#clojure – 2009-03-04

2009-03-04

00:26	slashus2	durka42: If the first is a symbol should it give an error?
00:26	durka42	i mean, would you ever use a symbol as doc
00:26	durka42	that seems strange
00:27	durka42	is doc required to be a literal string?
00:27	slashus2	It appears so.
00:27	slashus2	(def testing2 "test doc") and used testing2 and it didn't work.
00:31	hiredman	uh
00:31	hiredman	that is because you just def'ed resting2 to a string
00:31	hiredman	testing2
00:31	hiredman	if you want to add a docstring with def you have to use #^
00:33	ayrnieu	you can also add documentation after the fact, with .setMeta
00:34	arohner	so struct-map takes the name of a defstruct, and key value pairs. i.e. (defstruct foo :a 1, :b 2)
00:34	slashus2	I was talking about doing something like (defn testing testing2 [x y] y)
00:34	arohner	if I have a map of key value pairs, how do I call struct-map?
00:34	hiredman	,(doc defstruct)
00:34	clojurebot	"([name & keys]); Same as (def name (create-struct keys...))"
00:34	hiredman	,(doc create-struct)
00:34	clojurebot	"([& keys]); Returns a structure basis object."
00:34	hiredman	doesn't say anything about values
00:35	arohner	,(doc struct-map)
00:35	clojurebot	"([s & inits]); Returns a new structmap instance with the keys of the structure-basis. keyvals may contain all, some or none of the basis keys - where values are not supplied they will default to nil. keyvals can also contain keys not in the basis."
00:35	arohner	struct-map wants a list of key value pairs, rather than a map
00:36	lisppaste8	slashus2 pasted "sigs" at http://paste.lisp.org/display/76450
00:36	slashus2	durka42: How does this look?
00:37	durka42	seems to make sense...
00:37	durka42	but i'm not an authority on these things
00:40	slashus2	hiredman: Do you think that this is acceptable?
00:40	ayrnieu	arohner, (apply struct-map some-struct (interleave (keys some-hash) (vals some-hash)))
00:41	arohner	ayrnieu: thanks
00:50	arohner	is there a built in function to replace the name of a key in a map, keeping the value intact?
00:51	arohner	i.e. (dissoc (assoc my-map new-name (my-map old-name)) old-name))
00:51	durka42	i think you just wrote said function :)
00:51	cp2	heh
00:51	arohner	:-) yeah, I was just wondering if I needed to write it
00:52	Raynes	Someone said "Oh Clojure is Scheme for the JVM." I replied "Bite your tongue." Started a whole uproar as they thought I was dissing Scheme.
00:54	arohner	Raynes: where was that conversation?
00:55	Raynes	#Haskell-Blah Nice guys <3 the uproar only lasted until I said "I would never dis Scheme".
00:55	cp2	lol
00:56	ayrnieu	"I would never diss Scheme, but I do think it so shameful for Clojure to be called 'Scheme for the JVM' that I demanded you harm yourself after saying that."
00:56	ayrnieu	but it's a silly thing to say, anyway. There are already several schemes for the JVM.
00:58	Raynes	ayrnieu: You took that completely out of context.
00:58	Raynes	I never meant it as being shameful. Clojure is a world-a-way from being "Scheme for the JVM" that's all I meant.
01:00	ayrnieu	world-a-way meaning 'world-a-better': if it had been asserted that Clojure was "Forth for the JVM", you'd react differently. This assertion is simply wrong.
01:01	Raynes	ayrnieu: What are you even talking about?
01:01	Raynes	Had he have said "Fourth for the JVM" I would have asked him was he out of his mind.
01:01	Raynes	Of course I would of reacted differently.
01:01	Raynes	It doesn't take a genius to realize I'm not putting Scheme down.
01:02	Raynes	There is around 10 different ways one could take what I said.
01:02	Raynes	Not just "Scheme is shameful"
01:02	Raynes	Anyways, I need sleep.
01:02	Raynes	Good night.
01:03	ayrnieu	I suppose that it's conversations like this that led people to come up with 'EQ'.
01:04	cmvkk	EQ?
01:04	ayrnieu	Emotional Quotient. Like IQ.
01:04	cmvkk	ohh.
01:22	slashus2	durka42: I will have to change that a little bit. cond isn't defined by that point, nor is symbol.
01:24	redalastor	If I have a vector that that contain cartesian coordinates (something like [[3 17] [24 62] [21 99]]), how can I find the right-most or left-most value? If two values are right-most or left-most, any is fine.
01:25	redalastor	In Python I'd use min and max but in clojure this would not work.
01:26	cmvkk	(reduce (fn [x y] (min x (first y)) myvec) ?
01:26	redalastor	I could always sort it but that seems wasteful.
01:27	cmvkk	wait that wouldn't work
01:27	redalastor	cmvkk: Thanks!
01:27	redalastor	Why not?
01:27	cmvkk	well i think that might be slightly broken, hold on.
01:27	ayrnieu	sort it by the dimension of interest or else fold over it.
01:27	cmvkk	(reduce (fn [x y] (min (if (coll? x) (first x) x) (first y)) myvec)
01:28	cmvkk	otherwise when reduce calls it with the first two values it will choke on the first coordinate.
01:28	redalastor	thanks
01:29	cmvkk	oh and that will only return the left value, not the whole coordinate...
01:29	cmvkk	hmm...
01:30	cmvkk	(reduce (fn [x y] (if (< (first x) (first y)) x y)) myvec)
01:31	Chouser	(reduce (fn [x y] (if (pos? (compare x y)) x y)) coll)
01:32	cmvkk	ooh compare does things i didn't realize it did.
01:34	redalastor	Yes, interesting. I didn't know vectors were comparables.
01:34	cmvkk	if 'compare' can do that, i think 'min' ought to be able to do it too.
