SUB: BasicDBObject used to subclass HashMap
09:19	somnium	bit-rot sets in fast
09:20	G0SUB	somnium: hmm, so is that a breaking change?
09:20	chouser	this is actually the one part of cinc that worries me a little -- getting the (newly modular) host method resolution at compile and runtime right.
09:21	somnium	G0SUB: it shouldnt break anything, DBObject is the key interface
09:21	G0SUB	somnium: do we need to modify ClojureDBObject in any way?
09:23	somnium	G0SUB: no I guess not, just need to recompile it
09:23	G0SUB	somnium: OK
09:23	somnium	can drop two deprecated methods too
09:30	G0SUB	somnium: which ones?
09:32	somnium	G0SUB: scratch that, I was thinking of a clojure implementation, the java class doesnt have to do anything
09:32	G0SUB	somnium: hmm, a Clojure implementation should be cool.
09:33	G0SUB	interesting clojure project -> http://code.google.com/p/clojure-textflow/source/browse/src/textflow.clj
09:56	maacl	Anyone?
09:57	somnium	maacl: does elpa indicate slime is installed?
09:57	maacl	yes
10:00	maacl	somnium: yes
10:00	callen	how much pain am I in for if I come from lisp but not java to clojure?
10:00	somnium	I think there was an issue with the download failing and not getting cleaned up
10:00	Chousuke	callen: well, you'll need to learn how to read the java docs at least.
10:01	callen	Chousuke: reading documentation is fine, I know I'd have to leverage existing java functions.
10:01	chouser	callen: classpath is likely to generate some angst
10:01	rhickey	chouser: what did you mean by " getting the (newly modular) host method resolution at compile and runtime right" ?
10:02	callen	chouser: how so?
10:02	chouser	callen: but you'll get over both of those, and then you'll be golden.
10:02	maacl	somnium: ok, any way I can do it manually?
10:02	Chousuke	callen: the classpath causes headaches for everyone. :)
10:02	somnium	maacl: technomancy is the one to ask :)
10:03	callen	Chousuke: could you explain what it is exactly that you're talking aboot though?
10:03	maacl	somnium: yeah, he doesn't seem to be around :-)
10:03	callen	Chousuke: the path to the libs?
10:04	somnium	maacl: I got it to work by putting in a blank .emacs and making sure it asked me if I want to install clojure, but that was just shotgun debugging (I had a previous version installed)
10:04	liebke	maacl: can blow away your ~/.emacs.d/elpa/ directory, and start over?
10:04	maacl	liebke: yeah, about to try that
10:05	chouser	rhickey: It seems to me that the method resolution algorithm in Compiler.java embodies a complex set of rules that will be rewritten for cinc. That seems more likely than any other area I can think of to appear to be correct and working only to fail later in odd corner cases.
10:05	Chousuke	callen: paths, but yeah.
10:05	chouser	rhickey: amounting to regressions in currently working code.
10:05	rhickey	chouser: hopefully there will be a cross-language thingy for this coming from the JVM language community
10:05	Chousuke	callen: the classpath tells the JVM where to find classes
10:05	callen	Chousuke: interesting. also I feel like I've seen you around elsewhere before. Where else on fnode do you chill?
10:06	chouser	rhickey: hm, that's interesting. I think I read a thread you and headius were talking about related to that.
10:06	Chousuke	callen: ubuntu channels mostly
10:06	rhickey	I don't understand the DBObject problem from before, what do I need to reproduce?
10:07	callen	Chousuke: that'd be it then.
10:07	devlinsf	rhickey: Good Morning. I was working on assembla ticket 128 w/ chouser . Could you take a look at the patch? I was hoping to make 1.1
10:08	callen	really comforting to see the creator in the IRC channel, yeesh.
10:08	callen	can't count beyond one hand how often I see that.
10:08	devlinsf	rhickey: doesn't have to be now, but before 1.1
10:08	Raynes	callen: That shocked me as well the first time I came here.
10:09	chouser	somnium: did you try to reproduce the DBObject problem?
10:10	rhickey	devlinsf: functional patches added now won't have seen any usage prior to release - that's not good
10:10	callen	Raynes: It resembles how the mumble team works, and is thusly very encouraging for me.
10:10	devlinsf	rhickey: It's a bug fix
10:10	somnium	chouser: no, my repl is tied up
10:10	somnium	G0SUB: ping
10:10	rhickey	devlinsf: doesn't matter, it's function, not docs etc
10:12	devlinsf	rhickey: Hmmm... I'm gonna make one last appeal. Isn't one of the major points of an RC to find& fix bugs?
10:12	rhickey	people could work on any number of bugs and want to squeeze them in, at some point there needs to be a cutoff. Producing the RC was that point. I wouldn't hold up for other than a broadly impactful bug
10:12	devlinsf	rhickey: Okay, so it'll have to wait for 1.2?
10:13	rhickey	or 1.1.x patches
10:13	devlinsf	Alright
10:14	devlinsf	rhickey: I'll bring this up again after the release. Got to go now
10:14	rhickey	We haven't gotten much feedback on RC1 - I presume that's good ?
10:15	devlinsf	Transients & Sorted Collections
10:15	devlinsf	The don't work
10:15	pjb3	Anyone doing web dev with clojure? What are you using for HTML templating?
10:15	devlinsf	The concensus seems to be wait until 1.2
10:16	devlinsf	http://groups.google.com/group/clojure/browse_thread/thread/1281d1f7bfc6c5fa#
10:16	rhickey	devlinsf: not all collections do or will support transients, linked lists and binary trees (the sorteds) won't see much benefit vs the high-branching-factor collections
10:17	chouser	that sounds remarkably like what the transients doc page says. :-)
10:17	chouser	oh, but I should add hash-set there.
10:17	devlinsf	Right
10:17	devlinsf	Can we add a line explictly stating sorted don't work
10:18	devlinsf	gtg, really
10:24	maacl	pjb3 enlive
10:24	pjb3	maacl: yeah, I was just reading about that
10:24	karmazilla	maybe 'transient (and 'persistent!) could just return its input collection if it is not IEditableCollection?
10:25	chouser	karmazilla: no, their largely incompatible objects.
10:26	chouser	you ought to be able to assume you can conj! on whatever transient returns
10:26	pjb3	maacl: the readme is a little dense, do you know of any simple examples?
10:27	somnium	pjb3: have you tried clout? (formerly known as compojure.html)
10:28	karmazilla	chouser: ok, maybe I'm not updated. Had just understood that the interface was suppose to be the same.
10:28	maacl	pjb3: http://wiki.github.com/cgrand/enlive/getting-started
10:28	maacl	pjb3: Or was that what you were reading?
10:28	pjb3	maacl: thanks
10:29	pjb3	somnium: you mean hiccup?
10:29	somnium	thats it
10:29	pjb3	I'm not a fan of s-exps for html
10:29	pjb3	I like the idea of enlive much better
10:29	chouser	karmazilla: the interface is meant to allow similarly-structured code, but the actual fns supported are mostly different
10:29	karmazilla	chouser: is there a predicate to check for IEditableCollection?
10:29	pjb3	I'd rather write the html in html and have some method to transform the dynamic parts of it
10:30	somnium	pjb3: enlive is a good bet then
10:30	pjb3	somnium: yeah, I'm gonna read the docs on enlive
10:31	chouser	karmazilla: just (instance? IEditableCollection x) I think.
10:31	the-kenny	Is there something like gwt for clojure? (I know compojure, but I'm tired of writing html and css...)
10:32	the-kenny	or... something like a wrapper for gwt
10:33	chouser	the-kenny: there is a thing that's a bit gwt-like, but it has a somewhat generic name that I can never remember...
10:34	the-kenny	chouser: Yes.. I remember I've seen something which sounded like how I want it on github.. but I can't find it anymore
10:34	chouser	ah, there it is. http://www.webtoolkit.eu/jwt
10:35	chouser	the-kenny: http://www.nsa.be/index.php/eng/Blog/From-OO-to-Lisp-style-structuring-my-Clojure-Jwt-app
10:36	the-kenny	chouser: ah, thank you :) I'll have a look at it
10:36	karmazilla	chouser: editable? (or something) might be a nice addition to the set of predicates, looking at what is already there
10:49	fliebel	what is the difference between send and send-off? The documentation says they both return immediately, but one should be used for blocking actions.
10:50	danm_	send-off sends to a thread pool which can grow I believe
10:51	chouser	send is only for actions that are CPU-bound, not blocking on disk or network
10:52	fliebel	hmmm, okay...
10:53	danm_	if you send a bunch of blocking actions, you can end up needlessly blocking other agents
10:53	danm_	so you use send-off for those
10:54	bOR_	so with send-off, the functions send to the agent do not neccesarily happen in the same order that they are send to the agent?
10:54	chouser	bOR_: no, actions always happen in the order they were sent, regardless of send vs. send-off
10:55	johnmn3	So I'm working on hiredman's simple code editor.. making it a repl. There's one function which is a keyListener for the enter key, which gets the text from the text field and puts in an ArrayBlockingQueue. the other repl function, start-repl-thread runs clojure.main/repl with configuration options, specifically: :read (... ((binding [in (.take transfer-q)]... so, my problem is...
10:56	the-kenny	bOR_: send-off uses a different thread-pool in the background.
10:56	bOR_	ah, getting it.
10:56	the-kenny	send-off is for functions with io-blocking, send for "normal" functions.
10:56	johnmn3	I want to be able to write: (defn hi [name] <enter> and not have it evaluated yet.. I already have a good function to count parens.. I just need to use some intermediate structure to hold on to the string until we have a balanced form, and then send it to read
10:57	johnmn3	how might one set that up?
10:59	johnmn3	I'm guessing some kind of function that holds a queue or buffer and keeps on filling up until it reads a matching closing parens, then it flushes the queue to the repl's read.
11:00	fliebel	Why would you ever use 'do'?
