#clojure – 2011-10-03

2011-10-03

00:00	brehaut	(this analogy doesnt really extend very well does it)
00:00	ibdknox	lego technic was awesome
00:00	cemerick	brehaut: technic is strint; cl-format is mindstorms
00:01	ibdknox	hm
00:01	cemerick	the old contrib is most hightly ranked :-(
00:04	brehaut	whoa, strint is cool. i fear i will lose my mind if i try to understand cl-format
00:04	ibdknox	yeah cl-format seemed a bit crazy to me
00:04	brehaut	common lisp right?
00:04	ibdknox	brehaut: link to strint?
00:04	brehaut	https://github.com/clojure/core.incubator/blob/master/src/main/clojure/clojure/core/strint.clj
00:05	ibdknox	sweet
00:05	jli	what's mindstorms about cl-format? being able to include format strings in your format strings? :)
00:05	jli	(at least, I thought that was possible)
00:07	jli	http://www.gigamonkeys.com/book/a-few-format-recipes.html
00:07	jli	ah, it's slowly coming back now
00:09	jli	(format nil "~r" 1234) ==> "one thousand two hundred thirty-four"
00:09	jli	(format nil "~@r" 1234) ==> "MCCXXXIV"
00:11	jli	(format nil "~{~a~#[~;, and ~:;, ~]~}" (list 1 2 3)) ==> "1, 2, and 3"
00:12	jli	wow, this is really starting to go off the deep end
00:12	brehaut	jli: lets be honest, this is embeded apl isnt it
00:12	jli	brehaut: yeah. I just projected all the chars down into ascii and added parens.
00:13	brehaut	thought so
00:13	jli	"And there's more--I haven't mentioned the ~? directive, which can take snippets of control strings from the format arguments or the ~/ directive, which allows you to call an arbitrary function to handle the next format argument."
00:13	jli	this is way worse than regexes
00:14	jli	Let's say you have a problem and you decide to use cl-format. Now, kill yourself.
00:14	brehaut	im horrified that the blub paradox might extend to string interpolation functions
00:17	jli	brehaut: how do you mean?
00:19	jli	(coincidentally, I queued up Beating the Averages in firefox earlier tonight and I'm getting to it now)
00:19	amalloy	good read. that's one of two articles that got me over the hump and actually learning lisp
00:20	brehaut	jli: i cant actually explain it any better than that
00:20	amalloy	brehaut: i don't recognize the blub paradox. is it related to the lisp curse?
00:20	brehaut	amalloy: its another PG thing; any language more expressive than the one you know (blub) looks like ridiculous gibberish
00:21	amalloy	ah, that one
00:22	amalloy	i think he was talking about more like "pointless fluff" than "ridiculous gibberish", but i suppose it applies here too
00:22	brehaut	conversely, i cant remember the lisp curse
00:22	jli	brehaut: so you mean people who write their entire program in a single cl-format look down on us heathens trying to get by with strint etc.?
00:22	amalloy	brehaut: it's so easy to write an 80% solution in lisp that there are zillions of them, and no 100% solutions
00:23	amalloy	in blub, it's so hard to do anything at all that you get together a hundred people and eventually finish the 100%
00:23	jli	amalloy: haha. so basically, people inlining library code into their application?
00:23	amalloy	which leads people to prefer blub
00:23	brehaut	jli: well presumably there are people who have mastered cl-format and its clear and obvious to them, but to people like me with blub format functions, it looks unnecessarily complex
00:24	amalloy	something like that. "there's a crappy library to do basically what i want, but it's missing N things; i'll just write my own"
00:24	brehaut	amalloy: so its also the javascript curse :P
00:24	dnolen	amalloy: brehaut: I think it should be renamed to the JS curse.
00:24	jli	brehaut: well, most of its capabilities seem useful. it just seems like write-once, read-never code
00:24	dnolen	NIH syndrome way less in Clojure land.
00:25	brehaut	dnolen: yeah, one upon a time i would have called myself a javascript developer. these days i am not brave enough. thats a scary bad place
00:25	technomancy	brehaut: multiple major modes is still an unsolved problem for Emacs
00:26	technomancy	I am not sure it's worth the effort of implementing either
00:26	brehaut	technomancy: if you dont have to use something like djangos template system i'd agree
00:26	jli	amalloy: anyway. yeah, PG. I think my poorly reconstructed road to lisp is roughly: watch The Matrix a lot -> read ESR's How To Become a Hacker -> read PG -> common lisp
00:26	technomancy	brehaut: yeah, I'm really glad I don't have to use Ruby's ERB
00:27	technomancy	I tried to write an emacs mode for it, but it was awful
00:27	technomancy	it's just the wrong problem to solve
00:27	amalloy	jli: the other one, for me, was http://www.defmacro.org/ramblings/lisp.html
00:28	brehaut	technomancy: i just googled. that is about as far from a good idea as you could get
00:28	brehaut	technomancy: why not just use PHP
00:31	jli	enlive didn't make sense to me until I learned css selectors
00:31	jli	(at least the parsing bits)
00:31	brehaut	jli: ah yeah, without that foundation it would seem like crazy nonsense
00:32	jli	yeah. ":>" was pretty bewildering.
00:32	brehaut	thats the chicken selector, it finds all the poultry and foul in your document
00:33	jli	fowl? ;)
00:33	brehaut	jli: yeah sorry, and not pheasants. dunno why
00:33	amalloy	i wish there were something to find all the foul in your html
00:34	ibdknox	me too
00:34	ibdknox	lol
00:34	amalloy	it would either help you validate, or get you to use sexps instead
00:35	brehaut	it would be nice if hiccup and enlive were more easily able to work together
00:36	jli	isn't it kind of bizarre that paul graham's site has the yahoo favicon still?
00:38	amalloy	brehaut: data.xml is probably a pretty good bridge for that
00:38	amalloy	among other things it can convert hiccup vectors/etc to tag/attrs/content maps
00:39	brehaut	oh yeah, that might do it
00:39	amalloy	that is the one feature i added to data.xml :P
00:39	brehaut	hah
00:39	brehaut	i need to work out if its how im using data.xml or if its a bug regarding escaping content in output text
00:39	brehaut	(in emit)
00:41	brehaut	sorry bout the horrible english
00:41	brehaut	its getting late i should go eat
00:41	napping	The irony is delicious - lisp is the most powerful language * programmers can't recognize languages more powerful than their own
00:42	brehaut	napping: haskell programmers agree ;)
00:43	amalloy	brehaut: looks like there's still no official release of data.xml, right?
00:44	jli	brehaut: hm. I know some haskell - would you say it's more powerful? it has an interesting type system and is really expressive, but still no macros
00:44	jli	well, not /no/ macrs, but you know what I mean
00:45	brehaut	amalloy: actually, i think i was thinking of clojure.xml
00:45	brehaut	jli, perhaps you should look up template haskell
00:45	amalloy	jli: lazy-everything and auto-currying get you a surprising part of the way towards macros
00:45	napping	not terribly unlike "no macros", though
00:46	jli	lazy everything makes a lot of control flow macros moot, I guess
00:46	amalloy	brehaut: my understanding is that template haskell is a hideous shambles
00:46	brehaut	amalloy: haha yes :P
00:46	brehaut	but, its there
00:46	brehaut	the other thing that gets you a long way is the ghc compiler rewriting rules
00:46	gienah	some of the template haskell tutorials suggest that its best to read On Lisp to figure it out
00:46	brehaut	frinstance, theres a rule that looks for map f $ map g l and rewrites it as map (f . g) l
00:48	brehaut	also, the type system does let you do some insane things, the c style printf but typesafe implementation is amazing
00:48	brehaut	(and frightening)
00:48	brehaut	as is heterogenous lists (HList)
00:49	jli	brehaut: I read some paper about implementing printf with continuations. is that how haskell printf works?
00:49	jli	brehaut: also, HLists - ack! google google
00:49	napping	you'll say that after you find it too..
00:49	jli	is it based on existential types?
00:50	brehaut	jli: the code is far more massive than i can fathom
00:50	brehaut	http://hackage.haskell.org/package/HList
00:51	napping	QuickCheck is a saner example
00:51	brehaut	jli: i think its just typeclass magic (due to typeclasses being polymorphic on the return type of functions)
00:51	brehaut	combined with no parens for calls and curried functions
00:51	brehaut	i guess it could be continuations… dunnoh
00:52	jli	brehaut: what client are you using that keeps inserting unicode goodness?
00:53	brehaut	err, colloquay on a mac and me insisting on typographically ellipsis rather than three dots
00:53	jli	haskell is a pretty fun place to play with pointlessness, what with $, ., flip, etc.
00:53	jli	brehaut: oh, is it not auto-converting that?
00:53	brehaut	jli: fogus informed me once that if i thought thahthaskell was good for PF, i needed to try cat
00:53	jli	have you memorized the utf8 code point?
00:54	brehaut	jli: i dont think so. opt+; is …
00:54	jli	ah
00:56	gienah	a haskell type safe printf is described in section 4 of this paper: http://www.haskell.org/haskellwiki/Simonpj/Talk:FunWithTypeFuns
01:01	jli	mm, I need to spend some time with the git book. using mostly hg for months has wiped a lot of stuff.
01:47	napping	to print some examples with results, can I avoid something like this (defmacro example [form] `(do (println (quote ~form)) (println ";=>" ~form)))
03:19	Blkt	good morning everyone
04:01	kzar	ibdknox: Is there a way to access the request header / modify the response headers within defpage? Using middleware for some things seems over the top.
04:02	kzar	ibdknox: (Sorry I meant to say, using Noir)
05:01	justin`	if I have a map that I need to update 3 seperate values in, in there a better way to do it than just (-> m (update-in [:x] inc) (update-in [:y] dec) (update-in [:z] fn)) ?
05:06	mrh0057	No but if you are going to do it a lot you can write your function to perform that operation.
05:42	thorwil	justin`: perhaps you could do a for with [[:x inc] [:y dec] [:z fn]]
05:44	amalloy	it's not really possible with `for`. you can reduce, though
05:45	amalloy	&(reduce (fn [m k] (update-in m [k] inc)) {:x 1 :y 2 :z 3 :a 4} [:x :y :z])
05:45	lazybot	⇒ {:x 2, :y 3, :z 4, :a 4}
05:47	kzar	Is using .getBytes on a string idiomatic Clojure?
05:48	raek_	kzar: only if you supply the encoding parameter :)
05:48	raek_	kzar: what do you want to do?
05:50	kzar	raek: I'm trying to generate a HMAC-SHA512, I've got some of the way but figuring out the Java libraries is a puzzler
05:50	kzar	raek_: I was going from this example http://download.oracle.com/javase/1.4.2/docs/guide/security/jce/JCERefGuide.html#HmacEx
05:56	raek	kzar: ok, these algorithms operate on bytes (not chars) so you need to encode the string into a byte array (you can do this with (.getBytes string encoding) )
05:56	raek	but what byte value(s) does a character have?
05:56	raek	that completely depends on what encoding scheme you use
05:56	kzar	raek: I suppose it depnds on the encoding, hence the parameter
05:56	kzar	raek: Yea
05:57	kzar	raek: Is there a way to check encoding of a string?
05:57	raek	kzar: and that example is broken, since it will produce different hashes for the same string on different machines
05:57	raek	kzar: strings don't have encodings
05:58	raek	kzar: you can think of the connection between a byte array with, say, UTF-8 data and string as the conenction between a JPG stream and the picture you get on thescreen
05:59	raek	a picture can be stored in many formats
05:59	raek	likewise can a piece of text be stored in many formats
06:00	raek	kzar: so to for the hash function to make sense, it must either take an encoding argument or hard-code a certain encoding
06:02	raek	so you might document the function like "Computes the HMAC-MD5 hash of the bytes acquired by encoding the string s using encoding e"
06:03	kzar	raek: How come the string doesn't have an encoding though, how does it store the piece of text?
06:05	kzar	raek: Also the doFinal method for the Mac object seems to return a byte array, but mapping char over it throws an exception I think because some of the numbers are too high / low. How can I convert it back?
06:06	raek	kzar: yes, String uses some kind of encoding internally (which happens to be UTF-16) but that is an implementation detail
06:07	raek	kzar: ok, so the algorithm outputs a byte array. what does it contain?
06:07	raek	if it's not text, then you should not try to make a string from it
06:08	raek	strings on the JVM platform are for textual data only. there is no class for "byte strings" as in python3, for example
06:08	raek	only byte arrays
06:09	kzar	raek: Right OK, fair enough. I do need a string rather than a byte array though, so what do I do?
06:10	raek	that depends on which way you want to represent the bytes as a string
06:13	raek	if you are simply going to store this value in a file or send it over the network, OutputStream has a method that accepts byte arrays (so there's no reason to go from bytes to string back to bytes again in that case)
06:15	kzar	raek: Hmm I need to calculate the hmac-sha512 and then use it to check against the value they gave me. I'm not really sending the calculated value to a file or anything
06:16	raek	ok, so you received a string?
06:16	raek	is it a hex string or something else?