01:35	redalastor	min insists on comparing numbers
01:35	cmvkk	yeah. but clearly it would be trivial to write a min that works anywhere compare does
01:36	redalastor	Yeah, that's what Python and Ruby (I think?) do.
01:36	cmvkk	same with max, <, >, etc...
01:36	cmvkk	could be an efficiency thing I guess
01:37	redalastor	Probably. But they could be diffrently named.
01:37	redalastor	I kinda expect someone will tell us it already exists in contrib :)
01:42	cmvkk	man. defn is so much harder to write a macro over than defun was.
01:45	ayrnieu	yep. Clojure doesn't have some points of macro-hostility that Scheme has, but then it has superfluous syntax that must be checked and reconstructed properly.
01:47	cmvkk	i have a bad habit of doing things like this: (defn get-filter-stuff [thing & stuff] ...)
01:47	cmvkk	"poor variable naming habits"
01:48	ayrnieu	names are hard.
01:48	cmvkk	names are hard, let's go shopping!
01:49	cmvkk	in this case, though, [thing & stuff] is everything in a defn but the name. so "thing" could be a string, an arglist, a function body...
03:21	eevar2	(conj 10 [1 2 3])
03:22	ayrnieu	,[(conj [1 2 3] 10) (cons 10 [1 2 3])]
03:22	clojurebot	[[1 2 3 10] (10 1 2 3)]
03:41	eevar2	,(doc conj)
03:41	clojurebot	"([coll x] [coll x & xs]); conj[oin]. Returns a new collection with the xs 'added'. (conj nil item) returns (item). The 'addition' may happen at different 'places' depending on the concrete type."
03:41	eevar2	,(conj [1 2 3] 10)
03:41	clojurebot	[1 2 3 10]
03:41	eevar2	,(conj 10 [1 2 3])
03:41	clojurebot	java.lang.ClassCastException: java.lang.Integer cannot be cast to clojure.lang.IPersistentCollection
03:42	eevar2	(conj [1 2 3] 10)
04:13	BigTom	Hi
04:14	BigTom	does anyone have an example of reading a file, tweaking it, and writing it, using duck streams?
04:18	BigTom	I can do something with reading the whole file in processing it and writing it out but I was hoping to do it a line at a time
04:26	BigTom	Other question, can 'doseq' be considered a "strict" 'for'
04:26	BigTom	?
04:26	hiredman	nope
04:26	kotarak	BigTom: no. doseq returns nil and is purely for side-effects. for returns a seq
04:26	hiredman	for is not a loop
04:27	kotarak	clojurebot: for
04:27	clojurebot	for is not used often enough
04:27	kotarak	clojurebot: for
04:27	clojurebot	for is not a loop
04:28	ayrnieu	clojure doesn't need separate 'strict' procedures, anyway; you can dorun or doall a lazy one.
04:29	hiredman	eager is the term
04:30	kotarak	doall is so ugly....
04:30	ayrnieu	it turns out that people have many terms.
04:31	ayrnieu	kotarak - (def work doall) ?
04:31	ayrnieu	maybe (def WORK! doall)
04:32	kotarak	ayrnieu: I mean the concept itself. Not the name. I understand that for resource management, doall makes sense, but (dorun (map #(println ..) ..)) or (doall (map #(println ..) ..)) is ugly.
04:33	ayrnieu	oh, sure.
04:39	ayrnieu	kotarak: (defmacro dolist [[v list] & body] `(loop [~v ~list] (when ~v (let [~v (first ~v)] ~@body) (recur (next ~v)))))
04:41	kotarak	ayrnieu: what is the difference to doseq?
04:43	ayrnieu	if you mean to ask why anyone would use this instead of doseq: shrug.
04:44	kotarak	ayrnieu: I'm a bit confused, what you want to tell me with this macro.
04:46	ayrnieu	that's OK. I'm going to bed.
04:51	Lau_of_DK	,(int (* 100 (/ 1966 2864)))
04:51	clojurebot	68
04:51	Lau_of_DK	According to my head, that should give 69, why dont it ?
04:52	Holcxjo	Lau_of_DK: It doesn't to rounding
04:52	Holcxjo	,(/ 1966.0 2864.0)
04:52	clojurebot	0.6864525139664804
04:52	cmvkk	yep, the real result is 68.64, which int just chops off the decimal part of.
04:52	Lau_of_DK	Thats cheap
04:53	Lau_of_DK	Do we have a func for ceiling it ?
04:53	cmvkk	Math/ceil?
04:53	Lau_of_DK	or roudinging
04:53	Lau_of_DK	thanks
04:53	cmvkk	you still have to call int if you want an int though; that always tripped me up.
04:53	cmvkk	for some reason, (Math/ceil 68.64) will return 69.0 rather than 69.
04:54	Lau_of_DK	A little odd, but quickly fixed :)
04:54	Lau_of_DK	Thanks for the help guys
04:55	Holcxjo	,(int (+ 1/2 (* 100 (/ 1966 2864))))
04:55	clojurebot	69
04:55	cmvkk	heh, that's one way of doing it.
04:55	Holcxjo	truncation becomes rounding if you add 1/2 (or subtract for negative numbers)
04:56	Holcxjo	Math/round is the other option
05:01	kotarak	,(.split "c.c" ".")
05:01	clojurebot	#<String[] [Ljava.lang.String;@adb24>
05:01	kotarak	,(seq (.split "c.c" "."))
05:01	clojurebot	nil
05:01	leafw	Holcxjo: that (int (+ 0.5 x)) is the best way to round in java/c/c++, AFAIK, all errors considered
05:01	kotarak	Am I missing something=
05:02	kotarak	?
05:02	hiredman	,(count (.split "c.c" "."))
05:02	clojurebot	0
05:02	hiredman	,(count (.split "c.c" "\."))
05:02	clojurebot	Unsupported escape character: \.
05:02	hiredman	,(count (.split "c.c" "\\."))
05:02	clojurebot	2
05:02	kotarak	argh.