11:00	johnmn3	fliebel: to do two things where only one thing is expected
11:00	fliebel	hmmm...
11:00	chouser	fliebel: for side effects
11:01	johnmn3	,(if true (do (pr "hi") (pr "still here")) (pr "not evaluated))
11:01	clojurebot	EOF while reading string
11:01	johnmn3	er
11:01	johnmn3	,(if true (do (pr "hi") (pr "still here")) (pr "not evaluated)))
11:01	clojurebot	EOF while reading string
11:01	johnmn3	err
11:02	johnmn3	ah
11:02	fliebel	,(if true (do (pr "hi") (pr "still here") (pr "not evaluated))
11:02	clojurebot	EOF while reading string
11:02	johnmn3	,(if true (do (pr "hi") (pr "still here")) (pr "not evaluated"))
11:02	clojurebot	"hi""still here"
11:02	fliebel	finally :P
11:02	johnmn3	missing that last quote
11:02	fliebel	I see
11:02	johnmn3	where if was expecting one thing to happen
11:03	johnmn3	but like chouser said, people should mostly use it when they're are intentionally trying to induce side effects.
11:04	johnmn3	(if true (do (functional thing) (pr "this string is a side effect (for logging perhaps)")) (not evaled))
11:04	fliebel	so more like… (if (do (prn "hallo") true) (prn "world") (prn "not evaluated"))
11:05	ksm	,(do true false)
11:05	clojurebot	false
11:05	johnmn3	right
11:23	fliebel	Is there a less resource intensive way then a regex to determine if a string contains, starts with or ends with a certain string?
11:23	stuartsierra	,(.contains "foobar" "ooba")
11:23	clojurebot	true
11:23	chouser	,(.startsWith "foobar" "foo")
11:23	clojurebot	true
11:24	fliebel	ah, java things....
11:24	chouser	embrace java. :-)
11:24	fliebel	thanks… I was diging in the Clojure api
11:24	chouser	you're already suffering under any pain that Java is bringing. Might as well enjoy the benefits as well.
11:24	fliebel	:)
11:30	maacl	why would i get clojure-slime-config: Required feature `slime-autoloads' was not provided when trying to load swank-clojure?
11:30	fliebel	Benefit is relative in this case… Most languages offer their own solution that is documented in one place. But I must agree that Java also brings some power into Clojure.
11:44	chouser	fliebel: right. most languages reimplement strings from scratch. Clojure hasn't bothered.
11:47	johnmn3	I forgot. How do you do a counter with a closure?
11:48	chouser	,(-> 2 inc inc inc)
11:48	clojurebot	5
11:48	johnmn3	this doesn't work: (defn ab [] (fn [x] (let [a (atom 0)] (if (= x :inc) (swap! a inc) (swap! a dec)) @a)))
11:48	johnmn3	I mean where the function holds state for you.
11:49	chouser	:-(
11:49	chouser	why would you want that?
11:50	johnmn3	I'm trying to build a buffer that I can keep adding strings too.. "(asdf (asdf" until it gets a closing paren that matches the first open peren. then it returns the whole string.
11:50	lpetit	and ?
11:50	johnmn3	so I have to count parens up and down
11:50	lpetit	so count them, and pass the value in your loop
11:51	reify	johnmn3: move the let outside fn, like: (defn ab [] (let [a (atom 0)] (fn [x] (if (= x :inc) (swap! a inc) (swap! a dec)) @a)))
11:51	chouser	johnmn3: I'd recommend building a seq of strings, then passing to (apply str ...) to build your final string.
11:52	johnmn3	that's what I did. but I want to be able to make multiple calls to the function, because the user will be passing a whole function in, lines at a time, and I need a place to store those lines until the whole form is written
11:52	johnmn3	reify: ah
11:53	johnmn3	So, should I make a function that takes one char at a time?
11:54	lpetit	johnmn3: why not make this counter part of the intermediate state, having the function return {:buffer the-buffer, :open-parens the-count-of-open-parens]. And either publish the keys as part of your API, or provide helper functions to relieve the user from knowing the exact datastructure you use for holding the state ?
11:54	johnmn3	and feed each string into it with a (doseq [a argstring] (myfn a)) ?
11:55	lpetit	what does it mean to "Mark" a portion of text in emacs ?
11:55	lpetit	and how to do this ?
11:56	fliebel	Ok, I got another should-be-simple-but-can't-figure-it-out issue. I got an agent to which I want to "append" items, or set to a new list with more items, if that sounds better to you. I can't figure out how to do that… send seems to be a threaded apply or something like that.
11:56	johnmn3	then keep conjing the chars to a new string and, when a closing paren matches the opening one, flush and keep going
11:56	KirinDave	lpetit: You might wanna read the emacs programmer's guide.
11:56	KirinDave	lpetit: It's in every copy of emacs.
11:56	stuartsierra	lpetit: "mark" is emacs's version of select-copy-and-paste
11:57	lpetit	KirinDave: just trying to (very) quickly follow Phil's paredit-outline tutorial without having to dig into the rather comprehensive and big emacs doc
11:57	lpetit	http://p.hagelb.org/paredit-outline
11:57	chouser	,(let [a (agent [:foo])] (send a conj :bar) (await a) @a)
11:57	clojurebot	[:foo :bar]
11:58	chouser	johnmn3: like that?
11:58	KirinDave	lpetit: So, the thing is that emacs has a VERY different way of handling text. You literally keep using editing commands.
11:58	lpetit	stuartsierra: ok, would you mind give me the magical command combination ? :-)
11:58	fliebel	chouser: thanks… I was close, but not close enough, I was making conj an anonymous function.
11:59	lpetit	stuartsierra: and I promise not to RTFM you if you ever have a question on Eclipse :-)
11:59	fanatico	lpetit: ctrl-space
11:59	lpetit	fanatico: thx
11:59	stuartsierra	lpetit: Ctrl+Space sets the mark. Ctrl+w cuts between "point" (the cursor) and the mark. Alt+w copies between point and mark.
12:00	fanatico	and all region-specific commands work on the mark.
12:00	johnmn3	chouser: let me take that and explore it for a minute
12:00	lpetit	stuartsierra: but Phil states in his doc "Mark "declare rooms"" , so I guess mark can be placed on text selections as well as text offsets
12:01	fanatico	but I agree with stuartsierra, most people don't have the intuition for how emacs handles text (especially the kill ring), so it's best to read that section of the manual.
12:01	stuartsierra	lpetit: The mark is ONE position, not a selection. The area between the mark and the point defines the selection.
12:01	chouser	johnmn3: oh, sorry, I meant to address that to fliebel
12:01	johnmn3	oh
12:01	lpetit	stuartsierra: ok, so that's weird:
12:01	technomancy	"mark" as a verb means "make the region contain"
12:02	johnmn3	well, it applies anyway ;)
12:02	lpetit	open rooms.clj
12:02	lpetit	point at right before mire.rooms
12:02	lpetit	M-(
12:02	lpetit	Mark "declare rooms"
12:02	lpetit	Press (
12:02	stuartsierra	I dunno what that's about.
12:02	lpetit	I suppose Mark "declare rooms" means mark "d", go to after s, and then press (
12:02	lpetit	it's a tutorial showing the benefits of paredit
12:03	lpetit	http://p.hagelb.org/paredit-outline , line starting with ## 4. Wrangling (depth-changing)
12:03	fanatico	the point over the cursor isn't included.
12:06	fanatico	unless the mark is to the right of the cursor.
12:08	maacl	technomancy: think you could help me with a swank-clojure issue?
12:12	leafw	clojure new branch fails to compile, but ant repots "build successful". IS this normal? The exception claims that clojure.lang.RestFn is not found
12:12	leafw	Caused by: java.lang.NoSuchMethodError: clojure.lang.RestFn.<init>(I)V
12:12	probablyCorey	Would a clojure shell, like bash or zsh be a horrible idea?
12:12	leafw	actually, that is clojure.contrib ... oops, sorry
12:13	the-kenny	leafw: Do you mean clojure or clojure.contrib?
12:13	somnium	~RestFn
12:13	clojurebot	excusez-moi
12:13	chouser	leafw: do "ant clean" in contrib before building
12:13	KirinDave	Well, besides the inability to change cwd, there is nothing wrong with it, probablyCorey.
12:13	the-kenny	~clojure.lang.RestFn
12:13	clojurebot	ant clean and rebuild contrib
12:13	KirinDave	scsh is pretty popular.
12:13	leafw	I was mixing clj new branch with clj contrib master branch.
12:13	chouser	leafw: that's ok
12:13	probablyCorey	KirinDave: Cool, I'll take a look at scsh
12:18	leafw	clojure-contrib.jar fails to build ... with new branch on both repos
12:18	leafw	java.lang.RuntimeException: java.lang.ClassCastException: clojure.lang.PersistentVector cannot be cast to clojure.lang.Symbol (jmx.clj:89)
12:19	leafw	also lots of name conflicts: java.lang.Exception: Name conflict, can't def bytes because namespace: clojure.contrib.http.agent refers to:#'clojure.core/bytes (agent.clj:252)
12:24	chouser	leafw: use master branch of contrib
12:25	leafw	chouser: new branch of clj and master of contrib
12:25	leafw	ok
12:25	chouser	yeah. :-/
12:25	leafw	makes total sense xD
12:25	leafw	(i.e. it doesn't)
12:26	chouser	well, clojure 'new' is almost entirely backward compatible with clojure 'master', so contrib 'master' works fine with it.
12:26	chouser	contrib 'new' is where some people have started playing with clojure 'new's new features, but obviously they're not keeping up with it.
12:27	chouser	hm, I wonder if a merge from contrib 'new' would fix it.
12:27	leafw	it did build, clj new and contrib master. Thanks.
12:27	technomancy	probablyCorey: IMHO eshell is another excellent shell written in lisp; I use it nearly exclusively
12:27	leafw	by the way what is the timeline for clj new to become master? 1 month? 6 ? a year?