06:17	kzar	raek: Yea, so I'm attempting to write this function that takes a message, key (, and encoding I guess) and returns the result. I'm going to then use that as a piece of Ring middlewear, I get a value in the header and I need to check that it's correct
06:17	kzar	raek: That way we know the request came from our guys
06:18	raek	hrm
06:19	raek	unfortunately HTTP does not specify which encoding the headers are in
06:19	raek	which means that ideally they should be treated as bytes and not characters (since we do not know the encoding)
06:20	raek	but the servlet framework will decode the headers with some encoding scheme
06:21	raek	if you can guarantee that the header in question will only contain ASCII (for example by sending the data as hex or base64), maybe you could avoid this problem
06:25	kzar	raek: Hmm I think I'm probably more confused now
06:26	raek	kzar: much of what you have into interface with here is not encoding aware (which means that you get all kinds of nasty problems for non-ASCII text)
06:27	raek	kzar: is this header some sort of key?
06:28	raek	or is it a signature of the message?
06:28	kzar	raek: How do you mean? Well I look for a header like X-AUTH-HMAC or similar and then take that value and compare it to my calculated value
06:29	raek	ok, and what is the input to the HMAC-MD5 function? the request body?
06:30	kzar	raek: Yea, I'm actually doing a SHA512 instead of MD5 though. That's just the closest example I found
06:30	raek	ok
06:30	raek	so here's what I would do:
06:31	raek	make sure that the body (which is text, I presume) is always encoded in UTF-8
06:32	raek	let the X-AUTH-HMAC header value be encoded with base64
06:33	raek	so if you have the array [97 98 99], it will be transmitted as the ASCII string "YWJj" in the header
06:34	raek	then when you receive the body (string decoded from body with UTF-8) and the header (base64 string), get the bytes from the body by encoding it in UTF-8 again (with getBytes)
06:34	raek	and de-base64 the header to get the bytes stored in it
06:35	kzar	raek: OK, I've found an example that shows you how to turn the digest into a string as well. That's working
06:35	kzar	raek: Main thing that confuses me now, how to turn a string key into a java.security.key
06:35	raek	then you only have byte arrays which you can pass to the function and compare with java.util.Arrays.equals
06:38	raek	kzar: maybe you can use SecretKeySpec
06:38	raek	you need to give that the byte array of the key
06:38	kzar	raek: ahh
06:38	raek	(which you can obtain by de-base64:ing the header string)
06:39	raek	(if you choose to implement my approach)
06:39	kzar	raek: Yea I think I'll do it your way becaues you seem to be quite clear on the different encodings. Actually how did you learn all that? I always fudge my way through this stuff
06:40	raek	kzar: I'm a Swede. I see the results from encoding errors by simply using my native language. :P
06:41	kzar	aha
06:41	raek	well, it helps if you think of byte strings and character strings as icompatible types
06:41	raek	*in-
06:42	raek	and that somewhere a translation always has to be done
06:45	kzar	raek: Thanks for explaining all that, really helpful
06:46	raek	kzar: np. glad to help. :)
07:20	ilyak	Is there something like (let [foo (x) bar (y)] (body foo bar)), but which only evaluates body when foo and bar aren't null?
07:24	ejackson	nested if-let
07:28	michaelr525	hello!
07:28	Blafasel	Anyone related to 4clojure? The only guy I know that is a contributor is US based (or at least: sleeping, in a timezone where it's pitch black now). Site's dead :( I don't want to get back to real work :)
07:29	Blafasel	502 bad gateway sounds like the backend process for cgi/fcgi or something died..
07:33	ejackson	yeah, sounds like somebody needs to attend to a service
07:35	Blafasel	Ah well.. Back to C# then it is.
07:35	michaelr525	Blafasel: i think amalloy_ is related to 4clojure, try poking him maybe
07:36	michaelr525	...and back to C# as well ;)
07:40	michaelr525	amalloy_: ^^
07:44	Blafasel	michaelr525: He's the one that should sleep now :)
07:44	michaelr525	ah
07:44	Blafasel	I think. And the trailing _ seems to be the "I'm not here" marker :)
07:45	Blafasel	(ignoring that one could use /whois to check the away state)
07:49	ilyak	ejackson: Thanks
07:49	ilyak	when-let did the trick
07:50	ejackson	cool
07:53	lnostdal	any way to terminate the rogue thread created by an accidental send-off?
07:53	lnostdal	a rogue*
07:57	lnostdal	ns-unmap on the variable holding the agent didn't seem to work
08:25	fdaoud	good morning
08:25	fdaoud	anyone else getting 502 Bad Gateway on 4clojure.com?
08:25	zippy314	yep
09:03	tcrawley	There seem to be a few testing frameworks for clojure to choose from. Does anyone here have a preference they'd like to share?
09:04	ejackson	I use the standard test, but midje seems very popular
09:05	fdaoud	+1
09:07	tcrawley	ejackson: fdaoud: have either of you played with speclj? https://github.com/slagyr/speclj
09:07	ejackson	tcrawley: nope
09:08	fdaoud	I have not
09:09	tcrawley	speclj is attractive simply because I've done lots of rspec, but midje looks interesting
09:13	fdaoud	tcrawley: don't throw eggs at me-- I use Maven so having standard tests being executed within Maven test target is nice for me
09:14	tcrawley	fdaoud: heh - my tests will be executed by maven as well - are you using https://github.com/talios/clojure-maven-plugin?
09:16	fdaoud	tcrawley: yes, and with this piece of configuration, mvn test works: https://gist.github.com/1259078
09:17	fdaoud	reload, I put the whole piece
09:18	tcrawley	fdaoud: thanks!
09:18	dbushenko	what is the difference between fn and fn* ?
09:19	fdaoud	tcrawley: welcome
09:27	jbwiv	hello all. Anyone here using Intellij Idea for clojure dev with the la clojure plugin?
09:27	Arafangion	How does one get java programs to start faster?
09:27	Arafangion	Wait, it was javac -version I was checking.
09:28	Arafangion	java -version is only 10 or 20 times slower than python, that's not so bad.
09:28	jbwiv	Arafangion, you might read this article: http://blog.headius.com/2010/03/jruby-startup-time-tips.html. It's focused on Jruby, but would apply to any JVM lang
09:28	Chousuke	if you really need fast startup you can run a persistent jvm instance using eg. cake
09:28	Arafangion	Thanks. :)
09:29	Arafangion	Ooh, cake isn't neccessarily worksafe.
09:29	Chousuke	so basically it's just cheating by always running the jvm, but that might not be a bad solution
09:29	Chousuke	worksafe? :/
09:29	Arafangion	How do I google "cake" and not get, you know, cakes?
09:29	Arafangion	Chousuke: Lingerie cakes are result #4.
09:30	Chousuke	google clojure cake :P
09:30	Arafangion	Ok. :)
09:30	Arafangion	Ah, yep, that works. :)
09:30	fdaoud	mmm, cake.
09:30	andrewclegg	does "cake" mean something different outside the UK? what's a lingerie cake? I'd google it but I'm at work :-)
09:31	Arafangion	andrewclegg: I'm just going to hope it means "A cake in the shape of lingerie".
09:31	redinger	Yeah, googling cake for me doesn't result in any lingerie. :)
09:32	Arafangion	Well, google is known to give different results depending on where you are.
09:32	Arafangion	Just checked it out, and interesting, nothing to do wiht cake.
09:33	Arafangion	www.cakelingerie.com - not worksafe, but not exactly that outragious either, you'd do worse checking clothin advertisements. It's a website that 'congratulates you on your pregnancy'./
09:34	andrewclegg	aaaaaah :-)
09:35	Blafasel	Weird world in which this is not worksafe..
09:36	Arafangion	Depends on who is passing by behind you.
10:11	ejackson	I'm having a problem getting the evaluation time of arguments to a macro correct
10:11	ejackson	i've put up a gist https://gist.github.com/1259177
10:11	ejackson	any explanation or help would be most appreciated :)
10:13	ejackson	i'm unable to make the main argument available to the macro, very frustrating !
10:15	TimMc	Chousuke, chouser: Remember yesterday's discussion about the possible inadvisability of automatic subvec-like behavior with (vec (seq some-vec))?
10:16	TimMc	Chousuke, chouser: We just fixed a bug at work where String/substring and String/split held onto entire large documents even though we just wanted some small tokens -- which were then stored in a large in-memory hash table. Out of memory!
10:17	fdaoud	crap. I was really enjoying 4clojure.com and now it's down :(
10:18	TimMc	fdaoud: amalloy will be at work in about 4 hours, I would hazard.
10:18	Blafasel	ejackson: Disclaimer: I'm a lousy starter. Two questions: fm uses (symbol '=') ak uses 'and => Why the difference in quoting? More on topic: Why a macro? Couldn't it be a new method my-where taking a map and passing it on?
10:18	jkkramer	ejackson: if ak is a macro, then (cql/where (ak {:a 6 :m 9})) should work, i think
10:19	TimMc	fdaoud: Or... based on his last timestamp in #4clojure, maybe 5 hours. >_<
10:19	Blafasel	fdaoud: Heh. Just created a beautiful (yeah.. well..) solution, tested it locally and it went down before i could submit my pascal triangle :)
10:19	Blafasel	TimMc: /whois him :)
10:20	TimMc	Blafasel: What about it?
10:21	TimMc	The away timestamp means nothing without a time zone.
10:21	fdaoud	TimMc: thanks for the info
10:21	ejackson	Blafasel: quoting, no reason actually. macro: I was having trouble avoiding the evalution of the equalities, so I just wanted the forms. I'll try again though.
10:22	fdaoud	Blafasel: how many problems solved you at?
10:22	ejackson	jkkramer: Let me try that... thanks
10:22	ejackson	my head has been done in
10:22	Blafasel	TimMc: I was refering to "16:21:07 -!- idle : 0 days 4 hours 37 mins 7 secs [signon: Fri Sep 30 06:10:17 2011]"
10:22	Blafasel	fdaoud: No idea :)
10:22	Blafasel	Not a lot yet. 20? 30?
10:24	TimMc	Blafasel: Oh, cool. I just learned how to display that info.
10:31	ejackson	Blafasel: OK, I tried as a function https://gist.github.com/1259177#gistcomment-53987
10:31	Blafasel	To continue the stalking: It should be ~7:30 for him.. ;-p
10:31	ejackson	its passed the predicate as the list, but it where just returns it
10:32	ejackson	i don't understand why
10:36	ejackson	its a function argument so it should be evaluated by the time the where macro sees it, so its should just be a list. Yet it treats it as it if something unrecognised... aaaarg
10:36	Blafasel	ejackson: (defn better-where [p] (cql/where (ak p))) or (defn better-where2 [p] (apply cql/where (ak p))) won't work? Sorry if that's totally stupid?
10:36	Blafasel	I mean: Why the macro in the first place?
10:36	ejackson	Blafasel: lemme try
10:37	ejackson	Blafasel: Yeah, the macro, see my last mini rant about trying it a function, where just mocks it
10:38	Blafasel	Uhm.
10:38	gtrak	~no, kancho $1 is <reply>javabot teleports silently behind $1, crouches and yells 'a thousand years of pain!'
10:38	clojurebot	Ack. Ack.
10:38	Blafasel	Wrapping it explicitly as a function? #(ak p)?
10:38	ejackson	Blafasel: better-where2 doesn't work as where is a macro, so apply can't use it
10:39	gtrak	sry clojurebot, wrong window
10:40	ejackson	Blafasel: and better where is exactly the function I put in the comment, for some reason, which I don't understand, it doesn't work
10:40	ejackson	lemme
10:40	ejackson	try
10:40	ejackson	the
10:40	ejackson	wrapping as a function
10:43	theignorati	can I make leiningen put the jar it builds in a dir?
10:52	zippy314	Does anybody know why sometime I get classes reported back as Ljavax instead of javax? Specifically working with java mail I'm getting: [Ljavax.mail.internet.InternetAddress as a class type.
10:53	joegallo	that's an array of InternetAddresses, i think
10:53	twem2	zippy314: see http://download.oracle.com/javase/6/docs/api/java/lang/Class.html#getName%28%29
10:54	twem2	its an array
10:55	zippy314	THanks, thats exactly right.
11:04	jli	amalloy_: I don't suppose you could bring 4clojure back up now? :)
11:04	ejackson	wow, i didn't realise how popular that site is
11:07	fdaoud	jli: he's away
11:14	jli	ejackson: I keep crashing it
11:14	jli	fdaoud: ah, oh wel
11:15	ejackson	good work
11:15	jli	\o/
11:15	lnostdal	hm, i don't get this .. i got an app depending on a library, which in turn depends on hiccup-0.3.6.jar .. but looking in lib/ after a 'lein deps' in the app root folder shows hiccup-0.3.5.jar ...?
11:16	joegallo	the library and version that you're depending on would be helpful, if you're looking for diagnosis rather than commiseration :)
11:16	lnostdal	it's a local library here
11:17	lnostdal	it has :ring "0.3.11 and hiccup "0.3.6" in its :dependencies list (project.clj) .. if i run lein deps there it'll grab the right versions
11:17	brcrth	Hi, what's the advantage of having anonymous functions? How I'll "call" the function when I need it?