05:02	kotarak	hiredman: thanks
05:50	lenst	I don't like the new repl_utils. It seems to start some GUI junk.
05:51	cgrand	lenst: to start?
05:52	cgrand	can you be more precise?
05:54	lenst	I'm running Mac OS 10.4 and when I load repl_utils I get a clojure.main process in my dock.
05:54	cgrand	and no actual window?
05:55	lenst	No window. It probably loads some swing classes, and that makes the java process into a GUI process.
05:57	lenst	A
05:57	lenst	oops. And when it starts it becomes forground process and steals focus.
06:00	cgrand	lenst: can you try to evaluate javax.swing.JEditorPane in a fresh REPL to see if the simple loading of a gui class causes the process to appear in the dock?
06:01	lenst	It does.
06:03	lenst	clojure.contrib.javadoc.browse/open-url-in-swing might be the problem
06:05	cgrand	lenst: ok, can you open an issue, I'll look at it.
06:11	BigTom	Thanks for the correction on doseq and for
06:16	AWizzArd	clojurebot: max people
06:16	clojurebot	max people is 149
06:52	timothypratley	macro question:
06:53	timothypratley	I want to do something like this:
06:53	timothypratley	(defmacro fun [ & body]
06:53	timothypratley	`(let [name (javax.swing.JLabel. "hi")]
06:53	timothypratley	~@body)
06:53	timothypratley	(fun (.getText name))
06:53	timothypratley	<- contrived example
06:54	timothypratley	is that possible?
06:55	Lau_of_DK	sure is
06:56	kotarak	(defmacro fun [& body] `(let [~'name (javax.swing.JLabel. "hi")] ~@body))
06:56	kotarak	timothypratley: but you should clearly document the capture
06:56	timothypratley	:)
06:57	timothypratley	great!
06:58	kotarak	I hope I don't go to Hell for giving such tips. ;)
06:58	timothypratley	how does it work? ~' is clearly unquote quote ---- does that avoid the namespacing --- or something else?
06:59	kotarak	~'name is the same as ~(quote name)
06:59	clojurebot	'Sea, mhuise.
06:59	kotarak	You just have to remember, that ~ means "do the following in the context where the macro is defined".
07:00	kotarak	Where everything else is done in the context the macro is called.
07:00	kotarak	(rough, but as a general guideline)
07:01	kotarak	So ~'name means: "fill in (~) the quoted (') symbol 'name'"
07:01	kotarak	One could as well say: ~(symbol "name")
07:02	timothypratley	oooo
07:03	timothypratley	I almost understand that :)
07:03	kotarak	timothypratley: writing macros is actually simple. (I know, I know, ... let me explain)
07:04	kotarak	(defmacro foo [] `(bla blub))
07:04	kotarak	This means simply: replace (foo) with (bla blub) and continue compilation.
07:05	kotarak	Now first step: ~
07:05	kotarak	`(bla ~blub)
07:05	kotarak	This basically means: expand to a call to bla with the second argument replaced by blub from the enclosing context.
07:06	kotarak	blub could eg. be a macro argument.
07:06	kotarak	(defmacro foo [blub] `(bla ~blub))
07:06	kotarak	This can also be any computation: ~(symbol blub)
07:06	kotarak	Second step: ~@
07:07	kotarak	This is a bit more complicated, because it "removes" an outer set of parens. So the unquoted thing must be a sequence.
07:08	kotarak	`(bla ~@[1 2 3]) => (bla 1 2 3)
07:08	kotarak	This is maybe a simplified view, but I got me very far, when writing a macro.
07:08	timothypratley	so far so good
07:09	papermachine	Is there a mirror for the blip.tv screencast?
07:13	timothypratley	user=> (macroexpand-1 '(foo))
07:13	timothypratley	(user/bla user/blub)
07:13	timothypratley	user=> (defmacro foo [] `(bla ~'blub))
07:13	timothypratley	user=> (macroexpand-1 '(foo))
07:13	timothypratley	(user/bla blub)
07:15	timothypratley	we removed the name space qualifier, but presumably being in the current namespace blub and user/blub are the same?
07:15	timothypratley	(evidently not)
07:24	timothypratley	Oh I think I'm starting to see the light...
07:24	timothypratley	user=> (let [user/blub 2] blub)
07:24	timothypratley	java.lang.Exception: Can't let qualified name: user/blub (NO_SOURCE_FILE:12)
07:24	timothypratley	So I understand your solution better now kotarak thanks
07:28	djpowell	What are the recent checkins related to constants all about?
07:31	rhickey	representing constants as bytecode yields big startup improvement on lowly hardware (like Android phones)
07:31	Chouser	djpowell: they appear to be mostly about altering the indentation of else-if blocks
07:32	Chouser	or what rhickey said. :-)
07:34	djpowell	cool
07:34	rhickey	I was hoping for a size reduction too, but no joy there
07:36	djpowell	does it help startup time a lot then?
07:38	kotarak	timothypratley: ~'blub basically resolves blub in the context the macro is called, while simply blub is resolved in the context the macro is defined and then replaced by the fully qualified name. So when the macro is expanded, it references the same blub, that the macro knows about.
07:40	djpowell	just tried it, it does seem faster
07:42	timothypratley	kotarak: thanks for the help :)
07:42	kotarak	timothypratley: np :)
08:08	Lau_of_DK	Anyone here who has some experience with AXIS?
08:10	banisterfiend	Lau_of_DK: my grandfather was a nazi
08:11	Lau_of_DK	Im talking about the Apache servlet system
08:11	banisterfiend	oh
08:13	p_l	...
08:13	timothypratley1	Kotarak: I have another macro question :) From your explanation I understand why this doesn't work:
08:13	timothypratley1	(defmacro my-set!