12:29	chouser	leafw: I think we're still hoping for 1.1.0 final to be out by christmas. that would free up master to take on the 'new' patches.
12:29	chouser	not sure how fast rhickey will pull the trigger on that, but my guess would be a few weeks at most.
12:30	leafw	cool
12:30	leafw	that's good news. Thanks.
12:31	chouser	it's not so much news as a guess, so I hope I'm not off by too much. :-/
12:32	leafw	well: you just brougth it to the low end of the spectrum I threw (1 motnh to a year)
12:32	lpetit	stuartsierra: ok, certainly the text should have read : "select" instead of mark
12:33	technomancy	lpetit: they are synonyms
12:33	lpetit	technomancy: gaining knowledge of paredit by reading through your excellent paredit-outline intro text
12:33	lpetit	technomancy: well, that's not what I was told by Stu
12:34	lpetit	Must leave, there's a "toast" here at work :)
12:34	technomancy	lpetit: great! but I wrote that actually for notes to myself about how I should record a screencast. so it uses terminology that may not be appropriate for people getting started with Emacs. =)
12:34	lpetit	technomancy: ok, good to know that :)
12:35	lpetit	technomancy: how complete on the subject of paredit do you think the tuto is ?
12:35	technomancy	lpetit: it's everything I use, but there's still maybe 15% of paredit that it doesn't cover
12:35	lpetit	technomancy: and what would you suggest me as a tutorial for gaining even more knowledge ?
12:36	technomancy	the obscure bits that I haven't internalized myself
12:36	technomancy	lpetit: unfortunately there's not much apart from reading the source. that's why I started the project.
12:36	technomancy	(the project to create a screencast on it)
12:36	lpetit	technomancy: if there's just ~ 15% left, seems ok to me to hope have this rather sooner than later in ccw, then
12:55	jneira	hi people
13:09	technomancy	did a de-chunking mechanism ever get checked in?
13:09	technomancy	ISTR mentions of an experimental way go back to 1-at-a-time seqs
13:16	Chousuke	technomancy: not as far as I know :/
13:16	technomancy	yeah, looking back at it there are only mentions that such a mechanism is being discussed
13:17	Chousuke	I suppose a simple lazy-seq doing rests on the chunked seq is not enough? :/
13:18	technomancy	Chousuke: nope: (def x (seq (rest println (range 39))))
13:18	technomancy	that prints 0 - 31
13:18	technomancy	oops
13:18	technomancy	I meant (def x (rest (map println (range 39))))
13:22	lpetit	,(def x (rest (map println (range 39))))
13:22	clojurebot	DENIED
13:23	technomancy	the def is just to prevent the repl from printing it; you could wrap it in a do or some such
13:25	lpetit	my bad
13:25	lpetit	,(do x (rest (map println (range 39))))
13:25	clojurebot	java.lang.Exception: Unable to resolve symbol: x in this context
13:25	lpetit	I'm really a stupid guy sometimes
13:26	lpetit	,(do (rest (map println (range 39))) nil)
13:26	clojurebot	0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
13:27	fliebel	I just wrote some Clojure code… with a purpose and some threading for the first time :D ( defn ) http://github.com/pepijndevos/utterson/blob/master/src/utterson/core.clj
13:38	pdk	what would be the equivalent of doto for a class instead of an instance so you could call static methods in the doto block
13:45	stuartsierra	,(.parseInt Integer "123")
13:45	clojurebot	java.lang.IllegalArgumentException: No matching method found: parseInt for class java.lang.Class
13:46	Chousuke	hm
13:46	Chousuke	can't you do (doto someobj Integer/parseInt)?
13:51	hiredman_	Chousuke: that would return someobj and be rather pointless
13:51	hiredman_	~macros
13:51	clojurebot	Barking spiders!
13:53	chouser	pdk: I think you'd have to write your own macro
13:53	pdk	hm
13:55	chouser	(defmacro dostatic [cls & ms] (cons `do (map #(list* `. cls %) ms)))
13:56	fliebel	How can I get a regex to just evalulate to true or false? All of them seem to return lists and java stuff...
13:57	chouser	fliebel: re-find will return nil when it can't find a match
13:57	chouser	re-matches too, actually
13:57	fliebel	chouser: thanks, but that is nox exactly clear form the docs
13:57	chouser	and also re-seq, apparently. So take your pick! :-)
13:59	polypus	anybody know if google web toolkit is useable from clojure? i'm not really sure how GWT works internally
13:59	fliebel	chouser: which is the most efficient?
14:00	chouser	polypus: I've not heard of anyone using it successfully.
14:00	chouser	polypus: I believe the to-js compiler in GWT reads Java sources
14:01	cark	hum, wasnt it byte code
14:01	cark	?
14:01	polypus	huh, so direct java -> js. too bad it's not bytecode. even better would have been to just have the java generate js
14:01	chouser	fliebel: probably re-matches, though it requires the regex match at the start of the string.
14:01	cark	chouser : ah looks like you're right =(
14:01	chouser	cark: you may be thinking of android
14:02	fliebel	thanks
14:02	polypus	chouser: anyways, thx
14:02	cark	chouser : right
14:02	chouser	fliebel: doubt it'll make much difference for performance. re-find or re-matches seem clearest in their meaning if you're just treating it as true/false.
14:03	fliebel	chouser: I'm going to stick with re-find, since I reorganized my code to evaluate it only once.
14:04	chouser	ok
14:06	polypus	maybe it's possible to use GWT if only ever in development mode: "When an application is running in development mode, the Java Virtual Machine (JVM) is actually executing the application code as compiled Java bytecode, using GWT plumbing to connect to a browser window"
14:07	stuartsierra	it still uses Java source I think
14:08	chouser	polypus: I've wondered the same, though I didn't persue it.
14:08	polypus	stuartsierra: what, performs some extra analysis ontop of the bytecode compile?
14:08	chouser	polypus: Also wondered about a clojure-to-java-source compiler.
14:09	stuartsierra	polypus: No, I think GWT has its own compiler that reads Java sources.
14:09	polypus	chouser: yeah the latter crossed my mind too
14:09	chouser	but actually, when I played with it, the problem was Clojure's collection types use all kinds of unsupported (by gwt) classes.
14:09	chouser	gwt won't produce js for Clojure's own .java code without quite a bit of hacking at it.
14:11	chouser	stuartsierra: I learned JavaScript already, I just want macros.
14:11	hiredman	people want to leverage existing tools
14:11	chouser	and immutables locals and persistent collections.
14:11	hiredman	chouser: really? I just want parens
14:11	chouser	and the seq abstractions
14:11	polypus	stuartsierra: i actually don't mind javascript. but just trying to pick a toolkit and gwt has a nice crossbrowser canvas implementation
14:11	cemerick	I thoroughly dislike the entire web development experience, and js + browser dom is a big part of that distaste
14:11	polypus	which i need
14:12	chouser	js rivals ruby for my second-favorite language.
14:12	chouser	I wonder how hard it would be to compile js-like syntax into clojure.
14:12	polypus	chouser: ditto
14:13	hiredman	chouser: :(
14:13	chouser	hiredman: teehee!
14:14	cemerick	syntax is not the problem, browser quirks are IMO
14:14	chouser	browser quirks are an issue, though js itself has surprisingly few. The DOM makes up for it.
14:14	hiredman	I'm waiting for browsers to be rewritten in js
14:14	cemerick	right, the dom side of things is what I was referring to
14:15	chouser	indeed. And I tried js without the DOM once (thanks, Yegge). Turns out that's worse.
14:15	hiredman	really a browser could just be something like a smalltalk image
14:15	polypus	i'm waiting for browsers to standardize ontop of some VM so we can use whatever language we want.
14:15	stuartsierra	polypus: dream on
14:15	chouser	polypus: I think that's what hiredman just said.
14:15	hiredman	polypus: they did, it's called js
14:15	polypus	i expect to be dead before that happens though :)
14:16	hiredman	js is the browser bytecode
14:16	hiredman	which is why you have things like gwt
14:17	polypus	anybody ever write a VM in javascript?
14:18	hiredman	http://ejohn.org/blog/ruby-vm-in-javascript/
14:18	hiredman	etc, etc, etc
14:20	chouser	Nintendo emulator in pure javascript: http://benfirshman.com/projects/jsnes/
14:23	hiredman	http://mxr.mozilla.org/mozilla/source/js/narcissus/
14:23	hiredman	:D
14:24	jneira	user experience emulator in bastard javascript: http://www.microsoft.com/windows/ie/ie6/default.mspx
14:25	polypus	narcissus must be a speed demon :)
14:26	hiredman	:P
14:26	hiredman	modern js engines are fast
14:26	chouser	compared to? non-modern js engines?
14:26	hiredman	yes
14:27	hiredman	like last years :P
14:27	chouser	:-)
14:28	hiredman	so you just need a compiler to lower js to native code, and a nice ffi
14:29	fliebel	lol, jsnes is cool :P
14:30	alexyk	anybody here used clojure-twitter?
14:31	hiredman	huh, apparently yeggie's js in elisp is based in part on narcissus
14:32	fliebel	alexyk: tell me about it, I love writing little twitter apps :)
14:33	alexyk	fliebel: I'm just looking for tips on making it work
14:34	fliebel	alexyk: looks quite straightforward, what's wrong?
14:35	alexyk	fliebel: getting oauth going -- but I think I found that I should register a separate "client" for it
14:36	fliebel	alexyk: yea, you need to register your client at twitter before you can do oauth, they'll give you a token to work with.
14:36	alexyk	fliebel: it has "use Twitter for login," do I need that?
14:37	fliebel	wait a moment
14:37	fliebel	http://twitter.com/oauth_clients/new
14:38	fliebel	yes, you want to use twitter for login
14:38	fliebel	oh, wait
14:38	alexyk	fliebel: right, it has a checkbox saying "use Twtter for login," I wonder what that means
14:39	fliebel	no, i don;t think so...