11:17	lnostdal	brcrth, http://en.wikipedia.org/wiki/Anonymous_function#Sorting
11:17	joegallo	brcth: you pass them around as arguments, or set them to local names in let, for instance
11:18	brcrth	thanks :)
11:18	joegallo	so you'd call them like this (the-argument) or (the-name-that-i-let)
11:18	joegallo	which is to say, just like other functions
11:20	lnostdal	this doesn't make any sense .. grepping, "0.3.5" isn't mentioned anywhere
11:21	kzar	raek: What do you think? https://gist.github.com/1259359 It's not working right but I think it's nearly there
11:50	TimMc	fdaoud: 4clojure is up again, thanks to amcnamara
11:51	raek	kzar: I think you want to use this instead of the private sun classes (which are not guaranteed to be there): http://commons.apache.org/codec/apidocs/org/apache/commons/codec/binary/Base64.html
11:51	raek	kzar: [commons-codec "1.4"]
11:51	fdaoud	TimMc: thanks!
11:52	gtrak	I think contrib has a base64, no?
11:52	raek	kzar: and you want to decode the base64 string (not encode it).
11:53	raek	decoding the base64 ASCII representation gives you the original bytes
11:53	kzar	raek: Silly me for taking an example at face value with no sanity check
11:53	raek	(but decoding bytes with a character encoding scheme gives you a string)
11:53	raek	so decode/encode has opposite types :)
11:54	lnostdal	does lein have some sort of verbose mode? .. so i can see why it's doing what it's doing?
11:54	raek	kzar: I would also split out conversion from string and base64 form to outside that function
11:54	gtrak	lnostdal, what in particular is so unpredictable?
11:55	raek	it's not relevant for the hmac algorithm in what way the bytes were represented before it gets them
11:56	raek	(defn bytes-from-auth-header-value [base64-string] (Base64/decodeBase64 base64-string))
11:56	lnostdal	gtrak, i have project "B" which depends on ring 0.3.11 .. project "A" depends on project "B", but when i run 'lein deps' in context of project "A" it'll actually grab ring 0.3.8 for some reason not revealed
11:56	raek	(defn encode-in-utf-8 [string] (.getBytes string "UTF-8"))
11:57	gtrak	lnostdal, for that, use lein pom to write a maven pom.xml, then use mvn dependencies:tree to see why it does what it does
11:57	gtrak	not sure if there's a native lein way to do that
11:58	gtrak	lnostdal, then, once you find which dependency is pulling in the bad ring, you can add an exclusion to your project.clj
11:58	raek	kzar: also, you have a problem here: (.getBytes key)
11:59	raek	this uses the platform default encoding. you'd better fix it to something non-changing (e.g. a constant "UTF-8")
12:00	raek	if the key is automatically generated, you should use the original byte array form (assuming it's avaiable)
12:01	lnostdal	ok .. now i can't even get ring to build itself .. mvn install tries to pull stuff from some remote repository ...
12:02	raek	if you need to base it on a user-supplied string, you need to make sure that the path from the user to the .getBytes is encoding-aware (e.g. does not pass though a HTTP header)
12:02	gtrak	lnostdal, huh?
12:02	gtrak	let maven do its thing
12:03	lnostdal	kinda makes local patching, well, hard
12:03	gtrak	what does?
12:03	lnostdal	it pulling stuff from clojars.org or whatever seemingly at random
12:03	gtrak	don't actually use maven for building, just use it to build you a dependency graph
12:04	gtrak	to do that, it needs to download its own plugins, plus all the poms for all the dependencies
12:05	gtrak	i think the command is actually mvn dependency:tree
12:27	kzar	raek: OK I think I've made those changes right, does using the = function to check if two byte arrays are equal work though?
12:27	dnolen	so now that we have Browser REPL, what about an iOS REPL ?
12:28	raek	kzar: no, = means "identical" for mutable thing, including arrays
12:28	raek	kzar: use (import 'java.util.Arrays) (Arrays/equals arr1 arr2)
12:29	raek	this means "currently have the same value"
12:29	lnostdal	seems use of :main (in project.clj) and :gen-class was confusing lein/mvn (or me; whatever really.. sigh)
12:29	dpritchett_	is there a repl as cool as say iPython or irb+wirble yet?
12:29	dpritchett_	for clojure that is
12:31	lnostdal	emacs+slime, dpritchett_ ?
12:31	dpritchett_	I mean a no-dependencies repl
12:31	lnostdal	sure beats irb last time i tried that
12:32	dpritchett_	i like to work in vim with a REPL that i occasionally reload files with
12:32	dpritchett_	my two environments are vim + ipython/bpython and vim + irb
12:32	lnostdal	so you have a repl already then
12:32	dpritchett_	I have tried to love emacs a few times just because of clojure and it's never workedo ut
12:32	dpritchett_	lnostdal: Does the repl have line numbers, colorization, auto indents, etc?
12:32	lnostdal	..but if you want a cooler one i bet slime is cooler
12:32	dpritchett_	this is the kind of stuff i like about ipython/bpython
12:33	dpritchett_	plus bpython automatically supplies help text for whichever function i'm typing up at the moment
12:33	lnostdal	yes, except line numbers (does that even make sense in a repl?? -- i use buffers/files when stuff grow to, say, 5-6+ lines or so)
12:33	lnostdal	yup, it got that too
12:34	gtrak	dpritchett_, feel free to make an iclojure, I'd use it :-)
12:34	dpritchett_	so i could use slime purely for the repl and then do my actual editing in vim?
12:34	dpritchett_	how... odd
12:34	dpritchett_	i rather like the idea
12:35	lnostdal	might as well switch to emacs full-time then .. it's not that hard .. and elisp is a lisp, so you get more lisp while coding a lisp
12:35	dpritchett_	woah woah woah
12:35	lnostdal	:>
12:35	gtrak	yo dog, i heard you like lisp, so I gave you a lisp so you can lisp while you lisp
12:35	dpritchett_	I ask for a slightly juiced up cljoure repl and you jump to "might as well switch to emacs full time"
12:35	dpritchett_	cmon now
12:36	dpritchett_	Maybe if I were a full time clojure programmer I would be interested, but vim has served me faithfully on many different stacks so far
12:36	lnostdal	shrug .. perhaps there are some middle ground type thing out there, i don't know
12:36	dpritchett_	i've tried viper etc. and they didn't feel right
12:36	dpritchett_	i guess i could always try again
12:36	dpritchett_	i really like the simplicity of great repl + unconnected text editor
12:37	dnolen	dpritchett_: have you you looked at clooj? It the closest thing to something like Python's IDLE. It needs some work as far as syntax highlighting, but it has a lot of interesting features already.
12:37	dpritchett_	thank you dnolen, i will check it out
12:37	dpritchett_	IDLE is a big part of why i got into python to begin with
12:37	dpritchett_	its a great environment for playing around in
12:37	dnolen	dpritchett_: clooj definitely fits the bill of staying of your way.
12:38	dpritchett_	looks like clooj is swing-based though, i do my clojure work on linux VMs and my desktop here is win7
12:38	dpritchett_	i foresee headaches down that path
12:38	gtrak	why's everyone hate on swing all the time?
12:38	dnolen	dpritchett_: like it not understanding Linux/Windows conventions?
12:38	dpritchett_	i might try to run clooj on windows jsut to see how i like it
12:38	dpritchett_	yeah i would expect some of that dnolen
12:39	dpritchett_	I have nothing against swing gtrak, i just don't have a trustworthy java-speaking GUI handy
12:39	Bronsa	dpritchett_: http://i.imgur.com/eFaUa.png
12:39	Bronsa	isnt this tempting enough? :>
12:39	gtrak	dpritchett_, like netbeans or idea? those are in swing
12:40	dpritchett_	ok, i guess i failed to accurately frame my problem
12:40	dpritchett_	i haev a tiny VPS, i work via putty and tmux, i like REPLs and vim
12:40	gtrak	dpritchett_, ah, you should do it locally and upload instead :-)
12:41	dnolen	dpritchett_: why not vimclojure or slimv then?
12:41	dpritchett_	i havent really been back to vimclojure in a year or so, i remember trying 1.2 when it was finally released and it didn't sit well with me
12:41	dpritchett_	i haev never gotten slimv up and running, that's another thing to try
12:41	dpritchett_	I'm much better at this sort of thing than I was last October when vimclojure 1.2 came out
12:42	dpritchett_	perhaps I should just try everything one more time
12:42	dpritchett_	i still like having standalone repls
12:42	dpritchett_	i have built entire django sites with ipython in one pane and vim in the other this year
12:42	dpritchett_	and i really enjoyed that dynamic
12:43	dpritchett_	slimv looks al right though
12:43	dnolen	dpritchett_: I never liked the Python REPL experience.
12:43	gtrak	ipython was pretty nice, but eventually I just used eclipse for it all
12:43	dpritchett_	dnolen: Have you tried many of them? I'll admit that none of the python repls I've used were perfect
12:43	dnolen	reloading changes module w/ restarting everything was always a PITA
12:44	kzar	raek: OK I've made those changes, here's the test that fails https://github.com/kzar/ring-hmac-check/blob/master/test/ring/middleware/hmac_check_test.clj . I posted the full project this time - it's harder to post a small snippet after the refactor
12:44	dpritchett_	that is true, i hacked some init scripts into my .ipythonrc to reload my hottest modules with three-letter aliases
12:44	dpritchett_	and that only worked because my app was still fairly small
12:44	dpritchett_	and there are times when stopping and restarting the repl is necessary because i can't get rid of all of the objects from the last version of my class
12:44	dpritchett_	I can't fault you for finding that awkward; it is
12:45	dpritchett_	slimv seems like a good way to have my cake and eat it too
12:45	dpritchett_	considering how much mindshare swank has in this community
12:46	gtrak	it took me like 2 days to get effective on emacs and slime, but I don't have a vim habit to compare my effectiveness to
12:47	dpritchett_	i could certainly force myself to use emacs but it just seems silly to have my editor forced on me by a programming languages
12:47	dpritchett_	^language
12:47	dpritchett_	next thing you know people will be claiming they enjoy working in eclipse :P
12:48	dpritchett_	i'm rather susceptible to RSI which is another reason i avoid emacs... all of those key chords really hurt me after a day or two
12:48	kzar	dpritchett_: There are a few articles about emacs + rsi about
12:48	gtrak	dpritchett_, first thing is to switch caps-lock to be your ctrl
12:48	dpritchett_	There'd have to be, eh?
12:49	kzar	dpritchett_: I don't think there's causality. This book is worth reading by the way: http://www.amazon.co.uk/Mind-Body-Prescription-John-Sarno/dp/0446675156/
12:49	dpritchett_	I force myself to always use the opposite shift key from whichever key i'm modifying so that i don't have to hit two simultaneous keys with fingers ont he same hand
12:49	dpritchett_	i guess i need a right-capslock too
12:49	dpritchett_	kzar to be fair i was a rather avid gamer back then
12:49	jkkramer	or foot pedals ;)
12:49	dpritchett_	many causes to be sure
12:50	dnolen	dpritchett_: I've been using Emacs for 7 years now, I hate doing any long Emacs sessions w/o my Kinesis.
12:50	dpritchett_	is that the thing technomancy strapped to his pants?
12:51	kzar	dnolen: You should read that book as well, it's less fluffy than it sounds
12:52	gtrak	dnolen, which kinesis?
12:52	dnolen	Kinesis Advantage
12:52	gtrak	so sexy, I haven't tried a keyboard like that yet
12:52	dpritchett_	i had one of those MS split keyboards a decade ago and i loved it
12:52	gtrak	but i guess if this is my job for 50 years it's worth it
12:52	dnolen	gtrak: the best part is that it lets you leverage your strongest fingers - your thumbs.
12:53	dpritchett_	+1 gtrak, i was just weighing the benefits of bringing my own $300 KB to work
12:53	dnolen	I've had my Kinesis for 5 years now. Rock solid.
12:53	dnolen	can't say that for much tech I own.
12:55	dpritchett_	Do you work from home?
12:55	dpritchett_	It'd be sad having a sweet keyboard at home or at the office and then not having one one the other side
12:57	dnolen	dpritchett_: At one point I had two, work/home, but I gave my other to my brother who was having RSI trouble.
12:59	rbranson	i've taught myself how to type with my arms/hands at natural angles instead of bent, works great, even on a $10 keyboard
13:00	dpritchett_	What, like you have your arms nearly fully extended?
13:00	gtrak	you don't rest your palms?
13:00	dpritchett_	I adjust the pitch of my keyboard regularly and slide my KB forward and back regularly to find a good spot
13:00	dpritchett_	sometimes I rest a wrist on a makeshift support
13:01	dpritchett_	Oh I forgot rbranson has a sit/stand desk
13:01	dpritchett_	That'd help
13:01	rbranson	instead of bending at the wrist and aligning my fingers with the home keys
13:01	rbranson	my fingers do a little bit more work and my wrists stay straight
13:02	rbranson	i was using natural keyboards for like 10 years though, don't feel much of a difference either way
13:03	raek	7away
13:03	raek	eh.