08:13	timothypratley1	[obj constraint]
08:13	timothypratley1	`(set! (. ~obj ~(symbol (name (first constraint)))) ~(second constraint)))
08:15	timothypratley1	because constraint can only have first/second taken if it is a form at macro compile time
08:16	timothypratley1	so (my-set! obj (:gridx 1)) works fine, but (my-set! obj c) where c is '(:gridx 1) doesn't
08:16	timothypratley1	but how can I get around that constraint?
08:19	timothypratley1	I suspect I must mix a function and a macro in some clever way... I can't make it just a function because set! is a special form hence why I'm trying to trick it.
08:23	Holcxjo	Don't use the tilde in front of (symbol and (second if you don't want it evaluated at macro-expansion time
08:26	timothypratley1	Holcxjo: but then (set! does not work at all because it expects a symbol, not a form).... I guess what I'm trying to do is an impossible chicken/egg scenario.
08:33	Holcxjo	timothypratley1: set! does not evaluated the symbol expression to see what it is?!?
08:33	timothypratley1	nope :(
09:00	cemerick	Building my first multimethod hierarchy. It's a joy so far.
09:00	cemerick	Didn't isa? imply instance? at one point, though?
09:01	kotarak	timothypratley1: I'm afraid you can't. Macro cannot depend on such runtime information.
09:03	kotarak	cemerick: yes, instance? should also imply isa? (instance? String foo) <=> (isa? String (class foo)) (I'm not sure about the order for isa?, though)
09:04	rhickey	,(doc isa?)
09:04	clojurebot	"([child parent] [h child parent]); Returns true if (= child parent), or child is directly or indirectly derived from parent, either via a Java type inheritance relationship or a relationship established via derive. h must be a hierarchy obtained from make-hierarchy, if not supplied defaults to the global hierarchy"
09:05	kotarak	so it probably should be (isa? (class foo) String). I always confuse the order.... :\|
09:06	cemerick	ah-ha, I was being thrown off by the same classname floating around in my ns, but from a different package.
09:10	kotarak	timothypratley: that's a common mistake in macro development by the way. You correctly recognised the limits of a macro. Don't make the mistake of trying to hard to push this limit, eg. coming up with some eval using idea. This just gives more problems. See the limits of macros and you just improved your macro-fu by an order of magnitude. :)
09:12	timothypratley	:)
09:13	cemerick	is there a builtin that will return two colls given one coll and a predicate (where one of the returned colls is the result of filtering the input coll using the pred, the other is the result of removing)?
09:15	timothypratley1	cemerick: split-with
09:15	cemerick	timothypratley1: well, almost -- that uses take-while and drop-while
09:16	kotarak	recursive split-with until the seq is consumed.
09:17	kotarak	hmmm... maybe not. Why not simply (defn unzip-with [pred coll] [(filter pred coll) (remove pred coll)])
09:18	cemerick	yeah, I already wrote that, but I figured there'd be a builtin.
09:18	cemerick	also, I'm not so hot on pred being invoked 2n with approaches like that
09:18	cemerick	picky, picky :-)
09:19	kotarak	hmmm... How would this be done lazily w/o calling pred 2n times?
09:19	kotarak	You would need some kind of Queue, no?
09:20	kotarak	When I ask one seq and it finds an item, which doesn't belong there, it stuffs it into the Queue of the other.
09:20	Holcxjo	First create a (lazy) sequence of ((filter pref elem) elem) tuples and then split them?
09:23	Holcxjo	Something like this? (defn unzip-with [pred coll] (let [checked (map #([(pred %) %]) coll)] [(map second (filter first checked)) (map second (remove first checked))]))
09:27	lisppaste8	cemerick pasted "split without 2n pred invocations (sloppy first cut)" at http://paste.lisp.org/display/76458
09:28	cemerick	that's super-sloppy, but I gotta run for a bit
09:28	kotarak	Well, it's not lazy?
09:28	cemerick	right, not lazy, but I don't care about that in this case
09:30	cemerick	my predicate in this case is pretty expensive and all of the results will be consumed anyway
09:41	pjstadig	I got past the classloader issue
09:41	pjstadig	now i (when starting a second repl) i get an arity exception
09:42	pjstadig	wrong number of args passed to: core$map
09:42	pjstadig	it happens in clojure/main.java when it tries to require the clojure.main namespace on line 38
09:43	pjstadig	i can connect JSwat and step through the code, bouncing between the java code and code in core.clj
09:43	pjstadig	mind blowing
09:49	timothypratley1	wohooo I solved by set! woes: (clojure.lang.Reflector/setInstanceField gbc (name (first c)) (second c))
09:49	timothypratley1	who needs stinking special forms?
09:53	timothypratley1	pjstadig are you able to debug exceptions with JSwat or is it too late by then?
09:56	pjstadig	i can step through until the exception happens
09:56	pjstadig	and when an uncaught exception happens i can see the call stack
09:56	pjstadig	and poke around
09:56	timothypratley1	great!
09:57	pjstadig	this is weird
09:57	pjstadig	the problem appears to be in the 'spread function
09:57	pjstadig	actually 'nil?
09:58	danlarkin	well spread is private to core.clj
09:58	Hooke	hello
09:58	pjstadig	'nil? is returning true for (nil? ArraySeq instance)
10:00	danlarkin	(doc nil?)
10:00	clojurebot	Returns true if x is nil, false otherwise.; arglists ([x])
10:00	pjstadig	hehe
10:00	pjstadig	oh good point
10:01	pjstadig	so apply is calling (. f (applyTo (spread args)))
10:01	Hooke	I'm trying technomancy's emacs clojure-mode. The clojure-install command gives me the error "(Shell command failed with error)" can anyone help, plz
10:03	pjstadig	spread returns nil
10:04	pjstadig	so it tries to apply f (which is loadlibs) to null
10:04	pjstadig	which is where i'm getting my exception
10:05	danlarkin	spread returns nil because the collection you pass is nill
10:05	danlarkin	I'm assuming
10:07	shoover	danlarkin: what's the problem with jswat? are you on JDK 1.6?