14:39	alexyk	if oauth's idea is to be able to login without twitter user/password
14:39	fliebel	alexyk; do you know disqus?
14:39	alexyk	fliebel: yep
14:39	fliebel	alexyk: they let you login using twitter, if you are going to do that, check that box, else don't
14:41	fliebel	alexyk: after you registered you can go to http://twitter.com/oauth_clients/ and click your application to see the secret and key
14:42	alexyk	yeah, I see that... what I'm not clear about is how can I post to my account if I don't provide user/pwd
14:42	fliebel	alexyk; that is the whole point of oauth
14:43	alexyk	ok, lemme see
14:43	fliebel	alexyk: your application is going to ask twitter for access using a token, of you accept that, it gets another token, now every request with that token is known to be from a trusted source and is allowed access to your account.
14:44	fliebel	alexyk: I know it's a lot of black magic when you first look at it. I managed to get it working once, but for my own little apps it's more convenient to use basic auth…
14:45	alexyk	fliebel: yeah, well clojure-twitter apparently has only oauth
14:45	alexyk	and twitter threatens to stop basic auth
14:48	fliebel	alexyk: Got to run, success!
14:48	alexyk	thx!
14:55	lisppaste8	johnmn3 pasted "poor man's s-expr parser" at http://paste.lisp.org/display/92435
14:55	johnmn3	behold, it's hunchback glory
14:56	johnmn3	it needs some ironing, but it's coming along.
14:58	PrinceOfA	is there a clojure wrapper for com.sun.net.httpserver?
15:00	johnmn3	not sure.. know there is for jetty
15:00	johnmn3	I should probably add an :else (stop ignoring) to that cond
15:01	hiredman	http://github.com/hiredman/clojurebot/blob/master/hiredman/http_server.clj <-- not a wrapper around httpserver
15:05	quizme	how do i get clj to run like this: /opt/clojure-contrib/launchers/bash
15:05	quizme	oopens
15:05	quizme	oops
15:05	quizme	i mean
15:06	quizme	java -cp jline-0.9_0_5.jar:clojure.jar jline.ConsoleRunner clojure.lang.Repl
15:07	hiredman	depending on your version of clojure clojure.lang.Repl might still be around, but it has been deprecated for many months (feels like forever)
15:07	hiredman	use clojure.main
15:08	hiredman	clojurebot: clojure.lang.Repl is <reply> you mean clojure.main
15:08	clojurebot	You don't have to tell me twice.
15:08	hiredman	clojurebot: clojure.lang.Script is <reply> you mean clojure.main
15:08	clojurebot	You don't have to tell me twice.
15:08	johnmn3	also, your my jline version is 0.9.94
15:08	hiredman	"your my"
15:08	johnmn3	erm
15:09	johnmn3	/s/your/""
15:10	johnmn3	hmm. didn't know clojure.main replaced script too.
15:11	quizme	i guess i can just set up an alias clj='java -cp jline-0_9_5.jar:clojure.jar jline.ConsoleRunner clojure.main'
15:11	johnmn3	hiredman: so I've been tweaking your editor. Made it into a repl. just trying to fasten on that parser now.
15:11	hiredman	sure, but what about when you want to add something to your classpath?
15:12	quizme	i dunno ^^;
15:12	quizme	i'm new to this whole java thing
15:13	hiredman	quizme: what java version do you have?
15:13	hiredman	java -version
15:13	quizme	1.6.0_15
15:14	quizme	64 bit
15:14	hiredman	ok, make a directory called .jars in your $HOME
15:14	quizme	k
15:15	hiredman	put something like CLASSPATH="$HOME/.jars/*" in what ever bash uses (.profile? ,bashrc?)
15:15	quizme	.bashrc
15:15	hiredman	either copy any jars you plan to use to ~/.jars/ or make a syslink
15:16	hiredman	start clojure: java jline.ConsoleRunner clojure.main
15:17	quizme	xception in thread "main" java.lang.NoClassDefFoundError: jline/ConsoleRunner
15:17	hiredman	quizme: you are missing the jline jar
15:17	quizme	it's in ~/.clojure
15:18	hiredman	so?
15:18	quizme	oh
15:18	quizme	k lemme try
15:19	johnmn3	definitely doesn't work yet.
15:20	hiredman	johnmn3: what doesn't?
15:20	johnmn3	converted your code editor to a repl. now trying to bolt on multi-line support.. still banging on it.
15:21	hiredman	multiline support? it works for the select case
15:22	johnmn3	naw, I changed it up.. the input is just a JTextField now. I'm basically trying to get what a repl from the command like provides, but in a gui.
15:22	hiredman	that is going to be painful
15:22	johnmn3	yea, it is. once I have basic repl support, I'll try adding features on from there.
15:23	chouser	johnmn3: you didn't happen to look at texture, did you?
15:23	chouser	johnmn3: so far just a swing-based repl. I think the handles multiline input.
15:23	johnmn3	parens highlighting maybe.. after this though, I'm going to implement line history
15:23	chouser	s/the/it/
15:24	quizme	hiredman: thanks it works
15:24	stuartsierra	Is there a good Swing tutorial for use without an IDE?
15:25	johnmn3	chouser: yours, right? yes, I looked at it. unfortunately, it's just a little bit advanced for me. This is more of a learning thing right now.
15:25	johnmn3	I should look at textjure though, and see how it implements multi-line support.
15:26	chouser	johnmn3: yeah, that's my poor, sad, abandoned project. :-P
15:26	johnmn3	I've played with it.. I thought it was pretty cool.
15:27	johnmn3	I might pick it back up once I wrap my head further around these "bindings" and plumbing.
15:42	hiredman	clojurebot: simple setup is http://www.thelastcitadel.com/dirt-simple-clojure
15:42	clojurebot	You don't have to tell me twice.
16:00	drewr	what should the behavior of a failed :pre or :post be if assert is false?
16:01	drewr	both throw a plain j.l.Exception
16:04	stuartsierra	drewr: if assert is false the pre/postconditions are never evaluated
16:05	drewr	is binding sufficient for turning it off?
16:05	drewr	(binding [assert false] (constrained-sqr -1)) => Exception
16:05	drewr	(constrained-sqr from the doc example)
16:06	stuartsierra	I guess not.
16:06	stuartsierra	But it can be set!
16:07	Chousuke	drewr: it's a compile-time decision I think
16:08	Chousuke	since the value of assert is checked by the assert macro
16:09	chouser	Chousuke: that looks right.
16:10	chouser	I guess it makes sense, if the point of assert false is to avoid runtime checks.
16:10	drewr	do you supply is as a jvm param?
16:10	drewr	s/is/it/
16:10	Chousuke	drewr: no, but it's false by default :/ except in the repl.
16:11	drewr	interesting
16:16	lisppaste8	johnmn3 pasted "simple gui repl" at http://paste.lisp.org/display/92442
16:16	johnmn3	it works!
16:17	johnmn3	it needs a lot of cleaning up, but it actually works. (the multi-line support).. I'm sure it breaks under some conditions, but it's a start.
16:19	johnmn3	ah, already found one bug... but it'll have to wait till tomorrow.
16:22	liebke	johnmn3: nice, I'm looking forward to watching your progress. I'd love to have a simple multi-platform gui repl to include with Incanter
16:23	chouser	I'd love to have a simple multi-platform gui repl + clojure source editor to include with clojure
16:25	liebke	better yet
16:26	hiredman	http://i.imgur.com/7bAxb.png <-- this is pretty cool
16:26	hiredman	(for a repl)
16:26	johnmn3	ditto. I think basic repl support as a start is important.. then pump some steroids into the (completion, history, paredit?) and then extend it to do regular code editor like chores.. but keeping the focus on the repl
16:27	lpetit	I'm also interested in adding paredit support, maybe written in clojure
16:27	lpetit	to ccw.
16:27	johnmn3	hiredman: yea, that's the good thing about a gui repl... drScheme can do pictures too
16:28	hiredman	that would be really sweat for incanter
16:28	lpetit	Currently alternatives would be to reuse existing code in existing eclipse plugins for Scheme. Still not sure whether I want to adapt their code, or rewrite the whole stuff in clojure
16:28	johnmn3	lpetit: I've been thinking about what rules would be though.. have to look into how paredit works and decides on things.
16:29	hiredman	I dislike guis, so having a repl that can do gui stuff is nice
16:29	liebke	hiredman: yeah, it's a feature of R (at least on Mac and Windows) that I would really like to have
16:31	lpetit	somebody knows the principles behind paredit. Does it rescan the whole buffer content each time an opening symbol (paren, square bracket) is involved in the typing, or is it more sophisticated ? Does it work well with big files such as core.clj (presumably when editing stuff near the end of the file) ?
16:32	lpetit	johnmn3: paredit command list http://mumble.net/~campbell/emacs/paredit.html
16:32	johnmn3	that's an interesting problem.
16:32	lpetit	clojurebot: paredit commands is http://mumble.net/~campbell/emacs/paredit.html
16:32	clojurebot	Ok.
16:33	lpetit	clojurebot: paredit source is http://mumble.net/~campbell/emacs/paredit.el
16:33	clojurebot	Ik begrijp
16:33	hiredman	http://github.com/hiredman/clojure/blob/readerII/src/clj/clojure/reader.clj
16:35	lpetit	hiredman: is it just a reader for creating clojure datastructures, or does it preserve (maybe optionnally) all the properties of the initial source code (different possibilities for representing metadatas, preserving comments, etc.) ?
16:35	hiredman	lpetit: it is port of LispReader
16:36	johnmn3	from LispReader.java ?
16:36	hiredman	yes
16:37	stuartsierra	paredit was written for Common Lisp / Scheme
16:37	lpetit	hiredman: yes, so it's not tailored for textual editors problems, in that it does not preserve what the user is typing. And I guess that in the case of paredit's problem, it's also somewhat "overkill" since it recognizes structures paredit won't care about (meta, derefs, etc.)