13:05	dpritchett_	Has anyone here played with clansi? Looks like a neat way to add colorization to any old command line clojure tool
13:05	dpritchett_	like say `cake repl`
13:05	hiredman	I have
13:05	dpritchett_	have you hooked it into a repl? the clansi source is only 130 lines of clojure, i just need to figureo ut how to hook into it
13:05	fdaoud	dpritchett_: just catching up.. fwiw I use vim and regular command-line repl, works fine. don't like emacs.
13:05	hiredman	and I started to play with a repl that uses clojure.pprint and clansi for printing
13:06	dpritchett_	that sounds awesome hiredman
13:06	hiredman	shrug
13:06	dpritchett_	How'd you set up that repl?
13:06	dpritchett_	Just runtime invocations of pprint and clansi at startup?
13:06	dpritchett_	thanks fdaoud, that sounds like where i'm headed
13:06	dpritchett_	I have probably spent more time on clojure tools than on clojure programming... no one to blame but myself
13:07	hiredman	I can tell
13:07	dpritchett_	maybe i should be a hardcore greenscreen guy instead of even trying to make myself happy with tools
13:08	gtrak	dpritchett_, yea focus on punchcards first
13:09	dpritchett_	most of the folks in my office are greybeards with punchcards and flowchart stencils and other cool stuff
13:09	dpritchett_	i learn a lot from them
13:14	crazyFox	is there a naming convention for types defined with defrecord (or deftype or protocol for that matter)? Like, would i write (defrecord my-record-type ...) or rather (defrecord MyRecordType ...)?
13:15	dnolen	crazyFox: people tends to use Java naming conventions for protocols/types/records
13:17	crazyFox	dnolen: ok. thank you.
13:20	crazyFox	dnolen: does this also apply to the field/method names?
13:26	jbwiv	I'm trying to build clojure-contrib and getting this output from mvn package (https://gist.github.com/60301f4a98dc666d8a4b). Any ideas how to solve?
13:28	dnolen	crazyFox: no fields and protocols fns should follow normal Lisp naming conventions.
13:32	dnolen	jbwiv: looks like your trying to build with 1.3.0, that probably won't work since some of those libs are not 1.3.0 compatible.
13:35	dnolen	jbwiv: you could probably make that work by building w/o tests.
13:37	jbwiv	dnolen, ah, ok. what should I be using? 1.2?
13:37	jbwiv	sorry...clojure newb here
13:38	dnolen	jbwiv: don't use monolithic contrib, it's deprecated.
13:38	dnolen	jbwiv: the most useful bits were already incorporated into 1.2.0
13:38	jbwiv	dnolen, what's the best way of getting access to sql?
13:38	jbwiv	thought that was in clojure-contrib
13:38	dnolen	jbwiv: https://github.com/clojure/java.jdbc
13:39	jbwiv	dnolen, ok, thanks!
13:42	dnolen	no if only someone would implement all the Clojure persistent datastructures in C++ we could use them from V8, SpiderMonkey, JavaScriptCore ...
13:42	dnolen	s/no/now
13:42	seancorfield	org.clojure/java.jdbc "0.0.6"
13:43	gtrak	dnolen, I think you need GC for this, yes?
13:43	seancorfield	i need to push that up to 0.1.0 but there's a ticket i want to get in before then (resolving duplicate column names in a join)
13:43	gtrak	sounds like a huge pain in the ass without GC
13:46	gtrak	I guess you can do smart pointers, but no circular references right?
13:51	dnolen	gtrak: seems surmountable, http://code.google.com/p/jscocoa/wiki/GarbageCollection. I believe V8 has similar guidelines.
13:52	gcv	Is there any documentation on what exactly changed with exception handling between 1.2.1 and 1.3.0? I see some weird behavior where 1.3.0 chains an exception thrown from a Java library to a RuntimeException, which makes catch specificity impossible.
13:52	gtrak	I've done a little C++ and i'd be very very afraid
13:53	dnolen	gtrak: well Node.js isn't.
13:53	gtrak	dnolen, you mean they have structurally shared data structures in node?
13:54	dnolen	gtrak: no I mean they probably have to create C++ objects which are subject to GC there as well.
13:54	gtrak	sure, java has a GC, too, but the data structures are written in java, not C++
13:54	jbwiv	so I dropped back to 1.2 to compile clojure-contrib, because apparently the code I'm using requires that version of sql. I'm now getting this when building: https://gist.github.com/298cb5d4f6ef2a150b1a. Any ideas?
13:55	kharrington1	pretty sure the best way to do clojure data structures in C++ is to do it on top of Boost, then you can get GC for "free"
13:55	gtrak	kharrington1, not GC, reference-counted smart pointers
13:56	gtrak	it's a can of worms
13:57	kharrington1	mmm
13:57	dnolen	jbwiv: if you're using 1.2.1 there's no reason to build contrib from scratch.
13:57	jbwiv	dnolen, so under 1.2.1, contrib is already there?
13:58	dnolen	jbwiv: no it's a separate jar, but just drop into your classpath
14:01	jbwiv	dnolen, so, where would I find this jar? I don't see it after building 1.2.1
14:02	dnolen	jbwiv: contrib is not a part of 1.2.1. it's separate jar. And why do you need to build 1.2.1? It's also available for download.
14:03	jbwiv	dnolen, ok, thanks. I typically like to download source zips and build them myself, but in the interest of time I'll just download the bins
14:04	dnolen	jbwiv: btw, what code are you using that requires the older SQL lib?
14:05	jbwiv	dnolen, something another dev here wrote. may not require it, but I was getting "Could not locate clojure/contrib/sql__init.class or clojure/contrib/sql.clj on classpath" even after building java.jdbc
14:05	jbwiv	so I presume it requires the older lib, but it is just as likely to be my setup
14:05	dnolen	jbwiv: because the namespace it's pointing as is wrong now.
14:05	dnolen	s/as/at
14:05	jbwiv	dnolen, ah, k
14:06	dnolen	jbwiv: it's probably worth considering the minor hassle moving to java.jdbc since it's actively maintained.
14:07	jbwiv	dnolen, ok, thanks. in this particular case, I'm re-writing the code to ruby...but if it were staying clojure, i'd definitely follow your advice. I really appreciate the help
14:14	gcv	Could someone fill me in on the exact implications of the removal of checked exceptions in 1.3.0?
14:14	gcv	Does every checked exception thrown by Java code get chained to a RuntimeException?
14:21	dnolen	gcv: checked exception are a fiction from what I understand, http://tapestryjava.blogspot.com/2011/05/tragedy-of-checked-exceptions.html
14:24	gcv	dnolen: I totally agree that checked exceptions are broken by design, but something changed in Clojure between 1.2.1 and 1.3.0 — I call a Java method which defines a checked exception, but I can't catch it in Clojure 1.3.0. This used to work in 1.2.1. When I looked at the thrown exception closely, it turned into a RuntimeException with the original exception chained to it.
14:24	gcv	so I was hoping to understand the changes in 1.3.0 a little better — the changedoc has no details at all.
14:25	TimMc	I've heard some other people complaining about this too.
14:25	dnolen	gcv: checked exceptions used to be supported I think? that's since been removed.
14:25	technomancy	"Checked exceptions are the TSA of software. So what if there's no evidence we're safer, at least we're inconvenienced." - http://twitter.com/#!/posco/status/116535658515939328
14:26	gcv	guys, I'm not disagreeing that checked exceptions suck :)
14:26	TimMc	dnolen: And are instead wrapped in RTEs?
14:26	gcv	I'm just trying to understand why catching an exception declared as checked in external Java code doesn't work.
14:26	gcv	(anymore)
14:26	dnolen	TimMc: I think so I don't understand the details of the change.
14:28	TimMc	Neither does anyone, as far as I can tell.
14:29	TimMc	"Clojure does not throw checked exceptions" <- OK...
14:29	TimMc	gcv: Experimentally, what does clojure throw?
14:30	dnolen	gcv: bring the issue up on the mailing list, I'm sure someone has experienced this before.
14:30	gcv	dnolen: will do
14:31	TimMc	gcv: If you (throw (IllegalAccessException. "foo")), what does Java catch?
14:32	gcv	TimMc: no idea, I don't have any code which calls Clojure from Java
14:35	pjstadig	gcv: i think all exceptions are wrapped now into RuntimeExceptions
14:36	amalloy	dnolen: i read your link, followed one of the comments to http://james-iry.blogspot.com/2010/08/on-removing-java-checked-exceptions-by.html - pretty remarkable abuse of the type system to get around checked exceptions
14:37	gcv	pjstadig: I hope not, because that means typed catch is now useless, which weakens Clojure's Java interop story.
14:37	hiredman	pjstadig: are you sure about that?
14:39	hiredman	I believe that is only if you are using reflection
14:39	pjstadig	no i'm not sure
14:39	pjstadig	i'm likely wrong
14:39	hiredman	clojure's reflector rethrows as runtime exceptions to avoid having to declare thrown exceptions everywhere
14:40	hiredman	(in 1.3)
14:40	crazyFox	is there any other way to resolve cyclic load dependencies than putting code into a different file?
14:40	pjstadig	https://github.com/clojure/clojure/commit/8fda34e4c77cac079b711da59d5fe49b74605553
14:41	crazyFox	pjstadig: u r talking to hiredman?
14:41	pjstadig	yeah i mean checked exceptions are a fiction (like has been mentioned) there's no reason to wrap them in the bytecode that clojure emits
14:41	hiredman	right, but non-reflective calls still compile to normal java method calls, so exceptions bubbling out will not be caught and rethrown by the reflector
14:42	pjstadig	*wrap them in a RuntimeException
14:42	hiredman	well, the reason is the reflector is written in java, so it has to abide by the fiction
14:42	hiredman	and if the reflector has to then so does the compiler etc
14:42	pjstadig	yeah
14:42	hiredman	it ripples through the java codebase
14:43	pjstadig	right so i suppose anytime you cross that java barrier with clojure code, an exception thrown will get wrapped in Clojure's java code
14:43	amalloy	i guess the reflector couldn't be written in clojure because there are pieces of the runtime that depend on it? that's too bad
14:45	crazyFox	can declare be used to resolve cyclic load depedencies?
14:45	dnolen	crazyFox: no, you have to change your project structure
14:45	raek	crazyFox: between vars in a namespace, yes. between namespaces, no.
14:46	crazyFox	dealing with different namespaces so - ok ill change my files. not an ideal solution though.
14:46	duck1123	fixing load dependencies are a pain, but your project will be better for it anyway
14:47	crazyFox	duck1123: what do u mean by "you project will be better for it anyway" (- maybe i dont quite understand bcause im not a native entlish speaker)
14:47	raek	this is the result of some design decisions that makes it easy for macros to work the way they do (they can use any functions defined "so far" to do their job)
14:48	crazyFox	yeah i see
14:48	duck1123	I'm just saying that if you have load dependency issues, it usually means you haven't thought out the structure of your project enough
14:49	duck1123	I've moved so many functions around just to stay one step ahead of that issue, and my project is better designed because of that
14:50	amalloy	(inc duck1123)
14:50	lazybot	⟹ 1
14:51	amalloy	cyclic load dependencies recently gave me the choice of (a) dumping even more code into a "utils" file, or (b) breaking up that file into reasonable pieces
14:51	amalloy	chose b, but the important thing was there's no (c) ignore the fact that your utils file is a mess
14:57	crazyFox	duck1123: well i make my project structure up as i go along. so, yes its not thought out well so far :D
14:57	ibdknox	Noir made it into the trending repo list on github :) https://github.com/explore
14:58	dnolen	ibdknox: nice!
14:58	ibdknox	I was actually very surprised people cared lol
14:59	crazyFox	anyone aware of an editor thats has convenient functionality to help with cyclic load dependencies?
14:59	gtrak	crazyFox, you mean like a compiler?
15:00	crazyFox	gtrak: actually an ide
15:00	gtrak	if you try to compile your files and there's a cyclic dependency, it'll break
15:00	gtrak	isn't that enough?
15:02	kij	openjdk is broken for the entire clojure or just clojsurescript ?
15:02	crazyFox	it can be tedious. you could do nice things in an ide that automatically discovers cyclic load deps and e.g. suggest a solution (like which definitions have to move)
15:02	technomancy	kij: just clojurescript
15:02	kij	thanks.
15:03	TimMc	crazyFox: If yo uneed IDE support for that, you probably need to work on your abstraction boundaries.
15:03	dnolen	kij: OpenJDK should be fixed for ClojureScript now.
15:03	ibdknox	dnolen: when did that happen?
15:03	dnolen	ibdknox: today
15:03	ibdknox	dnolen: awesome :D
15:04	dnolen	ClojureScript is now using a newer version of Rhino, Brenton has dropped some serious awesomeness.
15:04	gtrak	crazyFox, if you don't write a ton of code all at once, you'll only have like one cyclic dependency to deal with at a time
15:04	upwardindex	Does "leon upgrade" also upgrade plugins?
15:04	ibdknox	upwardindex, no
15:04	upwardindex	"lein upgrade**
15:04	dnolen	ibdknox: print doesn't trigger the print dialog, JS stack traces from Chrome / FF
15:04	upwardindex	ibdknox: How do you upgrade plugins?