10:08	pjstadig	no the collection is not nil
10:08	pjstadig	the collection is [:require 'clojure.main]
10:09	danlarkin	shoover: java 1.5.0_16-133 on osx 10.5, when I launch it crashes with an NPE
10:10	shoover	danlarkin: I think it needs 1.6
10:11	danlarkin	pjstadig: spread "expands" to as many args as the function accepts, what's the arglist on the function you're trying to call?
10:12	danlarkin	shoover: that would explain it, then, thanks
10:12	cooldude127	does swank-clojure allow using jswat for debugging without writing a shell script to start clojure?
10:13	shoover	danlarkin: you'll need to wipe jswat and reinstall, taking pains to associate it with jdk 1.6. unfortunately I can't remember exactly how I did that
10:13	shoover	cooldude127: yes, it has a variable for extra jvm args
10:13	cooldude127	shoover: great
10:32	cooldude127	jswat is cool
10:47	pjstadig	nope it's not a problem with 'nil?
10:47	pjstadig	nil? is returning RT.F
10:48	pjstadig	for some reason it is not thinking that Boolean.FALSE == Boolean.FALSE across VMs
10:48	rhickey	pjstadig: ouch
10:48	Holcxjo	A very philosophical POV
10:48	rhickey	that will be a problem
10:49	pjstadig	it is probably something i'm doing wrong
11:14	kotarak	cemerick: how about this unzip-with? http://paste.pocoo.org/show/106457/
11:15	kotarak	general comments welcome
11:25	cgrand	kotarak: I think you don't need to use PersistentQueue: vectors can suffice
11:27	cgrand	in make-seq, replace the cons with a concat and the alter by a ref-set []
11:28	kotarak	cgrand: ah, ok. :) Hadn't thought of that.
11:28	kotarak	Just a sec.
11:30	kotarak	http://paste.pocoo.org/show/106463
11:50	rhickey	(defn unzip-with [pred coll]
11:50	rhickey	(let [pvs (map #(vector (pred %) %) coll)]
11:50	rhickey	[(remove nil? (map (fn [[p v]] (when p v)) pvs))
11:50	rhickey	(remove nil? (map (fn [[p v]] (when-not p v)) pvs))]))
11:54	kotarak	rhickey: yes. Holcxjo proposed something similar.
11:55	Holcxjo	rhickey: Can't the "(remove nil? (map ..." be done with some for ?
11:56	rhickey	kotarak: sorry, didn't see that
11:56	Holcxjo	Also: What if any element is nil ? It'll appear on neiher of the output lists
11:57	kotarak	rhickey: np :)
11:59	rhickey	Holcxjo: heh, just doing that, did you already post it?
12:01	lisppaste8	rhickey annotated #76458 "unzip-with" at http://paste.lisp.org/display/76458#1
12:01	Holcxjo	My solution was an ugly (defn unzip-with [pred coll] (let [checked (map #([(pred %) %]) coll)] [(map second (filter first checked)) (map second (remove first checked))]))
12:01	Holcxjo	Sorry for lack of formatting
12:02	Holcxjo	You version on pastebin is certainly nicer
12:02	Holcxjo	Hadn't thought of for until I saw your solution
12:04	Holcxjo	BTW: why "#(vector (pred %) %)" and not "#([(pred %) %])" ? Would you consider the latter too terse for readability?
12:04	kotarak	Holcxjo: #([]) is (fn [] ([]))
12:05	kotarak	Note the "call" to the vector.
12:05	Holcxjo	Bah -- the #() syntax is a bit too brittle for my brain
12:07	replaca	That is the one annoying thing to me about #(). It seems like about half the time I'm using it to create a quick vector. Well, that and the lack of nesting. :-)
12:07	kotarak	replaca: then simply use #(vector ....)
12:15	cemerick	kotarak, rhickey: thanks for the impls, I'll keep those under my belt for when I need laziness
12:18	replaca	kotarak: yeah, I do, it's just not as purty
12:26	te	rhickey: I just discovered your podcast on iTunes -- thanks!
12:27	te	rhickey: one other note -- you are a remarkably good speaker for a compsci guy
12:29	rhickey	te: thanks, although I'm not really a compsci guy
12:30	te	rhickey: you fooled me :)
12:30	te	rhickey: did i hear right that you're just a professional developer that got an itch one day?
12:30	mtz	:)
12:31	rhickey	te: yes, I'm just a developer
12:31	te	a good speaker AND humble
12:31	te	i'm beginning to like you more and more
12:31	te	:)
12:32	cemerick	wait, wait, wait -- a podcast, you say? :-D
12:32	te	indeed.
12:32	rhickey	cemerick: I think all the blip things end up as podcasts
12:33	te	It's just his video of clojre for lisp programmers
12:33	cemerick	ah, OK
12:33	te	or at least that's the one im watching right now
12:33	te	it would be cool to have someone comment weekly on the state of clojure, cool projects, etc.
12:33	te	community = win
12:34	te	err sorry: (= (community win))
12:35	mtz	(profit community)
12:36	pjstadig	te: i would listen to your postcast
12:36	pjstadig	*podcast
12:40	zakwilson	"I'm not really a compsci guy" -- rhickey
12:40	zakwilson	And that's why you can get stuff done in Clojure.
12:41	gnuvince	For a non-compsci guy, you sure are knowledgeable about it
12:42	Raynes	"I should probably learn LISP, but it seems so foreign..." - Paul Graham
12:42	zakwilson	I keep saying that about Haskell.
12:43	rhickey	gnuvince: I mean I'm not a researcher or anything, just a practitioner
12:43	gnuvince	ok
12:43	BigTom	is rhickey one of Zed Shaw's Coming Professional Masters ?