16:37	hiredman	lpetit: there is no reason it could not
16:37	hiredman	and it is written in clojure
16:38	hiredman	so you don't need to know java to add those features
16:38	lpetit	hiredman: indeed, you're thinking about a special mode that could preserve this information
16:38	lpetit	?
16:38	johnmn3	hmm. so something a little more sophisticated than my expression bagger function.
16:38	hiredman	lpetit: I am not thinking about paredit at all
16:39	lpetit	hiredman: maybe we're not talking about the same thing. What do you have in mind ?
16:39	hiredman	I'm just saying, if you are going to write a full reader for clojure, there you go
16:39	hiredman	paredit is must be much simpler
16:39	hiredman	I have't looked at the code, or really even used it that much
16:40	lpetit	hiredman: we agree, paredit must be much simpler since it only has to deal with trees and leaves without knowing too much about the different kinds of leaves
16:40	hiredman	but you just match the types of brackets, and keep brackets balanced for a certain set of operations
16:40	hiredman	paredit doesn't even need to know about trees
16:41	johnmn3	if I were to try to implement some rudimentary paredit features.. would it be heresy to opt for more regular key combinations, rather than emacs like key combos?
16:41	lpetit	hiredman: yes, just a stack of types of brackets + offset
16:41	hiredman	johnmn3: the main thing is balencing
16:41	hiredman	the other stuff is icing
16:41	lpetit	johnmn3: better separate core paredit functionality from the keys used to invoke the commands, no ?
16:42	johnmn3	lpetit: true
16:42	johnmn3	just saying, I tried paredit once and I had a hard time navigating the source
16:43	johnmn3	I think it would make a good learning tool
16:43	lpetit	johnmn3 : look at this http://p.hagelb.org/paredit-outline
16:43	lpetit	clojurebot: paredit tutorial is http://p.hagelb.org/paredit-outline
16:43	clojurebot	c'est bon!
16:43	lpetit	~paredit
16:43	clojurebot	Pardon?
16:43	lpetit	~paredit.*
16:43	clojurebot	Gabh mo leithscéal?
16:44	hiredman	I wonder if I broke something in clojurebot recently
16:44	hiredman	clojurebot: paredit?
16:44	clojurebot	Huh?
16:44	hiredman	clojurebot: paredit foo
16:44	clojurebot	Foo is not Baz.
16:44	hiredman	clojurebot: foo paredit
16:44	clojurebot	foo is is Short, Self Contained, Correct (Compilable), Example http://sscce.org/
16:44	johnmn3	aye, this will be good material
16:44	lpetit	hiredman: was just guessing. Would a regexp be accepted by clojurebot (if not broken ?)
16:45	lpetit	johnmn3: this tuto is made from note made by technomancy for his own use.
16:45	hiredman	lpetit: nope
16:47	hiredman	oh
16:47	lpetit	hiredman: did you compare the performance of your reader with the current java one, just for fun ?
16:47	hiredman	clojurebot: paredit tutorial
16:47	clojurebot	paredit tutorial is http://p.hagelb.org/paredit-outline
16:47	hiredman	lpetit: I did not
16:47	lpetit	hiredman: not tempted ? :-)
16:47	hiredman	I did not do any kind of type hinting or what have you
16:48	hiredman	lpetit: nope
16:48	hiredman	I wrote it, someone else can measure it
16:49	lpetit	please, could a paredit expert tell if, approximately, the following link seems to be a whole reference of paredit commands, or just a summary of the more useful ? http://mumble.net/~campbell/emacs/paredit.html
16:54	johnmn3	that technomancy doc was a good read.
16:55	johnmn3	will keep it bookmarked
16:55	lpetit	I've also put it in clojure bot:
16:55	lpetit	~paredit tuto
16:55	clojurebot	excusez-moi
16:55	lpetit	~paredit tutorial
16:55	clojurebot	paredit tutorial is http://p.hagelb.org/paredit-outline
16:56	johnmn3	btw, if anyone saw my code for the expression bagger, was using all those atoms bad form?
16:57	johnmn3	the function closes over 5 or 6 atoms
16:57	lpetit	johnmn3: can you remember the link ?
16:57	lpetit	s/remember/remind us/
16:57	technomancy	somebody talked to me about recording the paredit script as a screencast
16:58	technomancy	since I realized I don't have the editing skills to pull it off
16:58	johnmn3	http://paste.lisp.org/display/92442
17:00	johnmn3	I know it's not pretty
17:01	johnmn3	I still use a brute-force kind of style
17:01	lpetit	and I'm tired
17:01	johnmn3	aye
17:01	Anniepoo	,(in-ns 'tacos)
17:01	clojurebot	#<Namespace tacos>
17:02	Anniepoo	,(defn foo [] (println "for lunch"))
17:02	clojurebot	DENIED
17:02	lpetit	I don't understand: why do you do all this in order to create a GUI repl ?
17:02	chouser	johnmn3: have you seen any of rhickey's material on values vs. identities?
17:02	Anniepoo	(+ 1 2)
17:02	clojurebot	3
17:02	johnmn3	still learning. you don't have to actually try to understand it.. it's a bit messy.. Just wondering if that's bad form, or acceptable in this situation.
17:03	johnmn3	lpetit: why not?
17:03	lpetit	johnmn3: smells bat
17:03	lpetit	smells bad
17:03	Anniepoo	anybody else had trouble with enclojure and in-ns?
17:03	johnmn3	chouser: yea, I have.. are you getting at something?
17:03	lpetit	johnmn3: but does not come from a thourough analysis of your code. Just a smell
17:03	johnmn3	like i'm using too much state there?
17:03	johnmn3	aye
17:03	chouser	johnmn3: are each of those atoms properly an identity?
17:04	chouser	johnmn3: I haven't read everything you're doing there, but find it unlikley. :-)
17:04	lpetit	not much about state (state has its virtue, and maybe yours is just essential state, I don't know for sure)
17:04	lpetit	but smells like too much mutable storages
17:04	lpetit	as said chouser
17:05	johnmn3	aye.. my original function used a loop, but then I needed a function that was fed it's input one character at a time, which made a loop pointless.
17:05	chouser	guis are mutable state, and it can be hard to prevent that from causing the rest of your code to become infected with mutable state and imperative style. But it's worth the effort.
17:06	lpetit	chouser: beyond Rich's videos (which I've watched "eyes wide open" :-) ), do you know of any online tutorial to help continue "think" about it ?
17:06	chouser	lpetit: not really. it's something I still struggle with, moreso in code related to guis, sockets, and other "external" things like that.
17:07	lpetit	chouser: ok, so we're all in a learning curve here :-)
17:07	johnmn3	and I recured most of the changes back to the top of the loop, not needing any storage.. just used accumulators
17:08	johnmn3	but when being fed only one char at a time, with nothing to loop over on each run, I couldn't figure out an alternative
17:08	chouser	johnmn3: bagger is creating a bunch of state and then mutating it. try imagining all that state as a single value instead. then make a pure function that takes that value (along with any other needed data) and returns a new value.
17:08	lpetit	johnmn3: (fn [_] false) can be written (constantly false), btw
17:08	johnmn3	ah was wondering.. could do that in a lot of places then.
17:09	hiredman	how does clojureql's LIKE operator work?
17:09	chouser	johnmn3: then if you need to, you can store that value in a single reference type. This will keep more of your code "pure functional" and you'll get the best benefits from Clojure's features.
17:10	lpetit	{:newstr (constantly []), :started (constantly false), ...} and pass it in the loop, and use assoc/update/update-in
17:10	johnmn3	maybe I could still use a loop. just have it loop once per run, and just have one slot to hold state between loops.
17:11	lpetit	there's nothing wrong with a loop
17:11	johnmn3	right
17:13	lpetit	from paredit's source code:
17:13	lpetit	;;; This code is written for clarity, not efficiency. It frequently
17:13	lpetit	;;; walks over S-expressions redundantly. If you have problems with
17:13	lpetit	;;; the time it takes to execute some of the commands, let me know, but
17:13	lpetit	;;; first be sure that what you're doing is reasonable: it is
17:13	lpetit	;;; preferable to avoid immense S-expressions in code anyway.
17:13	hiredman	lpetit: pastebin
17:13	lpetit	good, maybe trying to do a raw port of it could be usable.
17:13	lpetit	hiredman: sorry
17:13	johnmn3	so make a map of my data and store it in a ref.
17:14	johnmn3	should be a good exercise
17:15	lpetit	I expect emacs lisp to be readable even if it's the first time I read emacs lisp, or am I fooling myself ?
17:15	lpetit	(I also have some CL knowledge)
17:15	technomancy	lpetit: it's pretty straightforward; yet another imperative lisp dialect
17:16	technomancy	very similar to CL, but more limited. (only dynamic scope)
17:16	lpetit	technomancy: ok, so maybe trying to port it, then write a comprehensive set of unit tests, and then rewrite it in idiomatic clojure
17:16	technomancy	I think it's closest to MacLISP, which CL is pretty closely related.
17:16	lpetit	technomancy: no lexical scope?
17:16	technomancy	lpetit: =(
17:16	technomancy	it's available in an experimental branch, but it's opt-in
17:17	technomancy	amazingly enough you can fake it for most purposes with a macro, but heaven help you if you ever try to read the expansion
17:17	lpetit	technomancy: ever thought about writing an emacslisp -> clojure translator (or the other way around) ? :-)
17:18	technomancy	lpetit: once cinc is done and the lexical scope/threading branches get merged to mainline Emacs I'll probably play around with it a bit
17:18	technomancy	it will be a pretty natural compilation target once thread support is finished
17:19	lpetit	technomancy: seems like copying paredit.el code in a .clj file and doing the translation may be a rather systematic (and boring, btw :-) ) task. Lots of symbols replacements, replacing setq with some set!, etc.