15:04	crazyFox	TimMc: well as i said - i make it up as i go along. dont have a whole lot of experience so far neither with clojure nor "bigger" projects
15:05	ibdknox	upwardindex: uninstall the old one and install the new version
15:05	ibdknox	dnolen: yeah I saw those :)
15:05	upwardindex	ibdknox: Ok, thank you!
15:05	gtrak	crazyFox, i honestly can't think of a case where it'd be so bad that a tool is worth it
15:06	gtrak	now a general refactoring tool, I could get behind that
15:06	gtrak	like a 'change all references to this function when i move it' sort of thing
15:07	technomancy	there's a refactorer for Emacs
15:07	technomancy	it can do things like "arrowize this form"
15:08	crazyFox	gtrak: yea but still. imagine an ide that tells u "definitions a b c are part of a cyclic load dep" - then u choose one of em, choose a file and the ide moves it automatically and adjusts all the ns forms. i think it wd be NICE
15:08	gtrak	crazyFox, that's the point. I'm imagining it, and it looks silly
15:09	gtrak	plus, it's not terribly hard to avoid writing them in the first place
15:09	ibdknox	crazyFox: out of curiosity, what language are you coming from?
15:10	crazyFox	idbknox: mainly c/java
15:10	gtrak	use a forward declaration as a last resort
15:10	ibdknox	crazyFox: using primarily eclipse?
15:11	crazyFox	gtrak: it might seem silly to some ppl, but it wd still help many i guess
15:11	crazyFox	ibdknox: lately yes
15:11	ibdknox	gtrak: I think it honestly depends on how you "grew up" programming
15:11	gtrak	yea... I guess I had some C and C++ before java
15:13	gtrak	you shouldn't write code like that in java, either, even though it lets you
15:13	ibdknox	sure, but what's to teach you not to? lol
15:13	crazyFox	"code like that"?
15:13	gtrak	circularly referenced stuff
15:14	gtrak	types that know about each other, etc..
15:15	gtrak	can't think of a case where that is the right choice
15:16	gtrak	ibdknox, lots of good programming books out there
15:16	ibdknox	gtrak: I think with types it's more obvious than it is with functions
15:16	crazyFox	gtrak: agree. i just - have a lot to learn and so i just started my project and ill see where it takes me
15:18	gtrak	ibdknox, with functions, wouldn't you be doing a mutual recursion by accident anyway?
15:18	ibdknox	gtrak: groups of functions
15:18	gtrak	you write a function that calls a function that calls you? seems like a stackoverflow
15:19	ibdknox	gtrak: sorry didn't mean inside of a function
15:19	ibdknox	I meant with a group of functions
15:19	gtrak	like how?
15:20	ibdknox	There are certainly logical groupings of functions where two files would depend on eachother
15:21	crazyFox	gtrak: over time u add on function after the other. u think "this one belongs to ... THIS module". then u compile and see - "oops, doesnt work this way"
15:22	gtrak	sure.. then you separate something out or add a forward declaration and it's fixed
15:22	amalloy	gtrak``: (defn a/foo [] (b/bar)) (defn a/baz [] 1) (defn b/bar [] (a/baz))
15:22	TimMc	crazyFox: Often the answer is to have a common utility namespace or such.
15:23	ibdknox	yeah. The fixes are usually straightforward :)
15:23	crazyFox	yea. and an ide could help me with this.
15:23	ibdknox	crazyFox: well, I would argue that it's better to not have the IDE in this instance
15:23	ibdknox	there's value in learning how to think in a manner that organizes things cleanly
15:24	technomancy	crazyFox: slamhound does that, sorta
15:24	technomancy	it won't untangle circularity, but it will fix your ns forms to reflect what's on disk
15:24	gtrak	you could also separate the names from implementations, which is what you do in C with .h's and .c's
15:24	TimMc	gtrak: ew
15:24	gtrak	well, you could :-)
15:24	ibdknox	haha
15:24	TimMc	ew
15:25	crazyFox	technomancy: sorta :D. in my ideal ide this wd be done more automatically
15:25	technomancy	crazyFox: should be possible to trigger a recomputing of the ns form when you encounter a compilation error.
15:26	gtrak	you could have an impl namespace that imports the vars
15:26	technomancy	just a matter of someone hooking up the pieces
15:26	gtrak	then binds itself onto the vars
15:26	kzar	ibdknox: Is there a way to access request and response header from inside defpage?
15:26	crazyFox	technomancy: do u plan to add functionality like "lein slamhound" -> slamhound replaces all the ns forms in all project files?
15:26	kzar	ibdknox: Or just the full request / response
15:27	ibdknox	kzar: you can set response headers by returning a map
15:27	technomancy	crazyFox: I have that implemented in elisp. it's really hard to implement in clojure.
15:27	kzar	ibdknox: Sometimes I want to do things myself, middleware is the only way I found but it seems silly to use sometimes
15:27	ibdknox	kzar, can you give me a case you think is silly?
15:27	technomancy	elisp lets you simultaneously treat code structurally and also as a string, but clojure doesn't
15:28	crazyFox	gtrak: i dont understand. can u elaborate?
15:28	gtrak	it's a hair-brained scheme I just pulled out of my ass :-)
15:28	crazyFox	oh. well then...
15:29	kzar	ibdknox: Well if there's a one off reason to access something in the header / request in general. Or if likewise for response it seems like overkil to make middleware. Often it's fine don't get me wrong
15:29	gtrak	you could have a namespace of function implementations that binds the implementations to the declaration vars on init
15:29	gtrak	that way you'd have all the names of functions defined, but implementations separated
15:30	gtrak	sort of like a dependency injection i guess
15:30	amalloy	ibdknox, kzar: perhaps you want to reimplement whatismyip.com
15:30	ibdknox	kzar: well I'm not sure what you're trying to do so it's a little harder for me to suggest the best method. There are a few different ways to get at the request. Pre-routes have full access. There's also the noir.request/ring-request function as of 1.2.0. Supplying response headers is just a matter of supplying a map as the return of your defpage. Take a look at what noir.response does
15:30	amalloy	one of your pages needs the client's IP address, but the rest don't
15:30	gtrak	I dont' recommend it as a a methodology, but just saying, it could be done
15:31	ibdknox	amalloy: you can grab that from noir.request now :)
15:31	amalloy	ibdknox: i assume you do this by adding a middleware that binds a var to the request?
15:31	gtrak	if all the names are declared in separate namespaces and decoupled from implementations, then you can't have circular references
15:31	crazyFox	gtrak: u mean like a (declare <list all your functions>) in a separat (like header) file that u use in all the other files?
15:32	gtrak	yea something like that
15:32	gtrak	a terrible terrible idea
15:32	redinger	ibdknox: When are you coming back to Durham to present the Clojure meetup with all this Noir hotness? :)
15:32	crazyFox	gtrak: y it does discourage a proper (layer) design
15:33	amalloy	gtrak: an idea that doesn't even work. you can't declare outside of your current namespace
15:33	crazyFox	thats even better. one less possibility to mess ^^
15:33	amalloy	and it's even more disgusting to write header.clj as: (ns foo) (declare myfn) (ns bar) (declare more-fns)
15:34	jweiss	,(map #(+ %1 ^{:my :meta}) [1 2 3])
15:34	clojurebot	#<ExecutionException java.util.concurrent.ExecutionException: java.lang.RuntimeException: Unmatched delimiter: )>
15:34	jweiss	that is a misleading error message to me ^
15:34	ibdknox	amalloy: yeah
15:34	jweiss	(i originally hit it copying metadata that already had a ^ into a with-meta form)
15:35	ibdknox	redinger: haha I'm not sure yet. :)
15:35	crazyFox	amalloy: i wonder how far bad style can be taken
15:36	kzar	ibdknox: OK, so if I return a map instead of the body the map I return is the full response that's sent?
15:36	gtrak	amalloy, I think I said declare in one ns, bind in another
15:36	ibdknox	kzar, well it's parts of it
15:37	ibdknox	kzar, other parts will get added to it by middleware as it comes back out
15:37	gtrak	and i said it's totally a bad idea
15:37	ibdknox	kzar, but yes
15:37	kzar	ibdknox: OK gotya, thanks
15:37	hiredman	jweiss: misleading in what way? given the way the reader works it makes sense
15:37	jweiss	hiredman: should users of clojure know how the internals of the reader work?
15:37	hiredman	yes
15:38	crazyFox	gtrak: i wont quote you on it
15:38	ibdknox	lol
15:38	ibdknox	no
15:38	hiredman	absolutely
15:38	ibdknox	that's the opposite of pragmatism
15:38	kzar	ibdknox: BTW IMHO I think the new way to access a request is a bit ugly, might be better to just pass it as a parameter. Like I say it's just my opinion. Otherwise I really like Noir
15:38	ibdknox	kzar: I think it is by far the 10% case, and so it was designed as such
15:39	kzar	ibdknox: Hmm fair enough, at least there's a way when you need it
15:39	ibdknox	kzar: yep :)
15:39	kzar	ibdknox: Things like rails really limit you there, when you're doing anything more advanced. Like I was making a REST API and simple things where so hard without full access to request and response
15:39	gtrak	amalloy, yes it has downsides, but circular references is not one of them :-)
15:40	jweiss	hiredman: while i agree it helps to know the internals, that should not be a requirement
15:40	ibdknox	kzar: yeah, I agree. I think the default should be plain and simple, but you should be able to do the advanced stuff without needing to do crazy things :)
15:40	jweiss	no consumer of a library is expected to know implementation details, just the contract
15:41	amalloy	jweiss: the contract is that ^{...} forms need to be attached to something?
15:41	hiredman	jweiss: thats one of the huge benefits of clojure, the implementation, compared to most other languages is very simple
15:42	jweiss	amalloy: yes, i know that, but i accidentally didn't do it and the error message was wrong
15:42	hiredman	I would argue you should know the details of any language you use, and it's easier to do so with clojure
15:43	jweiss	ok, if the idea is that people want the reader to be simple and not give better error messages, i won't argue. but the error message here is wrong
15:43	jweiss	the delimiter is not unmatched.
15:43	hiredman	it is
15:43	hiredman	^{} causes it to be unmatched
15:44	jweiss	it is if you know the reader consumes the paren with the meta
15:44	jweiss	and your mistake wasn't putting the ^ in there by accident in the first place
15:44	srid	how do I set .innerHTML in clojurescript?
15:44	srid	or set any attr
15:45	amalloy	set!
15:46	dnolen	srid: you probably want to use something from google closure to do that if you're worried about cross browser issues.
15:50	crazyFox	when i have a (defrecord MyDataType ...) in one file, do i :use, :import or :require it in another file?
15:51	crazyFox	well i can actually ask slamhound
15:52	srid	(.set! (. innerHTML content) (. innerHTML tab-content))
15:52	srid	-> TypeError: Cannot read property 'content' of undefined
15:52	gtrak	if you require, you'll have to qualify it by the alias
15:52	gtrak	crazyFox^
15:52	hiredman	gtrak: no
15:52	amalloy	srid: args to . are backwards
15:52	gfredericks	where did clojure.contrib.json go for 1.3? I thought there was a clojure.data.json ns, but apparently not
15:52	gfredericks	googling doesn't help
15:52	dnolen	crazyFox: default ctor fns are provided for you in 1.3.0, in 1.2.0 you provide your own.
15:52	hiredman	crazyFox: in 1.3 defrecord creates factory fns and you should use them
15:53	srid	amalloy: sure? http://java.dzone.com/articles/adventures-clojurescript says otherwise
15:53	srid	looks like I should not have a space after .
15:53	srid	with that, I get: TypeError: Object Nothing has no method 'set_BANG_'
15:53	hiredman	gfredericks: use cheshire
15:53	hiredman	https://github.com/dakrone/cheshire
15:54	gfredericks	hiredman: okay I will try, thanks
15:54	srid	oops, I used .set! instead of set!
15:55	duck1123	gfredericks: data.json is at https://github.com/clojure/data.json
15:55	gfredericks	duck1123: thanks
15:55	kzar	ibdknox: Oh, I meant to ask you. How does the mapping to static files in resources/public work?
15:55	srid	dnolen: not sure if closure has a wrapper for setting innerHTML
15:56	dnolen	srid: it does
15:56	kzar	ibdknox: I couldn't figure it out, but I spotted a bug. If you're filename has spaces in they are escaped in a url to be %20, but then the translation doesn't decode it back to a space
15:57	kzar	ibdknox: So it tries to open a filename of "dumb%20example.jpg" instead of "dumb example.jpg"
15:57	hiredman	cheshire is far superior to contrib's json handling
15:57	srid	dnolen: it has .setTextContent, but none for setting HTML content from what I can tell by reading http://closure-library.googlecode.com/svn/docs/closure_goog_dom_dom.js.html
15:58	hiredman	round tripping through cheshire is faster than round tripping through clojure's printer/reader
15:58	hiredman	(which you can see from the nice table in the cheshire readme, data.json doesn't give any benchmarks, which is telling)
15:58	dnolen	srid: goog.dom.htmlToDocumentFragment
15:58	gfredericks	hiredman: in my case, development time is more important
15:59	duck1123	I might have to give cheshire a look then
15:59	srid	dnolen: hmm, ok. actually, it looks like I need to "move" a html chunk instead of copying it, else all the id's in it would become meaningless.