12:43	BigTom	http://www.zedshaw.com/essays/master_and_expert.html
12:43	zakwilson	Though I think I may finally do it just to purify my mind after working on an especially bad PHP project (http://paste.lisp.org/display/76132).
12:43	Raynes	zakwilson: Haskell is fun.
12:45	zakwilson	Raynes: I keep hearing that, but I haven't quite figured out why monads work. I almost get it, but not quite.
12:45	Raynes	zakwilson: When they say it's a mindfuck - They mean it.
12:46	zakwilson	But I need to do something to purge the evil from my mind (look at the paste, really).
12:46	Raynes	However, I never got to the monads chapter. Haskell gains and loses my interest several times over time.
12:46	Raynes	Clojure stole my interest.
12:46	Raynes	My god.
12:46	Raynes	That's some nasty PHP.
12:47	zakwilson	I get FP. Really strong static typing makes sense, though I'm not good at using it. Monads are the part I really need to get.
12:47	zakwilson	Raynes: that printf call appered in three different files, slightly modified. The rest of the code is about as bad.
12:47	danlarkin	you can write bad code in any language
12:48	Raynes	danlarkin: But it's less hard to write it in Haskell.
12:48	Raynes	In Haskell you either write good code or don't write it at all. There aren't many options.
12:48	zakwilson	danlarkin: yes you can, and I'm not blaming PHP for the above badness. You could, in fact write essentially the same thing in Haskell.
12:48	danlarkin	or clojure for that matter
12:49	Raynes	Clojure is a different story.
12:49	zakwilson	Anything with printf or something like it... which is pretty much every general-purpose language.
12:49	Raynes	It's easier to write bad code in Clojure than Haskell I feel, because you aren't held back by Haskell's strongness.
12:49	gnuvince	I agree
12:49	Raynes	Clojure will pat you on the back, Haskell will smack your face.
12:49	Lau_of_DK	Rich might smack you aswell Raynes - Good evening all =)
12:50	gnuvince	Haskell is like Gunnery Sergeant Hartman from Full Metal Jacket
12:50	Raynes	You didn't see the context, it wasn't anything bad about Clojure.
12:50	gnuvince	It's gonna yell at you until all your shit is nice and square
12:50	Lau_of_DK	I know Raynes , I was just making a funny
12:50	Raynes	:p
12:51	Raynes	Good evening Lau_of_DK
12:51	zakwilson	Yes... Haskell will refuse to compile many common errors and warn you about a host of others.
12:52	BigTom	Hi
12:52	kotarak	gnuvince: and finally it gets shot?
12:52	BigTom	I am processing a log file with for, which is nice and simple and works
12:52	Lau_of_DK	leafw or someone today, mentioned that (int (+ 1/2 x)) is the best way of rounding in c and java - Why is that?
12:52	BigTom	now I want to sometimes generate a couple of lines for each log line
12:52	BigTom	is there a neat way of doing that?
12:53	cmvkk	hmm, so the for returns a list of lines?
12:53	BigTom	yup
12:53	gnuvince	kotarak: don't push analogies too far
12:53	Lau_of_DK	gnuvince: he always does that :)
12:53	cmvkk	i guess you could wrap each output in an extra list, then do (map concat ...) on the result
12:53	leafw	Lau_of_DK: (int (+ 0.5 x)) -- AFAIK, involves the least assembly instructions and does the right thing.
12:54	Lau_of_DK	isnt that a mapcat cmvkk ? :)
12:54	Lau_of_DK	ah ok
12:54	kotarak	BigTom: (mapcat (fn [one-line] [one-line1 one-line2 one-line3]) log-lines)
12:54	BigTom	cmvkk: that's a thought
12:54	BigTom	Hi kotarak :-)
12:54	leafw	Lau_of_DK: I guess the clojure version would be (int (+ (float 0.5) x))
12:55	BigTom	Thanks, that'll get me going again
12:55	Lau_of_DK	leafw: no 1/2 works fine
12:55	leafw	Lau_of_DK: but 1/2 is not a primitive.
12:55	Lau_of_DK	,(+ 1/2 (* 100 (/ 68 20)))
12:55	clojurebot	681/2
12:55	Lau_of_DK	oops, forgot int
12:56	Lau_of_DK	But yea, true, its not a primitive
13:11	Lau_of_DK	Gents - Ive made a little webservice using compojure and I want the result to be an image which contains 1 colored bar and a pie-chart - i which direction should I be looking to accomplish this?
13:12	danlarkin	definitely an image? or will you use a javascript api for charting?
13:12	kotarak	Lau_of_DK: dejcartes?
13:12	Lau_of_DK	dejcartes is for swing apps I believe
13:12	Chouser	google charts
13:12	Lau_of_DK	danlarkin: I think I need to supply and image, which another sige can just incorporate ala <img src="myservice?arg=1"/>
13:12	kotarak	Lau_of_DK: it uses JFreeCharts, which can export images
13:13	Lau_of_DK	oh ok - I'll look into it
13:13	Chouser	you could use dejcartes to produce a .png, but google's service would allow you to let their servers do the work.
13:13	Lau_of_DK	But is it fast and stable?
13:13	gregh	you can have myservice?arg=1 redirect to google charts with the appropriate chart creation parameters
13:14	Lau_of_DK	I cant imagine GC being as flexibl as I need - The design its very clearly described
13:14	gregh	it's probably at least worth a look: http://code.google.com/apis/chart/
13:15	Lau_of_DK	Thanks greg
13:37	candera	Anyone have a clue how a ref can get completely trashed? As in, I can still do (class @foo), but @foo results in an NPE. I can isolate so that commenting out one line of code makes the problem go away, but the line of code in question doesn't try to modify the ref...
13:38	candera	I hesitate to suggest a Clojure bug (since I know how often it turns out to be the app, not the RT), but that's my best theory at present.
13:38	jayfields	clojurebot url
13:38	jayfields	hmm, how do I get the url for pasting code?
13:38	candera	There's paste.lisp.org - is that the one you're thinking of?