17:20	technomancy	yeah, you'd probably want a pretty literal translation at first, then try to clean it up with a bit more FP once it was working
17:20	lpetit	technomancy: yes, make it work, write a decent test suite, make it right
17:21	lpetit	technomancy: yes.
17:22	spuz	lpetit: :o
17:22	technomancy	where would we be today if people let that stop them?
17:22	spuz	technomancy: finished?
17:23	lpetit	technomancy: yes, "They didn't know it was impossible, that's why they succeeded in it" (or something like that, in french "Ils ne savaient pas que c'était impossible, c'est pour cela qu'ils l'ont réussi")
17:23	lpetit	;)
17:23	lpetit	ok, time to go to bed, bye guys
17:23	doublindirection	is it possible to specify a throw clause to a :gen-class :method? I cannot find anything in the docs
17:24	chouser	doublindirection: I don't think so
17:24	lpetit	doublindirection: you don't need to do so
17:24	hiredman	paredit?
17:24	clojurebot	paredit source is http://mumble.net/~campbell/emacs/paredit.el
17:24	lpetit	or do you ?
17:24	chouser	Clojure doesn't have much patience for checked exceptions.
17:25	lpetit	it works again !
17:25	hiredman	lpetit: what you wanted never worked
17:25	hiredman	(but it does now)
17:25	lpetit	paredit.*
17:25	lpetit	~paredit.*
17:25	clojurebot	paredit commands is http://mumble.net/~campbell/emacs/paredit.html
17:25	hiredman	no not that
17:25	doublindirection	I'm generating API for java coders, they will want checked exceptions
17:26	lpetit	hiredman: just dreaming :-)
17:26	hiredman	asking "paredit?" and expecting the fact for "paredit source"
17:26	lpetit	ok
17:26	lpetit	would it be a good idea for it to return everything that starts with paredit ? (multiple lines ?)
17:26	lpetit	~paredit source
17:26	clojurebot	paredit source is http://mumble.net/~campbell/emacs/paredit.el
17:26	lpetit	~paredit commands
17:26	clojurebot	paredit commands is http://mumble.net/~campbell/emacs/paredit.html
17:27	lpetit	~paredit tutorial
17:27	clojurebot	paredit tutorial is http://p.hagelb.org/paredit-outline
17:27	lpetit	paredit tuto?
17:27	lpetit	~paredit tuto?
17:27	clojurebot	paredit commands is http://mumble.net/~campbell/emacs/paredit.html
17:27	lpetit	I'm confused
17:28	lpetit	ok, must really leave now, cu
17:28	hiredman	I am ok with that
17:33	Anniepoo	,(reduce #(+ %1 %2) '())
17:33	clojurebot	java.lang.IllegalArgumentException: Wrong number of args passed to: sandbox$eval--4276$fn
17:34	Anniepoo	expected something other than that
17:35	chouser	yeah, that's a bit weird.
17:36	chouser	but it's documented
17:36	Anniepoo	I guess the converse question is, what should it return?
17:36	chouser	If coll contains no items, f must accept no arguments as well, and reduce returns the result of calling f with no arguments.
17:36	chouser	from the docstring
17:37	Anniepoo	maybe it's just the choice of exception to throw
17:38	hiredman	reduce doesn't throw any exceptions
17:38	hiredman	fn's throw exceptions if misapplied
17:38	Anniepoo	ah, ok, now makes sense
17:51	jweiss	can someone recommend a key-value db to use with clojure? i would like to implement a bug tracking system and i don't want to be stuck dealing with rdbms schemas, i want it to be flexible
17:52	technomancy	jweiss: couch is easy. I've heard mongo is nice too if you need it to handle absurdly large datasets.
17:53	jweiss	technomancy: i doubt i need to handle absurdly large dbs. redhat's very old and very large bug tracking system is only up to 500k records
17:54	kwertii	jweiss: Berkely DB?
17:54	kwertii	*Berkeley
17:54	Anniepoo	http://w-hat.com/httpdb/
17:54	Anniepoo	deadly simple
17:56	technomancy	if you want a web-based system, slapping a JS UI on top of couch is the simplest thing that could work
17:56	technomancy	but it doesn't involve clojure. =)
17:56	chouser	Anniepoo: ooo
17:56	Anniepoo	???
17:56	jweiss	technomancy: i want to build an api/cli first and add a web ui later
17:56	chouser	I've written a few underpowered versions of httpdb for various little ajax projects.
17:57	Anniepoo	had the impression the request was for simple and flexible
17:57	chouser	Anniepoo: apparently I should have been using httpdb. :-)
17:57	kwertii	Is anyone working on an Elephant-like transparent persistence system for Clojure yet?
17:58	jweiss	flexible, not necessarily simple. would be nice if there's an existing redhat or debian package (which is +1 for couch and berkeley)
17:58	Anniepoo	well, if you trust your data to w-hat, go for it
17:58	jweiss	i was under the impression that clojure's json-fu is not very strong
17:59	hiredman	jweiss: very wrong
17:59	jweiss	oh i stand corrected then
17:59	hiredman	there are two json encoder/decoders for clojure, and another encoder as the demo for fnparse
17:59	hiredman	one of the json libs is in contrib
18:00	hiredman	anyway, derby is nice and easy to use
18:00	jweiss	ok so whose couchdb lib should i try first, danlarkin's or thekenny's
18:00	jweiss	there's also this redis thing: http://github.com/ragnard/redis-clojure
18:03	hiredman	http://db.apache.org/derby/
18:04	drewr	can anyone compare parsec and fnparse? are they similar?
18:04	hiredman	http://bret.appspot.com/entry/how-friendfeed-uses-mysql
18:15	maacl	jweiss: Clutch has worked out good for me
18:15	kwertii	Are there any good examples of non-trivial open source Compojure apps out there?
18:16	technomancy	kwertii: clojars may be a good one
18:17	kwertii	technomancy: thanks, will take a look
18:19	jweiss	maacl: so it looks like you can query directly in clojure using clutch?
18:20	maacl	jweiss: yep - it has a view server
18:20	jweiss	maal - interesting, i will look into that, thanks
18:20	jweiss	maacl: ^
18:26	lazy1	How can I get the name of a function which is passed as a parameter? (actually a vector of functions)
18:28	hiredman	lazy1: functions don't have names
18:29	lazy1	So I need to explicitly add metadata?
18:29	hiredman	functions don't take metadata
18:29	lazy1	The var the points to the function?
18:29	lazy1	s/the/that/
18:29	hiredman	yes
18:30	hiredman	vars do have names
18:30	hiredman	,(prn #'+)
18:30	clojurebot	#'clojure.core/+
18:30	hiredman	,(prn-str #'+)
18:30	clojurebot	"#'clojure.core/+\n"
18:31	lazy1	But if I get an array of functions, I can't use #' on them (or do I?)
18:32	kwertii	technomancy: is technomancy.us written in Clojure?
18:32	Chousuke	lazy1: you don't need #' to use functions as values
18:33	lazy1	I know, sorry for not being clear
18:33	lazy1	I've tried something like http://paste.lisp.org/display/92451, but I get an error on var
18:33	Chousuke	lazy1: #'foo is short for (var foo), so it won't work on values of anything.
18:33	technomancy	kwertii: nope: http://p.hagelb.org/Rakefile.html
18:34	Chousuke	the argument to var is a symbol (foo)
18:34	lazy1	Ah
18:34	lazy1	Is there a way to write nameof?
18:34	Chousuke	for arbitrary functions? no :/
18:34	Chousuke	some functions don't even have names.
18:34	kwertii	technomancy: ah, interesting. having grown frustrated with merb and rails, I am now weighing Ramaze versus Compojure.
18:35	technomancy	kwertii: I vote for static files. =)
18:35	kwertii	technomancy: would love to, but it has to front a SQL database :)
18:35	lazy1	Is that why we have deftest?
18:38	hiredman	lazy1: what do you mean?
18:38	lazy1	So you'll be able to get the name of the failed test
18:40	hiredman	lazy1: I haven't really used deftest, but from what I recall it doesn't print out "names" of tests ever
18:41	hiredman	the test runner runs the tests for a namespace and blows up if they fail
18:42	lazy1	OK, then let me get to a higher level
18:43	alexyk	mattrepl: figured it's easier to chat here than over github mail! :)
18:43	lazy1	I'm trying to distribute functions (tests) between several agents, and I'd like to print an error if one of them fails. How can I do that?
18:43	lazy1	(that = name of the function that failed)
18:44	Chousuke	pass the name of the function to the agents as well :P
18:44	alexyk	how do we do default arguments in clojure?
18:45	lazy1	I tried to avoid that, but I like the simplicity of that
18:45	Chousuke	alexyk: overload by argcount usually
18:45	lazy1	Thanks
18:45	alexyk	ah ok
18:45	alexyk	anybody else familar with oauth here?
18:46	Chousuke	(fn foo ([] (foo 1 2)) ([x y] (+ x y))) is a common pattern
18:46	Chousuke	except with defn, obviously. :)
18:47	alexyk	Chousuke: so optional will be nil in one branch; no real way to distinguish optional from default with non-nils, right?
18:47	Chousuke	alexyk: well, it's whatever you pass from the overloaded lesser arity.
18:49	ksm	wait, so a function can have multiple bodies?
18:49	hiredman	gah
18:50	Chousuke	ksm: yeah. overloaded on arity
18:50	defn	hmmm -- where'd clojure.test go?
18:51	ksm	neat
18:56	defn	am i crazy? did clojure.test vanish?
18:57	lazy1	defn: (require 'clojure.test) works for me
18:58	lazy1	,(require 'clojure.test)
18:58	clojurebot	nil
18:58	defn	yeah for some reason that ns doesn't exist for me?
18:58	lazy1	which version of clojure?
18:58	lazy1	IIRC it moved to clojure core after 1.0
18:58	defn	so maybe im running 1.0?