15:59	srid	all examples for tabbar does 'copy', but not 'move'
16:01	ibdknox	kzar: is that with 1.2.0
16:04	dakrone	hiredman: yea, I need to add benchmarks for data.json comparisons
16:05	jbwiv	anyone here using intellij with la clojure?
16:05	crazyFox	dnolen, hiredman: sorry, i dont understand yet. Using clojure 1.3.0. file A contains (defrecord MyDataType), file B creates instances of that type (via (MyDataType. ...)), file C contains a function that deals with those instances. How do i go about?
16:05	jbwiv	I'm trying to inspect variable values in intellij when I hit a breakpoint, but watches don't seem to work
16:07	dnolen	crazyFox: import the ctor fns for those types into file C
16:09	crazyFox	dnolen: what are ctor functions? what do they look like? or rather: is this explained somewhere?
16:10	kzar	ibdknox: 1.1.1-SNAPSHOT
16:10	jbwiv	what's the preferred way to debug clojure code? do any ides support debugging well?
16:11	kzar	ibdknox: I'll upgrade and see what happens
16:13	kzar	ibdknox: Yea, it happens with 1.2 as well
16:14	S11001001	crazyFox: http://dev.clojure.org/display/design/defrecord+improvements
16:16	Raynes	ibdknox: I think you're right. My tolerance for Sublime's poor indention is quickly running out.
16:20	kzar	Is there something like re-seq that returns the context, the bits in between the matches? Or maybe something that returns the position of the matches
16:21	kzar	I need to do a regexp replace, but by giving a function the match as a parameter instead of replacing with a static string
16:21	amalloy	kzar: clojure.string/replace takes a function
16:22	amalloy	&(clojure.string/replace "test" #"s" (fn [m] (str (char (inc (int (first m)))))))
16:22	lazybot	⇒ "tett"
16:22	kzar	amalloy: Schweet, that's exactly what I needed
16:23	amalloy	worst function definition ever there, btw. doing arithmetic on characters is such a pain :P
16:27	crazyFox	S11001001: thank you! i will read and digest :)
16:34	pyr	so I have a locally deploy maven artifact
16:34	pyr	in $HOME/.m2/repository
16:35	pyr	when referencing it in my project.clj
16:35	pyr	it seems as though there's still an attempt to download from http repositories
16:36	pyr	is that because i did not reference the artifact correctly
16:36	pyr	or is it expected
16:42	duck1123	pyr: if it's a SNAPSHOT, it'll try to update it from time to time
16:53	kzar	How can I make regexp match over new lines?
16:54	TimMc	&(clojure.string/replace "test" #"s" (comp str char inc int first))
16:54	lazybot	⇒ "tett"
16:54	TimMc	kzar: DOTALL or MULTILINE or something
16:55	TimMc	You have to use Pattern directly to do that, I think.
16:56	amalloy	$javadoc java.util.regex.Pattern
16:56	lazybot	http://download.oracle.com/javase/6/docs/api/java/util/regex/Pattern.html
16:58	kzar	TimMc: If the regexp literal is that class could I call a method on it to enable multi line matches?
16:58	pyr	duck1123: pb came from me
16:58	TimMc	kzar: Looks to be immutable. There seems to be a way to enable those flags inline, though.
16:59	TimMc	,(re-seq #"_._" "a_\n_b")
16:59	clojurebot	nil
16:59	TimMc	,(re-seq #"(?s:_._)" "a_\n_b")
16:59	clojurebot	("_\n_")
17:00	TimMc	^ It's not pretty, but that's how to use Clojure's RE reader syntax and still get those flags.
17:00	TimMc	Or, do (Pattern. (str #"...") flags...)
17:00	TimMc	Pattern/compile, rather
17:01	kzar	TimMc: I don't understand the ?s example, shouldn't it be ?m ?
17:02	TimMc	DOTALL allows . to match \n
17:02	TimMc	MULTILINE makes ^ and $ match line boundaries.
17:03	duck1123	it's a shame re-pattern doesn't accept flags
17:04	kzar	TimMc: Sweet I got that working, thanks
17:06	TimMc	kzar: Word of warning: Pattern might not be smart enough to optimize that "global override" into a global flag.
17:06	TimMc	So this approach could slow down your regexes, maybe.
17:06	Blafasel	http://pastebin.com/e0Py3ELe - What is leiningen trying to tell me?
17:06	duck1123	Is there a way to access the command history in CCW's repl views?
17:07	ibdknox	Blafasel, looks like a Clojure version mismatch
17:07	TimMc	duck1123: Have you tried C-p and C-n?
17:07	duck1123	Blafasel: Most likely, you have a jar that was compiled against a different version than you're using
17:08	duck1123	TimMc: yep. (once I remembered that esc-p is not alt-p) didn't work
17:08	Blafasel	duck1123: ibdknox: This is a host that has never seen clojure before (micro instance from amazon). Other than the deps it didn't download anything.. Weird.
17:09	duck1123	Blafasel: one of your deps is going to have a bunch of class files in it, that's most likely the problem
17:10	Blafasel	duck1123: Checking. Thanks
17:10	srid	clojurescript's map doesn't work?
17:10	srid	(map #(...) (dom/getChildren (dom/getElement "maintab_content")) <- here the function was not invoked at all
17:11	ibdknox	srid, you're getting a node collection back, which isn't a seqable thing
17:11	srid	ah
17:11	srid	doseq, loop and seq didn't work, for that reason
17:11	TimMc	because mutable?
17:11	chouser	I think you and use primative-seq or something on it
17:11	srid	ibdknox: so how do I "iterate" this node collection?
17:12	ibdknox	srid: use pinot ;)
17:12	srid	later, maybe :-)
17:12	chouser	(prim-seq node-collection 0)
17:12	ibdknox	srid, or more seriously, you can either turn it into a collection using the array methods, implement ISeq on that type, or use what chouser said
17:13	chouser	srid: try prim-seq ^^^
17:14	chouser	(map #(...) (prim-seq (dom/getChildren ...)))
17:14	chouser	(map #(...) (prim-seq (dom/getChildren ...) 0))
17:14	duck1123	also, don't forget that map is lazy
17:14	srid	hmm, prim-seq doesn't see to work
17:14	chouser	:-(
17:15	chouser	I used it a bunch in traction. Not sure why.
17:15	ibdknox	chouser, traction?
17:15	chouser	er, I mean, not sure why it wouldn't work for you.
17:16	chouser	ibdknox: it's not documented or released: https://github.com/Chouser/talk-cljs/blob/master/src/traction.cljs
17:16	Blafasel	duck1123: Weird.. http://pastebin.com/24Cwh9Pe / except for clojure itself there's only the vimclojure server jar that contains .class files. And I'm pretty sure that very same thing worked on my other machine quite well.
17:18	srid	chouser: prim-seq worked with doseq, but not map!
17:18	chouser	srid: are you taking into account that map is lazy?
17:18	srid	facepalm
17:18	srid	,(dec srid)
17:18	clojurebot	#<CompilerException java.lang.RuntimeException: Unable to resolve symbol: srid in this context, compiling:(NO_SOURCE_PATH:0)>
17:26	srid	ibdknox: i should probably take a look at pinox sooner, for right now i have the goog.ui.tabbar code in disparate locations (hiccup code to initialize html template, hiccup css includes, and cljs event handling)
17:29	crazyFox	clojure 1.3.0: how do i import the record constructor functions? i tried (ns ... (:import myproject.myfile.map->MyRecord)), but this gives me a ClassNotFoundException
17:29	brehaut	crazyFox: use or require the namespace that defines it first
17:30	dafra	hi there
17:30	dafra	i've just tried swank, it's clojure 1.2
17:30	dafra	how to use clojure 1.3 ?
17:31	brehaut	crazyFox: secondly, if you can provide utility functions (constructors, predicates etc) instead of requiring a namespace consumer to explicitly depend on the record type itself, thats better
17:33	brehaut	dafra: i know basicly nothing about swank etc, but do you know of clojure-jack-in ?
17:33	dafra	brehaut: i use clojure-jack-in, and it launches a 1.2.1 repl
17:33	crazyFox	brehaut: u mean like (ns ... (:use [myproj.myfile]) (:import myproj.myfile.map->MyRecord))? gives me the same exception
17:34	brehaut	crazyFox: where did you get the arrow from? leave that out
17:34	raek	dafra: change the clojure version in your project.clj file
17:34	raek	and run lein deps in the project dir
17:35	dafra	:dependencies [[org.clojure/clojure "1.3.0"]]
17:35	dafra	raek: but that did't help
17:35	amalloy	brehaut: i think that's correct in 1.3
17:35	brehaut	amalloy: real? huh
17:35	dafra	raek: maybe rebooting emacs ? let's try
17:36	brehaut	completely missed that change
17:36	crazyFox	so what should my ns form look like?
17:36	amalloy	oh, it looks like my previous message got dropped though
17:36	amalloy	crazyFox: it's a plain old clojure function. just require/use it
17:36	crazyFox	ok, ill try that
17:36	amalloy	brehaut: in 1.3 (defrecord Foo []) has the side-effect of defining two additional functions, map->Foo and ->Foo
17:36	srid	why would a lamina channel be automatically closed?
17:36	brehaut	oh right
17:37	raek	dafra: are you running clojure-jack-in while visiting a file or directory in the project?
17:37	dafra	raek: yes
17:37	brehaut	(inc amalloy)
17:37	lazybot	⟹ 16
17:37	raek	dafra: a listing of the files in your lib/ directory would also be helpful
17:37	dafra	raek: lein's lib ?
17:38	crazyFox	amalloy: long journey comes to an end. Thank you!
17:38	raek	dafra: yes, the lib/ in the project directory
17:38	brehaut	crazyFox: sorry for leading you astray
17:39	dafra	raek: clojure 1.2.1 :)
17:39	dafra	raek: thanks, look like i know what to fix !
17:39	raek	and you have no other depencencies?
17:39	crazyFox	brehaut: no problem. after two hours of reading searching trying this was insubstantial
17:39	dafra	raek: not for now, jsut toynig with the ide
17:39	raek	you shouldn't have 1.2.1 there if you specified 1.3.0 in the project.clj file... :/
17:40	raek	dafra: the lib/ directory is populated from what you put in the :dependencies thingy
17:41	duck1123	Do records always have a no-arg constructor? Or is that just in the cases I've been using them for?
17:41	raek	duck1123: they always take one argument per field
17:42	raek	(defrecord Foo [a b c]) must be constructed as (Foo. 1 2 3)
17:42	dafra	raek: I've copied clojure-1.3.0.jar into lib and the repl is 1.3 as expected
17:42	raek	dafra: but you shouldn't have to do that!
17:43	dafra	raek: how to tell lein that I want 1.3 ?
17:43	raek	dafra: are you sure you saved the project.clj file before you ran "lein deps" in the project directory?
17:43	duck1123	raek: Ok, In that case, I need to think about one of my libraries a bit more. :(
17:43	dafra	changing the porject.clj wasnt enough
17:43	dafra	raek: lein deps ?? let's do it
17:44	brehaut	dafra: porject.clj or project.clj?
17:44	amalloy	duck1123: you're on 1.3 now though, right?
17:44	dafra	brehaut: right spelling (created by lein new)
17:45	duck1123	amalloy: yes
17:45	brehaut	dafra: figured it probably was, but wanted to check
17:45	amalloy	duck1123: so you should be able to use (map->MyRecord {}) to get an "empty" version of a map
17:45	amalloy	er, record
17:45	dafra	brehaut: could be, thanks
17:45	raek	dafra: if you change the dependencies you need to run lein deps to "apply the changes"
17:46	dafra	raek: just did lein deps, obviously better than copying by hand :)
17:46	technomancy	raek: (unless you set :checksum-deps true in project.clj or ~/.lein/init.clj)
17:46	raek	dafra: yeah, copying by hand is no fun for depencencies, dependencies of dependencies, dependencies of dependencies of dependencies, etc ;-)
17:48	dafra	raek: see you guys
17:49	raek	dafra: happy hacking :-)
17:56	srid	in clojurescript, (doseq [index [0 1 2 3 4]] ...) -> " doseq requires a vector for its binding" error
17:56	srid	i don't understand the error, as the deseq syntax looks alright
17:58	srid	the doseq comes within a doto macro, if that matters
17:58	ibdknox	that matters
17:59	srid	ah yes
18:13	ibdknox	oh no's! Lazybot is gone
18:14	gtrak	,#(print "OH NO")
18:14	clojurebot	#<sandbox$eval27$fn__28 sandbox$eval27$fn__28@1dec77f>
18:14	gtrak	,(#(print "OH NO"))
18:14	clojurebot	OH NO
18:14	ibdknox	,(print "oh no :p")
18:14	clojurebot	oh no :p
18:14	ibdknox	# is unnecessary :)
18:14	gtrak	duh
18:15	gtrak	we all have moments where we think something is a level less abstract than it is
18:15	Blafasel	The whole machine is down, it seems
18:16	Blafasel	4clojure.com try-clojure.org and the guys that have a bouncer on that machine pinged out.