13:39	jayfields	no, there's some way to get the clojurebot in this room to give you a url.
13:39	jayfields	@clojurebot url
13:39	jayfields	#clojurebot url
13:40	Chouser	lisppaste8: url
13:40	lisppaste8	To use the lisppaste bot, visit http://paste.lisp.org/new/clojure and enter your paste.
13:40	Chouser	candera: what does (class @foo) return? Can you do (def tmp @foo)?
13:40	jayfields	thanks Chouser.
13:41	Chouser	jayfields: np
13:41	candera	(class @foo) => clojure.lang.LazySeq
13:42	candera	I can (def tmp @foo), but evaluating tmp throws the same NPE.
13:42	replaca	candera: yeah, cause it's lazy til you try to get the head
13:43	replaca	candera: so when you try to evaluate the head (with @foo) the evaluation throws an NPE
13:43	replaca	candera: that's my guess anyway
13:43	candera	Makes sense.
13:43	candera	Still leaves me guessing as to what the problem is, however. :)
13:47	Chouser	you should get a stack trace from the NPE that shows the root cause
13:49	candera	Chouser: So, this is a game. The error that's thrown is from the next deref of the ref. The update occurs previously. If I make a change in the update code, it no longer barfs.
13:53	AWizzArd	what is an easy way to convert a java HashTable or HashMap into a Clojure map?
13:53	Chouser	right, the update code is apparently creating a new lazy seq
13:53	AWizzArd	is into the best way?
13:53	Chouser	candera: it's in the evaluation of one of the items in the lazy seq that the NPE is thrown.
13:54	candera	Chouser: Ah, duh. Man I thought I had finally internalized this laziness stuff...
13:54	candera	Chouser: Thanks! I'll proceed down that road.
13:55	Chouser	candera: here's a simple example, in case it helps: (def foo (lazy-cat [1 2 3] (.x nil)))
13:55	Chouser	(take 3 foo) will work fine
13:55	Chouser	but just foo by itself blows up, because the repl will evaluate the whole thing in order to print it.
13:56	Chouser	AWizzArd: unless you're going to "update" the map, you should be able to just pass around the Java collection.
13:56	candera	Chouser: yep. That makes sense. Thanks again!
13:56	AWizzArd	Chouser: well, in reality I have a HashMap with strings as keys. In Clojure I would prefer to have them translated into keywords
13:56	Chouser	AWizzArd: otherwise, (into {} the-map)
13:58	Chouser	AWizzArd: (zipmap (map keyword (keys the-map)) (vals the-map))
13:59	Chouser	(reduce (fn [m [k v]] (assoc m (keyword k) v)) {} the-map)
13:59	AWizzArd	btw, does keys and vals guarantee the same order?
14:00	Chouser	if the underlying map does. I think but do not know that that's the case.
14:00	AWizzArd	provided no other thread writes to the map while we want to zipmap them, is it guaranteed that the key/value pairs will still match?
14:01	Chouser	AWizzArd: you should verify that or use the reduce version
14:06	lisppaste8	jayfields pasted "user defined globals" at http://paste.lisp.org/display/76474
14:06	lisppaste8	jayfields annotated #76474 "using printer.clj" at http://paste.lisp.org/display/76474#1
14:08	lisppaste8	jayfields annotated #76474 "failing version" at http://paste.lisp.org/display/76474#2
14:09	jayfields	okay, so I want to use (defonce flag ...) to allow users to set a flag when consuming my library, but when I try to set the flag and consume the printer class I get a failure (code available at the last link from lisppaste8)
14:10	jayfields	I can get around this issue by defining the flag in the same namespace, but that seemed like poor form, or is that the correct way?
14:13	tashafa	jayfields: dont know what you are trying to do, but have you tried "require" instead of "use" in the ns form
14:14	jayfields	nope, I check the docs. I don't know the difference.
14:17	lisppaste8	tashafa annotated #76474 "using require" at http://paste.lisp.org/display/76474#3
14:19	jayfields	that code executes without throwing an exception, but 'hello' is printed
14:19	tashafa	i think 'use' infers all the symbols in the corresponding namesapce into the current namespace
14:19	jayfields	because (I assume) should-print from helpers.printer is being used by print-hello
14:20	jayfields	what I want to be able to do is define a flag that will cause the print not to happen.
14:22	tashafa	i have no idea what should-print does... maybe someone more knowlegabel can help
14:23	jayfields	should-print is just the name of my flag
14:23	jayfields	you could just as easily use my-flag
14:23	jayfields	there's nothing special about it.
14:23	tashafa	and if that flag is set println shouldnt print?
14:23	tashafa	set to false that is
14:23	kefka	If I call a function at a top-level (using emacs shell) that blocks, is there any way to stop it without killing the whole toploop?
14:24	jayfields	did you see helpers.printer?
14:24	ayrnieu	kefka - no.
14:24	tashafa	yeah
14:25	jayfields	oh, sorry, I missed some code, one second.
14:25	lisppaste8	jayfields annotated #76474 "print-hello that notices should-print" at http://paste.lisp.org/display/76474#4
14:26	jayfields	sorry, missed the conditional logic the first time.
14:27	jayfields	so, given the last annotation, what I'm trying to do is let someone defonce should-print to false if they don't want the print to happen.
14:27	tashafa	i think you may want to use a ref
14:29	tashafa	or be more functional
14:31	Raynes	Needs moar functional.
14:31	tashafa	Raynes: ha, you know what i mean
14:32	hiredman	why use a ref?
14:32	jayfields	okay. so the problem is, I need to define a test, and in one context, immediately run that test, and in another context, don't do anything.
14:32	cemerick	my first day seriously using a hierarchy -- I seem to be writing a lot of code like (some #(isa? (class foo) %) [::type1 ::type2 ::typeN]). Is there a better (and likely faster) way to work with determine if an object is a match against some set of hierarchy "types" that are only available at runtime?