19:00	lazy1	Maybe, there is a stub left in clojure.contrib.test for backward compatibility
19:02	defn	yeah im running old clj
19:14	headius	rhickey: around?
19:39	alexyk	Chousuke: so what's the syntax for optional params like (fun x :and y :maybe z)?
19:41	the-kenny	alexyk: for defn? [foo bar & [a b c]]
19:42	alexyk	so :only is not a map key?
19:43	the-kenny	hm.. I don't know what you mean.
19:43	the-kenny	Sorry
19:43	defn	sweeet -- testing is fun
19:43	defn	:X
19:44	kylesmith	can anyone explain why (read-string "abc123") works, and (read-string "123abc") throws a NumberFormatException?
19:45	the-kenny	kylesmith: Because the reader tries to interprete 123abc as a number, but fails to parse the a?
19:46	kylesmith	yes, but why doesn't read-string return the entire string when happens?
19:46	kylesmith	,(doc read-string)
19:46	clojurebot	"([s]); Reads one object from the string s"
19:47	the-kenny	hm.. it the same when you type "123abc" in the repl.
19:47	kylesmith	"123abc" is certainly an object, so why not 'read' that and return it?
19:47	the-kenny	kylesmith: but read-string just sees 123abc, not the " around it
19:47	the-kenny	,(read-string "\"123abc\"")
19:47	clojurebot	"123abc"
19:50	the-kenny	If you type 123abc in the repl, it also doesn't gets interpreted as a string
19:51	kylesmith	true, but the doc doesn't say "Attempts to read the first possible object, or throws an exception"
19:55	technoma`	kylesmith: you can't start a symbol with a digit
19:59	kylesmith	what I'm really trying to do is read all objects from a string by wrapping the string in parens, and calling read-string.
20:05	technoma`	well if you want symbols back then you'll need to wrap it in a quote as well
20:05	technoma`	but do you really want symbols back?
20:06	rhickey	headius: hey
20:07	kylesmith	technoma`: you talkin to me?
20:07	technoma`	kylesmith: right
20:08	kylesmith	I want a seq of strings to be read, not symbols
20:08	kylesmith	it would be nice if read took a 'template' of types
20:09	chouser	kylesmith: 'read' is for converting text to clojure objects, just the way clojure source is read.
20:10	kylesmith	i.e, tell the parser to interpret the first object as an x, the second as a y, the third as a vector of [foo bar baz] etc
20:10	mattrepl	alexyk: there's defnk in c.c.def, alternatively see this page: http://stackoverflow.com/questions/717963/clojure-keyword-arguments
20:11	mattrepl	heh
20:11	kylesmith	chouser: yes, but read can sometimes be abused for parsing files ;)
20:13	JonSmith	why do people go crazy about typing 2 extra parens/curly-brackets/square-braces?
20:14	defn	because they're crazy
20:15	defn	JonSmith: it really depends on where and why
20:16	defn	i can see why someone would be annoyed by typing them where they're clearly not needed or don't represent anything of importance
20:16	defn	and yet are required
20:18	JonSmith	I always have thought the CL &key &rest &optional &aux etc. were confusing
20:25	defn	when i use (deftest mytestname) -- does mytestname need to be unique
20:25	defn	err let's say I have (defn mytestname [])
20:25	defn	can i have (deftest mytestname) as well?
20:25	defn	or will those overlap?
20:26	technoma`	defn: they're all vars
20:26	defn	technoma`: thanks
20:30	developernotes	Has anyone ever seen the message "Cannot open load file, swank-clojure-autoload" when attempting to launch slime with clojure?
20:30	technoma`	developernotes: you're probably using a deprecated configuration
20:31	technoma`	that file hasn't existed for quite some time now
20:31	developernotes	I originally installed clojure in emacs through ELPA, is there a good way to upgrade this?
20:33	technoma`	developernotes: you can remove all your old configuration. if you're using elpa then simply installing the packages will be enough
20:33	technoma`	in particular calling clojure-slime-config is no longer needed
20:33	ohpauleez	Does anyone else compile and run clojure on jvm 1.7?
20:34	developernotes	technoma`: okay, I'll try that, will that still allow me to run SBCL, I assume so, they shouldn't be tied to that right?
20:35	technoma`	developernotes: in that case you might have to keep around some of your old config, but not clojure-slime-config
20:35	technoma`	you should be able to use it in harmony with CL, but I haven't tried that myself
20:35	developernotes	technoma`: okay, I'll give it a go.
20:35	developernotes	technoma`: thanks.
20:36	technoma`	np
20:47	developernotes	I just installed swank-clojure for emacs, but everytime I attempt to run M-x slime, it just loads SBCL instead of clojure, any thoughts as to what might cause this?
20:49	developernotes	I was under the impression that M-x slime would launch clojure with swank-clojure installed and that I would need to use C-u M-x slime to provide an alternate process to run (i.e., sbcl).
20:49	developernotes	Any thoughts on getting clojure to run within emacs through slime, I already have swank-clojure installed, and also clojure itself.
20:51	defn	developernotes: you're using jochu's swank-clojure -- use technomancy's elpa version of clojure-mode, swank-clojure
20:51	defn	remove any configuration you've done
20:51	developernotes	defn: I installed swank-clojure from ELPA.
20:51	defn	and clojure-mode?
20:51	defn	did you run M-x clojure-install?
20:52	defn	you'll also need slime from elpa
20:53	developernotes	defn: when I run M-x clojure-install, it prompts for the directory I previously installed it to, but when I say 'y', it says that this function is deprecated in favor of swank-clojure.
20:53	developernotes	defn: thoughts?
20:54	defn	developernotes: not really -- it seems like you have everything working, but that there's another swank-clojure floating about
20:54	developernotes	defn: I have slime from ELPA installed, but not slime-repl from ELPA, do I need that?
20:55	defn	yes
20:55	defn	get that
20:55	defn	Can anyone tell me how to get ("abc") out of the list so it's just "abc"?
20:57	hiredman	seriously?
20:57	hiredman	,(first '("abc"))
20:57	clojurebot	"abc"
20:58	defn	bahhh sorry hiredman -- i accidentally had (first (str '("abc")))
21:01	kwertii	anyone know why C-c C-k produces "Wrong number of args passed to: basic$eval--1058$compile-file-for-emacs"?
21:01	ssideris	,(first (first '("abc")))
21:01	clojurebot	\a
21:02	kwertii	afaik I am using clojure 1.1 nightly from git, as well as the latest swank-clojure and clojure-mode
21:04	headius	rhickey: sorry, missed ya, around now?
21:05	rhickey	headius: hi
21:05	headius	hey there
21:05	headius	I think we're talking past each other on jvm-l
21:05	rhickey	a bit
21:05	headius	I think everyone considers it a given that stub generation is one way to make this work, since that's what groovy's joint compiler does, for example
21:05	ssideris	,(
21:05	clojurebot	EOF while reading
21:05	rhickey	my main concern is not building a mega-compiler out of all our compilers joined by the coordinator
21:06	headius	yeah, I don't want that either
21:06	rhickey	I think making type info available to tools like IDEs via APIs is fine
21:06	headius	what I want is really just a compiler into which I can make an upcall and say "I need this type and I don't know how to find it", and it will either look at existing classes or query other languages being compiled at the same time
21:06	rhickey	but when I try to think through the compilation workflow that way, it seems complex
21:06	headius	it may be
21:06	headius	and it may be too complex
21:07	rhickey	we need C++ header files :)
21:07	headius	yeah, that's basically what this ends up as
21:07	headius	virtual header files
21:07	rhickey	there are some benefits to that
21:08	headius	I know the inference case is a little obscure too, but I know the scala guys have endless battles getting IDEs to support cross-language type management, refactoring, etc
21:08	rhickey	inference is very scary - I think we need them to step up rather than speculate about what might work
21:08	headius	refactoring and friends are certainly out of scope for this, but being able to just get a "map" of what types a set of files provides and consumes would provide the underpinnings for all that as well as a joint compiler
21:09	rhickey	all of these concepts are really pretty foreign for langs that don't use classes in the ordinary way - as John Cowan said
21:09	headius	certainly, but that's a separate discussion
21:10	headius	this is for joint compiling languages that want to present Java types and have cross-language, possibly-circular dependencies
21:10	headius	and I'm trying to stop the spread of one-off joint compilers
21:10	headius	there's a groovy/java compiler that's mostly separate from the groovy/scala/java compiler, which is separate from the scala/java compiler
21:10	headius	so many wheels being reinvented
21:11	rhickey	and those langs are the most similar
21:11	headius	it won't be long before jruby can enlist in those compilers almost as easily, and I don't want it to need another one-off
21:12	headius	at any rate, I wanted you to understand I'm trying to gather all potential requirements, then figure out what's reasonable and in scope, and then decide on which approach to take
21:12	headius	rather than just diving into an approach
21:12	rhickey	what's most appealing about Per's suggestion is that each compiler can work as-is (i.e. non-joint), given some separate way to enumerate types, and another to generate stubs
21:13	rhickey	but yes, we need some use cases and scenarios
21:13	headius	the stub-based thing is certainly the easiest, but it's also the "dumbest" in that it doesn't provide a lot of useful tooling beyond belching out classes
21:13	headius	and that may be fine
21:13	rhickey	the pathological cases require topological sorts etc
21:14	headius	I also have a side motive, which is to try to get scala guys to figure out the hard parts of cross-language inference in a common compiler so I don't have to write any of that for duby
21:14	rads	using the following code, when I try to call (.setName (new cljbot.pircbot) "test"), I get no method found: http://gist.github.com/261442 anyone know why it doesn't work?
21:14	headius	my ideal setup would be a service which knows how Java types are structured and how method lookup works that I can just toss mirrors into and get mirrors out
21:15	rads	(I just want to make Pircbot#setName public)
21:15	headius	I've started to hack that a bit in Duby, but it's all in Ruby and not useful in general
21:15	headius	I suppose one way to explore this would be to write it in Java and base it on the mirror types, which has been planned for Duby for a whil
21:17	rhickey	a prototype never hurts
21:22	headius	I'll see if I can mock something up
21:23	rhickey	cool
21:23	defn	headius: Duby?