18:21	ibdknox	haha
18:22	ibdknox	amalloy jinxed it talking about how he hasn't really had any issues with hosting
18:28	wdouglas	Hey all, trying to get (floor(log_2 x)). Didn't want to loop recur so ended up using swap! but I don't really like it. Any better options for this?
18:28	wdouglas	FYI code is just:
18:28	wdouglas	(let
18:28	wdouglas	[a (atom 0)
18:28	wdouglas	b 30]
18:28	wdouglas	(while (> (bit-shift-right b @a) 0)
18:28	wdouglas	(swap! a inc))
18:28	wdouglas	@a)
18:28	Blafasel	What's wrong with Math/*?
18:29	wdouglas	log base e/10
18:29	wdouglas	Unless I missed something
18:29	wdouglas	Or just do the log conversion do you mean?
18:30	Blafasel	Didn't know about the missing log, I meant both
18:31	ibdknox	why didn't you want to use loop/recur?
18:32	ibdknox	it'll be much more efficient than what you have right now
18:32	hiredman	using an atom for that there is dumb
18:32	ibdknox	though I'm with Blafasel, you can do that with normal math functions, can't use?
18:32	ibdknox	you*
18:32	wdouglas	Hrm guess I could time it, was thinking the division is going to make the log base conversion a bit slower.
18:33	Blafasel	wdouglas: Depending on precision: http://everything2.com/title/log2%2528x%2529+%253D+log10%2528x%2529+%252B+ln%2528x%2529
18:33	wdouglas	I just thought the loop/recur was ugly for that
18:33	hiredman	you are wrong
18:33	ibdknox	wdouglas, loop/recur was designed to solve problems like that
18:33	ibdknox	lol
18:34	wdouglas	Blafasel: Oh nice on that top one.
18:34	dabd	how can I construct a ratio from a string? For example from "13/17"
18:34	wdouglas	Well that's the way of it then. loop/recur it is. Thanks all =)
18:36	dabd	the Ratio type does not seem to have a constructor or parser...
18:36	hiredman	,(/ 13 17)
18:36	clojurebot	13/17
18:37	ibdknox	dabd: split the string, map Integer. over it and then do what hiredman did
18:37	dabd	ok thx!
18:38	dabd	But why there is a type without a constructor?
18:38	brehaut	dabd: there is a constructor, its /
18:38	dabd	well / works as a constructor obviously
18:38	dabd	yes
18:40	ibdknox	I don't think I expected this:
18:40	ibdknox	,(/ 2)
18:40	clojurebot	1/2
18:40	brehaut	ibdknox: makes sense as a dual to ##(* 2)
18:41	Blafasel	,(doc /)
18:41	clojurebot	"([x] [x y] [x y & more]); If no denominators are supplied, returns 1/numerator, else returns numerator divided by all of the denominators."
18:41	ibdknox	and doesn't when compared to +
18:41	ibdknox	,(+ 2)
18:41	clojurebot	2
18:41	Blafasel	See the docs :)
18:41	ibdknox	I know
18:41	ibdknox	I was reading the source :p
18:41	ibdknox	I meant more from a conceptual standpoint
18:41	S11001001	wdouglas: if you find loop distasteful, you may consider iterate
18:41	ibdknox	I didn't expect it to work that way
18:42	S11001001	(weirdly enough, that exact phrase could also be used when talking about common lisp, though loop and iterate mean completely different things there :)
18:42	brehaut	ibdknox: i disagree; in all 3 cases, the unary implementation respects the identity value for the operation
18:43	ibdknox	brehaut, I just expected (/ 2) = 2
18:43	brehaut	ibdknox: consider in the context of the varargs imp: (/ 1 2) is 'really' (/ 1 1 2)
18:44	brehaut	(ibdknox i do get that its not necessarily obvious, but it seems correct to me)
18:45	ibdknox	brehaut, I don't follow your varargs argument
18:45	brehaut	ibdknox: / folds the varargs down; implicitly starting with the identity value for the operation
18:46	ibdknox	no
18:46	ibdknox	not based on the source
18:46	brehaut	ibdknox: thats because its hand optimized to be fast
18:46	brehaut	the naive implementation would be (defn vararg-divide [& args) (reduce binary-divide 1 args))
18:46	dabd	I'd like to change all the values in a map and zipmap seems to do the job but I am concerned that 'keys' and 'vals' don't return the elements in the same order since maps have no order
18:47	ibdknox	brehaut, mm
18:47	ibdknox	dabd, why are you concerned?
18:47	dabd	should I use map instead and then into {} on the result?
18:47	dabd	ibdknox: well maps have no order
18:47	wdouglas	S11001001: I didn't think about iterate at all. I'll see how that looks, thanks
18:48	brehaut	ibdknox: the one that really tripped me up is (< 1)
18:48	ibdknox	dabd, maps don't guarantee order no matter which way you do it
18:48	gtrak	a treemap does
18:48	gtrak	in javaland
18:48	ibdknox	gtrak: the clojure {} map doesn't
18:49	ibdknox	brehaut, yeah... I still trip up on those sometimes
18:49	brehaut	ibdknox: i had to reprogram my brain; they aren't 'less than' etc
18:49	hiredman	keys and vals will always return things in the order you want
18:49	gtrak	there is a sorted-map
18:50	hiredman	rich has guaranteed it
18:50	brehaut	ibdknox: but rather increasing sequence
18:50	ibdknox	brehaut, yeah, I liked technomancy's idea of renaming them to always decreasing
18:50	gtrak	,(source 'sorted-map)
18:50	clojurebot	#<ClassCastException java.lang.ClassCastException: clojure.lang.PersistentList cannot be cast to clojure.lang.Symbol>
18:50	brehaut	ibdknox: link?
18:50	dabd	ok then (zipmap (keys m) (map my-fn vals)) should do the job
18:50	brehaut	or was it irc?
18:50	gtrak	,(source sorted-map)
18:50	clojurebot	Source not found
18:50	ibdknox	brehaut, irc
18:50	gtrak	,(doc sorted-map)
18:50	clojurebot	"([& keyvals]); keyval => key val Returns a new sorted map with supplied mappings."
18:50	dabd	oops I meant ok then (zipmap (keys m) (map my-fn (vals m))) should do the job
18:51	gtrak	(into (sorted-map) existingMap)
18:51	brehaut	ibdknox: incidentally the doc string for < never mentions 'less than' only 'monotonically increasing order'
18:51	dabd	do I really need to use the sorted-map isn't zipmap on a hash-map enough?
18:52	ibdknox	you shouldn't use zipmap
18:52	hiredman	dabd: as I said, vals and keys will return things in an order that lines up keys with values
18:52	hiredman	you are good to go with zipmap
18:52	dabd	hiredman: thx
19:05	dabd	I am getting Java out of memory. I am using lein swank. Does anyone know how to increase heap space?
19:10	technomancy	dabd: export JVM_OPTS=-Xmx1024m
19:10	technomancy	or :jvm-opts ["-Xmx1024m"] in project.clj if you want it to apply to everyone working on the project
19:12	dabd	technomancy: thx
19:32	gtrak``	dabd, use a profiler if you're not expecting that much data
19:34	dabd	gtrak: Not the case I need to read a 52MB file that contains a clojure map :-P. Some trick to memory map this file would be useful though
19:35	hiredman	sounds like a bad format decision
19:36	hiredman	using a single json or clojure map for large data dump is icky
19:37	hiredman	you end up needing something like a what do you call it, pullparser?
19:37	hiredman	better to use small maps
19:54	wink	is there a better way to test for nil than (is (string/blank? (whatever)))?
19:55	ibdknox	wink: hm?
19:55	lynaghk	Ah, Chris, I was just looking to see if you were in here.
19:55	lynaghk	I've got a ClojureScript question for you
19:55	ibdknox	lynaghk, shoot
19:56	lynaghk	I'm getting jsc_duplicate_input errors
19:56	wink	well, testing that whatever returns nil for my given input
19:56	lynaghk	and saw that you ran into that stuff on your cljs-watch project.
19:56	ibdknox	,(doc nil?)
19:56	clojurebot	"([x]); Returns true if x is nil, false otherwise."
19:56	ibdknox	lynaghk, ah, yes. It happens when a cljs file is both on the classpath and in the compilation path
19:57	lynaghk	What's the compilation path?
19:57	ibdknox	how are you running the cljs compiler?
19:57	lynaghk	I'm using cake
19:57	lynaghk	and just invoking a brief script
19:57	ibdknox	what's the script do?
19:57	lynaghk	(closure/build ...)
19:58	wink	ibdknox: hm, silly me. so it's a little about learning what's in core.
19:58	ibdknox	the compilation path is the first arg to build
19:58	ibdknox	probably "src/"
19:59	lynaghk	Ah, so I should put my external cljs dependencies on the classpath (so that the compiler can find them), but when I'm actually compiling my own code should keep it out of the classpath?
20:00	lynaghk	I'm not sure if that makes sense, I must be missing something.
20:00	ibdknox	lynaghk, basically yes
20:00	lynaghk	uh oh
20:00	lynaghk	= )
20:00	ibdknox	so the way I fixed this for cljs-watch is by not including src/ in that jvm's classpath
20:00	ibdknox	but still including lib
20:01	ibdknox	lynaghk: speaking of, out of curiosity, why aren't you using cljs-watch?
20:01	lynaghk	I shoved clojurescript support into my existing Rakefile-based toolchain
20:01	lynaghk	which handles HAML and SASS compilation
20:02	lynaghk	I've got a lot of stuff setup in those already, so I haven't switched to the Clojure analogs
20:04	ibdknox	gotcha
20:04	lynaghk	Just as I was getting settled down with my auto-compiled minified CoffeeScript, ClojureScript gets released
20:04	lynaghk	= )
20:04	ibdknox	I'm not sure you're going to be able to do this from Cake
20:05	ibdknox	since cake's JVM will definitely have src on the classpath
20:05	lynaghk	You can config it out. I have src/clj, src/cljs
20:05	ibdknox	ah :)
20:08	lynaghk	yeah, all of this is a bit tricky
20:08	lynaghk	I've been meaning to sit down and try and figure it out
20:08	lynaghk	I just found out about the :foreign-libs option from Luke VanderHart's blog yesterday
20:09	ibdknox	it didn't exist until about a week ago
20:09	ibdknox	and wasn't merged in until this weekend I think
20:09	ibdknox	though really we shouldn't be packaging libs that are universal...
20:09	lynaghk	what do you mean?
20:09	ibdknox	they should be coming from a CDN
20:09	lynaghk	things like jQuery?
20:10	ibdknox	yeah, otherwise we're defeating the point of caching
20:10	lynaghk	true, though you might save a lot of bytes from the dead-code elimination
20:11	lynaghk	plus, it's nice to have everything in one monolithic JS---I definitely try to send single HTML file visualizations to clients when I can
20:15	lynaghk	ugh, D3 stops working with :optimizations :simple
20:15	hiredman	clojurebot: ping?
20:15	clojurebot	PONG!
20:16	lynaghk	I need to sit down and hash out my clj replacement.
20:16	ibdknox	lynaghk, I'm still on board to help with that :)
20:17	lynaghk	thanks
20:17	lynaghk	yeah, I've been letting it stew in my mind for a bit
20:17	lynaghk	spent the weekend reading up on SVG
20:17	ibdknox	I've had plenty to do, so that's fine with me :)
20:18	lynaghk	heh
20:18	lynaghk	are you going to the Conj?
20:18	gtrak``	lynaghk, what's a clj replacement?
20:18	ibdknox	I am indeed
20:18	lynaghk	gtrak`` I am thinking about implementing a visualization libary in Clojure(Script)
20:19	ibdknox	gtrak``, he meant cljs replacement for D3
20:19	lynaghk	heavily-inspired by D3/Protovis
20:19	gtrak``	ah nice
20:19	lynaghk	ibdknox, well we'll definitely have to grab a beer in November then
20:20	ibdknox	lynaghk, sounds good :)
20:20	gtrak``	how about processing.js?
20:20	lynaghk	I haven't used it very much
20:20	ibdknox	processing isn't data driven
20:20	lynaghk	or processing
20:21	ibdknox	as an example, see th visualization section of the pinot readme: https://github.com/ibdknox/pinot
20:21	lynaghk	ibdknox, what is keeping you so busy?
20:22	ibdknox	well, got Noir 1.2.0 out this weekend, started working on a SQL layer called Korma, and I have a couple secret projects in the works :)
20:22	lynaghk	WELL WELL
20:22	lynaghk	yeah, I saw that on HN
20:22	lynaghk	props
20:23	gtrak``	do you run a consulting company or something?
20:23	ibdknox	gtrak``, are you asking me?
20:23	gtrak``	yea
20:23	ibdknox	nope :)
20:23	gtrak``	i saw you had a 'prototype your site in 2 weeks thing going' :-)
20:23	lynaghk	is anything about Korma online?
20:24	ibdknox	lynaghk, not yet, I'll put something up in the next couple of days
20:24	lynaghk	I've been toying with ClojureQL a bit; I like the idea of actually using relational algebra with my databases.
20:24	ibdknox	gtrak``, I'm currently under the employ of ReadyForZero :)
20:24	gtrak``	ah
20:24	ibdknox	lynaghk, ClojureQL has some weirdness to it
20:25	lynaghk	yeah?