21:24	headius	my ruby-syntaxed static-typed, local-type-inferred, compiler-pluggable language
21:24	headius	a toy, but maybe more someday
21:25	defn	headius: cool
21:25	headius	http://kenai.com/projects/duby
21:26	headius	grr....javax.lang type names are almost the same as com.sun.mirror
21:27	defn	http://blog.fogus.me/2009/12/21/clojures-pre-and-post/
21:28	bettsp	Newb question, what does a '*' after a function name mean?
21:29	bettsp	http://github.com/ragnard/redis-clojure/blob/master/src/redis/internal.clj on line 38 for example
21:29	defn	bettsp: depends on the function
21:29	defn	it's not a pointer or anything
21:29	dnolen	bettsp: has no real meaning, it's a convention really.
21:29	bettsp	defn: Ok, so there's no idiomatic meaning though
21:29	defn	not that im aware of, ive seen it used in many different cases
21:29	bettsp	defn: accepted throughout Clojure that is
21:30	defn	usually it's a way of differentiating myfn from a special myfn*
21:30	defn	err not usually, in some cases that ive seen, that's how it behaves
21:30	dnolen	bettsp: usually it differentiates between a macro and the function that does the work.
21:31	bettsp	defn: But that's not the same as *s in (set!), which is a convention throughout the language
21:31	bettsp	Like (set! some-variable true)
21:33	PrinceOfA	rhickey: According to wikipedia, you funded the development of Clojure out of your own pocket for 2,5 years. Has Clojure turned out to be a financial success for you?
21:34	rhickey	PrinceOfA: um, not even close
21:35	PrinceOfA	rhickey: has it somehow changed your professional life or do you still do the same stuff?
21:35	PrinceOfA	or same stuff + clojure development
21:35	rhickey	PrinceOfA: mostly Clojure now
21:39	headius	hmm
21:42	PrinceOfA	rhickey: has clojure turned out to be what you expected it to be when you started the effort? or did it turn out to be something else, i mean both technologically and as a community
21:44	rhickey	PrinceOfA: I never expected it to hit the way it did
21:47	tolstoy	Anyone have a recommendation for an xmlrpc client? Has someone wrapped that all up in a clojure lib yet?
21:47	PrinceOfA	rhickey: ok, thanks. i'm asking because clojure as a codebase and a tool, the github thing and the community all have a very positive feel to it. i'm wondering how much of it is planned and how much "just happened" :)
21:47	tolstoy	There's one for Apache, and there's this Redstone one....
21:48	ssideris	rhickey: if it's not a financial success for you personally, how do you manage to mostly do clojure these days? i mean financially if you don't mind me asking
21:48	ssideris	i'm asking mainly because i would like to do the same
21:48	ssideris	(give up the dayjob and work on my own projects)
21:53	rads	anyone know why (.setName x) still doesn't work using this code? http://gist.github.com/261442
21:53	tolstoy	rads: Yep.
21:54	tolstoy	rads: That method is protected.
21:54	tolstoy	You can use (in contrib java-utils) wall-hack-method to get at it.
21:54	rhickey	ssideris: http://clojure.org/funding
21:55	tolstoy	rads: I just worked on that very problem yesterday! Also see hiredman's clojure bot (githup) to see how he worked it.
21:55	rads	tolstoy: cool, thanks
21:56	tolstoy	(wall-hack-method org.jibble.pircbot.PircBot :setName [String] bot name)
21:58	ssideris	rhickey: ok :-)
22:00	kwertii	It seems that "lein swank" does not include the src/ directory in the classpath, but "lein repl" does. Am I doing something wrong, or is this a bug?
22:04	alexyk	is there a way to list names defined in repl?
22:06	mitchellh	alexyk: (ns-public ns) will list all the public ones
22:06	alexyk	mitchellh: thx
22:06	alexyk	hm... Unable to resolve symbol: ns-public in this context
22:07	alexyk	should I import it from somewhere?
22:07	mitchellh	its part of clojure.core so you shouldn't have to
22:07	mitchellh	but just to be sure try
22:07	mitchellh	(clojure.core/ns-publics ns)
22:07	mitchellh	alexyk: http://richhickey.github.com/clojure/clojure.core-api.html#clojure.core/ns-publics
22:07	alexyk	ah, ns-publics
22:07	mitchellh	oh, did I misspell it previously? My mistake
22:08	alexyk	np, works now
22:08	alexyk	cuute
22:08	alexyk	rhickey: you got something really special going! :) I like it better every day
22:08	mitchellh	There are others, like ns-interns and such but I just started with clojure so I'm not sure which you want :)
22:09	polypus	i just noticed something. M-x swank-clojure-project ignores the :library-path in project.clj. to get it to work i had to symbolic link in the project root to my lib folder
22:09	alexyk	_ato: did you really have just two months Clojure experience before doing WF2, per twbray?
22:09	alexyk	it's hard to beieve
22:11	ssideris	alexyk: what's the project you're referring to?
22:11	alexyk	ssideris: Wide Finder 2
22:12	alexyk	clojurebot: ~Wide Finder
22:12	clojurebot	Huh?
22:12	alexyk	clojurebot: ~"wide finder dammit"
22:12	clojurebot	Excuse me?
22:12	alexyk	clojurebot: ~"wake up hiredman"
22:12	clojurebot	when I write functions I go blind and let the lambda do its thing
22:12	alexyk	http://www.tbray.org/ongoing/When/200x/2009/09/27/Concur-dot-next
22:13	ssideris	oh yes, i read about wide finder 2
22:13	alexyk	furthermore, that would put making clojars at day 1 or 2 of clojure, which is harder to believe still
22:15	hiredman	what?
22:16	clojurebot	What is meta
22:16	alexyk	hiredman: what what?
22:16	alexyk	clojurebot's on strike, not googling
22:16	hiredman	~google stuff
22:16	clojurebot	First, out of 35800000 results is:
22:16	clojurebot	The Story of Stuff
22:16	clojurebot	http://www.storyofstuff.com/
22:16	hiredman	?
22:16	alexyk	how do we google "wide finder"?
22:17	hiredman	clojurebot: google wide finder
22:17	clojurebot	First, out of 5240000 results is:
22:17	clojurebot	Wide Finder Home - Wide-Finder - wikis.sun.com
22:17	clojurebot	http://wikis.sun.com/display/WideFinder/
22:17	alexyk	ah... so ~ is for single word?
22:17	hiredman	~ is short for clojurebot:
22:17	clojurebot	for is not a loop
22:18	alexyk	ah! the mysteries unravel
22:18	hiredman	you have to tell clojurebot you want to google
22:18	alexyk	clojurebot: I want to google something
22:18	clojurebot	rationale is http://clojure.org/rationale
22:22	JonSmith	clojurebot: google foo
22:22	clojurebot	First, out of 3010000 results is:
22:22	clojurebot	Foobar - Wikipedia, the free encyclopedia
22:22	clojurebot	http://en.wikipedia.org/wiki/Foobar
22:26	PrinceOfA	does the proxy macro require the functions required by the interface to be declared where the proxy macro is used? for example, com.sun.net.httpserver.HttpHandler has "void handle(HttpExchange exchange)". when creating a proxy, am i forced to do (proxy [HttpHandler] [] (handle [HttpExchange] ...))) or can i somehow refer to a handle function elsewhere?
22:27	PrinceOfA	*handle function which is defined elsewhere
22:28	ohpauleez	has anyone been having issues with lein repl?
22:30	kwertii	ohpauleez: lein repl works fine here, it's lein swank that's causing me problems :)
22:31	ohpauleez	ah, gotcha, thanks for letting me know kwertii
22:40	arohner	(and foo (every? identity foo))
22:40	arohner	is there a more concise way to write that? it feels wrong
22:42	arohner	actually, that doesn't do what I want either, it fails on an empty list
22:43	arohner	I want 1) foo is a seq 2) foo is non-empty 3) every value of foo is true
22:47	JonSmith	you could use filter i guess
22:47	JonSmith	meh, that's kind of silly though
22:49	JonSmith	,(= () nil)
22:49	clojurebot	false
22:49	JonSmith	,(= '() nil)
22:49	clojurebot	false
22:49	JonSmith	,(= nil false)
22:49	clojurebot	false
22:49	JonSmith	,(= false false)
22:49	clojurebot	true
22:50	JonSmith	,(reduce #(and %1 %2) (list true true true true))
22:50	clojurebot	true
22:51	JonSmith	,(reduce #(and %1 %2) (list true true false false))
22:51	clojurebot	false
22:51	JonSmith	,(reduce #(and %1 %2) (list))
22:51	clojurebot	java.lang.IllegalArgumentException: Wrong number of args passed to: sandbox$eval--4356$fn
22:51	arohner	,(reduce #(and %&) (list))
22:51	clojurebot	nil
22:52	arohner	,(reduce #(and %&))
22:52	clojurebot	java.lang.IllegalArgumentException: Wrong number of args passed to: core$reduce
22:52	arohner	,(reduce #(and %&) nil)
22:52	clojurebot	nil
22:53	arohner	I think that works. thanks
23:20	arohner	is there a way to introspect a multimethod to see what values it will work on?
23:21	arohner	nm
23:21	arohner	(.getMethodTable multi-method)
23:21	arohner	reading the source is good for you!
23:25	ohpauleez	arohner: (methods) should also work, right? http://github.com/richhickey/clojure/blob/a9cb831fe714966e225a5297ef844513fe4a373a/src/clj/clojure/core.clj#L1066
23:25	ohpauleez	which is just sugar .getMethodTable
23:26	arohner	oh, nice
23:27	arohner	that's not mentioned on the multimethods page