20:28	ibdknox	though my main motivation is actually I want it to handle relations for me
20:28	ibdknox	and I think there's a much prettier language waiting to come out of it
20:29	lynaghk	I hope you mean that literally
20:29	lynaghk	and borrow heavily from the APL characterset
20:29	ibdknox	lol
20:29	ibdknox	:p
20:31	lynaghk	Well that sounds interesting, I'll definitely look out for it
20:31	lynaghk	I'm not yet super into ClojureQL
20:31	ibdknox	the first implementation will just look like prettier sql though :)
20:31	lynaghk	uh oh, paved with good intentions right there
20:31	lynaghk	be careful
20:31	brehaut	i think clojureql is a good proof of concept
20:31	ibdknox	some of the SQL it generates hurts me
20:31	chouser	ibdknox: as long as parts of a query are values that can be composed at runtime, I'll be interested. :-)
20:32	ibdknox	chouser, they definitely are. tenant number 1 right there :)
20:33	lynaghk	Do you have any interest in implementing a 'lil DB in Clojure as well, instead of just compiling to SQL?
20:33	lynaghk	because then I can finally get my functional-relational-programming on.
20:33	lynaghk	(and compile it to JS)
20:33	gtrak``	lynaghk, isn't the point of a DB to be mutable?
20:34	lynaghk	sure, in very constrained ways.
20:36	ibdknox	err, I meant tenet above. Don't mind the stupidity :)
20:37	ibdknox	I'm not sure I'd want to implement a db
20:37	ibdknox	sounds like work :p
20:37	brehaut	ibdknox: but you could get on the leading edge of the coSQL wave
20:38	jli	co?
20:38	clojurebot	compilation is see compile
20:38	hiredman	broSQL
20:38	lynaghk	ibdknox, maybe, but you could wave yoru hands and say it's all in-memory and just implement everything with table scans
20:38	ibdknox	lol
20:38	lynaghk	maybe I'll take a crack at it sometime
20:38	ibdknox	I think fewer problems need databases
20:38	brehaut	yeah, erik meijer
20:39	lynaghk	in all seriousness though, I really would like to find some implementations of FRP
20:39	ibdknox	datastores are often more correct I think
20:39	brehaut	's strange loop keynote is about how SQL and noSQL are duals
20:39	brehaut	please hold; finding the link
20:39	brehaut	http://queue.acm.org/detail.cfm?id=1961297
20:40	lynaghk	brehaut, does that mean we can easily get a lot of funding if we make a "primal sql".
20:40	brehaut	hiredman: broSQL; the first choice in storage for low grade beer and grain spirits
20:41	brehaut	lynaghk: thats gone over my head
20:41	lynaghk	eh, math joke.
20:42	brehaut	I'm too stupid to understand math jokes
20:42	lynaghk	sayith wikipedia: "Every linear programming problem, referred to as a primal problem, can be converted into a dual problem, which provides an upper bound to the optimal value of the primal problem."
20:42	lynaghk	Oh, me too. I just try to get a few from each discipline so I can defend myself at cocktail parties.
20:43	lynaghk	That paper looks interesting though, thanks for the link brehaut
20:43	brehaut	lynaghk: see also https://github.com/strangeloop/2011-slides/blob/master/TOC.md
20:45	lynaghk	Hmm, these conference organizers are on it
20:56	vkurup	xb
20:57	vkurup	oops. That was supposed to be C-xb :)
21:06	shep-home	I evidently have 7 infinite recursion threads going from clojure-mode C-x C-e... any tips on how to stop them?
21:06	seancorfield	has anyone used Enlive with Clojure 1.3.0 successfully?
21:07	seancorfield	i just tried the basic getting started example and got: NullPointerException net.cgrand.enlive-html/load-html-resource/fn--326 (enlive_html.clj:53)
21:08	seancorfield	which is: (with-open [^java.io.Closeable stream stream]
21:09	shep-home	Ah, I got it - M-x slime-thread-kill
21:15	seancorfield	bah... that means i can't use deftemplate since that relies on the same broken loading process
21:24	jli	yay, kernel.org is back
21:35	budu	hi
21:36	budu	is it normal that swank-clojure and 'lein repl' are still using clojure 1.3?
21:36	budu	or did i installed them unproperly?
21:37	brehaut	budu: I'm not a swank user, but aren't they dependent on the project.clj clojure dependancy?
21:37	jli	"still"?
21:37	budu	yes, but i'm talking about using them outside of a project
21:37	budu	sorry for the lack of preciness
21:39	budu	nobody is doing that it seems! ;-)
21:39	gtrak``	you can install it globally, and remove it from project.clj's. it's recommended now
21:39	brehaut	i didn't even realize lein repl worked outside a project
21:49	duck1123	budu: which version are you using, I think the new one specifies 1.3
21:50	duck1123	although it might be multi dep
21:51	duck1123	hmm... it is multi dep, but 1.2 seems to be the default
22:11	mcoffbyone	hi, what is the proper way to write a ?function?macro that takes a method name as an argument? I got this (defn metfun [method] (eval `(~method "string")))
22:12	mcoffbyone	it is called (metfun 'toUpperCase) -> "STRING"
22:24	duck1123	(defn metfun [f] (f "string"))
22:24	duck1123	it's that easy
22:25	duck1123	although, I'm not sure how well that works with interop
22:26	duck1123	(defmacro metfun [f] `(~f "string"))
22:41	ibdknox	what is the correct way to specify compatibility with clojure version >= 1.2.1 for a lib?
22:42	brehaut	ibdknox: I'm using [1.2.1],[1.3.0] in necessary evil, but i think something like [1.2.1) is available too
22:42	ibdknox	ah, that works. Thanks :)
22:43	mcoffbyone	thank you duck1123, do you know why/if macros should be prefered over eval for such simple cases?
22:45	glob157	anyone use hadoop in here with clj ?
22:46	brehaut	mcoffbyone: the trite answer is that macros are idiomatic and eval is not
22:47	brehaut	mcoffbyone: macros are part of the standard compilation process; eval fires up the compiler for each call. you could cache the result, but thats more work compared to letting a macro reduce down to the underlying behavior once
22:49	ibdknox	glob157: sadly I do not, but just as a heads up, you're more likely to get an answer if you ask the question you really want the answer to
22:50	mcoffbyone	brehaut: does this mean that the output of a macro is in no way different from code forms?
22:50	brehaut	mcoffbyone: exactly
22:51	mcoffbyone	and is (name some-symbol) like eval?
22:51	brehaut	no
22:51	brehaut	name is just a function
22:51	brehaut	that works on strings and clojure.lang.Named instances, of which symbols and keywords are instances
22:52	mcoffbyone	I got it it happens after the compilation so it is not some run-time introsepction
22:52	ibdknox	as a general rule of thumb eval is evil outside of developer tooling, and even then there are often other solutions
22:52	brehaut	mcoffbyone: huh?
22:53	mcoffbyone	I thougt (name) is inspecting the name of a symbol at runtime
22:54	mcoffbyone	but the name is already an instance so it can have just a property with its name
22:54	mcoffbyone	argh the symbol not the nam is an instance
22:54	brehaut	mcoffbyone: yes; its just a function
22:54	brehaut	,(name (symbol "foo"))
22:54	clojurebot	"foo"
22:55	ibdknox	if it were a macro this would work:
22:55	ibdknox	,(name foo)
22:55	clojurebot	#<CompilerException java.lang.RuntimeException: Unable to resolve symbol: foo in this context, compiling:(NO_SOURCE_PATH:0)>
22:57	brehaut	mcoffbyone: just to underscore how uncommon eval is in clojure code, clojure/core.clj is 6500 lines long and a quick grep has "(eval" occuring 4 times (once in the definition of eval)
22:57	brehaut	and thats a piece of code that you might expect to find a lot of evals
22:58	mcoffbyone	:)
23:00	brehaut	ibdknox: http://docs.codehaus.org/display/MAVEN/Dependency+Mediation+and+Conflict+Resolution#DependencyMediationandConflictResolution-DependencyVersionRanges
23:00	ibdknox	brehaut: ah, thanks
23:01	brehaut	ibdknox: rather than cargo culting the crap i scrabbled together
23:01	brehaut	sorry it took so long to find; i had a google fail
23:01	ibdknox	technically I think the [1.2.1],[1.3.0] is the correct one
23:01	ibdknox	since I can't guarantee there won't be breaking changes in the future of clojure
23:01	ibdknox	lol
23:02	brehaut	thats what i guessed too :)
23:02	amalloy	brehaut: core.clj contains two actual uses of eval, fwiw
23:02	ibdknox	amalloy: I was trying to think up an example outside of creating a new execution context where eval should ever be used, can you think of any?
23:03	brehaut	ibdknox: whenever you want the old hands on #clojure to yell at you :P
23:03	ibdknox	lol
23:03	amalloy	maybe if you've decided life is no longer worth living
23:04	brehaut	amalloy: if thats the case, why just eval when theres PHP
23:04	ibdknox	lol
23:04	amalloy	ibdknox: eval: death by stoning
23:05	amalloy	php: self-mutilation
23:05	amalloy	er, brehaut: ^
23:05	brehaut	i'll defer to your experience in this matter
23:05	brehaut	its been years since i cut myself
23:05	ibdknox	haha
23:06	ibdknox	listen. PHP used to be all there was... unless you wanted to use perl.
23:06	brehaut	ibdknox: there was asp with vbscript
23:07	ibdknox	hah
23:07	brehaut	(a language that appeared in the mid 90s on a 32 bit platform and by default has 16 bit signed ints)
23:08	ibdknox	makes sense to me :p
23:08	brehaut	if you think that explicit return is a dumb idea, it has explicit return by assigning to the function name (but not subs; subs cant return anything)
23:08	amalloy	brehaut: how else would you implement a number with 16 sign bits? perfect for describing n-dimensional hypercubes
23:09	brehaut	haha
23:10	brehaut	amalloy: http://xkcd.com/37/
23:14	khaliG	is it educational to read through clojure source code, amalloy?
23:14	ibdknox	khaliG: very
23:14	brehaut	despite not being amalloy, its education for me
23:14	amalloy	whoa, since when am i the authority on that?
23:15	brehaut	khaliG: do you know about the source function?
23:15	duck1123	you're probably going to find a couple of weird things, but generally it's a good place to look
23:15	khaliG	since it seems you've read it :)
23:16	amalloy	khaliG: well, i'm armed with an OS that has grep built in :P
23:16	amalloy	i think it depends on whether you mean clojure source in general, or the source of clojure. both have a wealth of information, but especially in the latter case you need to be careful; there's a lot of stuff in core.clj that is not "good" code; it's there because the language is being built as you go
23:17	brehaut	khaliG: the clj source is a bit more readable and enlightening than the java source, which is pretty weird looking and does some amazingly clver tricks to get around certain limitations in java
23:18	lynaghk	Ibdknox, do you know offhand what the Closure compiler does in simple mode?
23:19	lynaghk	in particular, does it just concatenate and run everything in the compilation path?
23:19	ibdknox	munges local names, concats
23:19	lynaghk	Is there a designated entry point / namespace that gets run?
23:19	ibdknox	no
23:19	ibdknox	code at the top level will be run
23:19	lynaghk	of all namespaces
23:19	ibdknox	yup
23:20	lynaghk	ah
23:20	lynaghk	that's what I was afraid of
23:20	lynaghk	=)
23:20	gtrak``	why?
23:20	clojurebot	why not?
23:20	duck1123	then you can create a -main somewhere and point lein to it for lein run
23:20	ibdknox	advanced mode will do the same
23:20	lynaghk	okay, that's good to know
23:21	lynaghk	I've been just hacking around the build system for the past few weeks
23:21	lynaghk	I need to get things all ship-shape
23:21	lynaghk	I might put together a this-is-how-to-do-it repo
23:21	lynaghk	it = ClojureScript + Clojure macros + external JS libs
23:21	duck1123	well, lein new <projectname> should get you most of the way there
23:22	duck1123	nvm
23:22	lynaghk	oh, this is for ClojureScript
23:22	gtrak``	too bad no lein for cljs
23:24	ibdknox	well
23:24	ibdknox	when lein 2.0 comes out
23:24	ibdknox	I'll do some fun things for CLJS in it
23:24	lynaghk	is 2.0 going to have persistent JVM support?
23:24	lynaghk	because that is the main reason I am using Cake
23:25	lynaghk	Lein seems to have a lot more community backing though
23:27	ibdknox	not that I know of
23:27	ibdknox	the main thing is that it will use 1.3
23:27	lynaghk	ah
23:28	ibdknox	which is necessary for me to do cljs as a plugin
23:28	lynaghk	you are going to integrate your watcher?
23:28	ibdknox	that would make sense
23:28	ibdknox	I think
23:29	lynaghk	yeah, I think people would find it pretty helpful.
23:30	glob157	any clojure dojos on the east coast any time soon ?
23:30	lynaghk	that's the main reason cake is nice though, because cljs changes are instantanous
23:30	glob157	i need to get my skills up
23:30	lynaghk	that + the livereload browser plugin, though my understanding is that ClojureScript has some kind of Browser REPL now
23:30	lynaghk	which does all this stuff automagically for you