14:32	hiredman	if you are using a ref you shouldn't use **
14:32	jayfields	I'm writing some tests that I want to execute one at a time sometimes, but also as part of a larger suite at other times
14:33	hiredman	** is used to mark things that are intended to be dynamicaly rebound
14:33	jayfields	I created a macro that defines the test (using deftest from test-is) and runs the test when I want to execute individually
14:33	ayrnieu	jayfields - when you put this a library and have someone else use it, their defonce'd should-print will appear in another namespace. If you also export a (def should-print true) , then they could (bindings [should-print false] ...) to disable that. If you want a toplevel setting, then offer a ref for them to set or export a 'config' or like function that they can pass keywords to.
14:34	Raynes	I need to buy some ink so I can print the CA.
14:34	tashafa	hiredman: true, no **, @your-flag
14:34	jayfields	but I also want to not execute the individual test and run them all as a suite.
14:36	tashafa	Reaynes: CA?
14:36	lisppaste8	jayfields pasted "run individual tests" at http://paste.lisp.org/display/76476
14:36	shoover	jayfields: have you looked at binding? by definition it establishes a binding per thread, but it works great for doing different things in different contexts
14:37	hiredman	^-
14:37	hiredman	binding and set!
14:37	lisppaste8	jayfields annotated #76476 "with execute-individual-test set to do nothing" at http://paste.lisp.org/display/76476#1
14:37	hiredman	re-defing is bad, don't do it
14:37	ayrnieu	jayfields - ^ my last answer
14:37	jayfields	I get binding, but it doesn't work in my case because I want the code to execute without doing anything extra when run in isolation.
14:38	hiredman	,(doc set!)
14:38	clojurebot	java.lang.Exception: Unable to resolve var: set! in this context
14:38	hiredman	bah
14:38	ayrnieu	jayfields - OK, now read that last part. Use a ref internally but export a nice function by which they can control your library's behavior.
14:38	hiredman	ugh
14:43	tashafa	what ayrnieu said... hiredman: are you disagreeing?
14:44	jayfields	okay, so in my test_runner namespace I'll use a ref, but I'll also create a function that changes that ref, right? And, couldn't I just use an atom?
14:45	hiredman	,(binding [in 1] (set! in 2) (println in))
14:45	clojurebot	2
14:45	hiredman	Yes
14:45	ayrnieu	jayfields - yes, for the sake of offering decent interface that you can extend later. You could.
14:46	hiredman	the mechanism used for setting things like that in core.clj is set!
14:47	hiredman	,(do (set! in 2) (println in))
14:47	clojurebot	java.lang.IllegalStateException: Can't change/establish root binding of: in with set
14:47	hiredman	I see
14:49	hiredman	well
14:49	hiredman	cough
14:49	shoover	jayfields: can you paste examples of calling into helpers.test_runner?
14:49	hiredman	if you do go the ref route, you will most likely want an atom instead
14:50	candera	Chouser: your advice was helpful in tracking down the problem. However, it's still weird: if I alter a vector via concat without first materializing it via vec, it still blows up.
14:50	hiredman	unless there is the potentional for lots of contention
14:50	shoover	jayfields: re atom vs. ref, yes, atom is fine instead of ref if you don't need to synchronize with other state
14:50	candera	That is, (alter foo concat (additional-items)) blows on me, but (alter foo (fn [v] (vec (concat v (additional-items))))) does not.
14:51	ayrnieu	,(concat [1] [2])
14:51	clojurebot	(1 2)
14:52	lisppaste8	jayfields annotated #76476 "example using test_runner" at http://paste.lisp.org/display/76476#2
14:52	jayfields	shoover: there's a simple example
14:55	shoover	ok, and from the REPL you'd want that def-test to define and run the test, but from a suite you'd want to define everything and then run them all at the end?
14:56	jayfields	shoover: I'm doing it in IntelliJ with La Clojure, but I believe the result is the same.
14:56	Victorr	hello, I can't get Slime to work with clojure, I get "java.lang.Exception: Unable to resolve symbol: lazy-seq in this context (core.clj:70)", does anyone have any hint?
14:57	shoover	how about defining the default to actually run an individual test, so that works with IntelliJ, then define a suite-runner function that binds execute-individual-test to do nothing and inside that binding runs all the tests?
14:57	Victorr	I'm using the latest slime/swank/clojure from their repositories
14:57	hiredman	Victorr: it has come up before
14:58	AWizzArd	,(map identity {:a 1, :b 2})
14:58	clojurebot	([:a 1] [:b 2])
14:58	hiredman	Victorr: can you start the repl by itsself?
14:58	AWizzArd	Is there a function that will put the key/value pairs of a java HashMap or HashTable into a vector, as in that example above?
14:58	Victorr	hiredman: I can
14:59	Victorr	hiredman: it seems to me that perhaps lazy-seq moved? It is in clojure-contrib, yet I don't see a (:use) statement for it in swank's core.clj
14:59	hiredman	nah
15:00	shoover	jayfields: actually, I see that because your macro expands to code to run the test right away, you will need to use an atom or find a way for the suite to actually load tests.heartbeat's code in a binding that nulls out execute-individual-test
15:00	hiredman	Victorr: I would check again that you have the latest everything
15:00	shoover	or manually def it in your IntelliJ interactive session to actually do something
15:00	Victorr	I just checked out from their repositories just now :-(
15:01	hiredman	I don't use emacs, but other people have come in here with that Exception before
15:01	Victorr	google is not turning up anything, unfortunately
15:01	hiredman	I have four or five hits for that exception in my irclogs
15:02	Victorr	ok, I'll see if I can irc logs for this channel..
15:02	shoover	Victorr: lazy-seq is defined in clojure.core in versions of clojure as of a few weeks ago. If it doesn't exist, swank is not running in a recent enough version of clojure
15:03	shoover	perhaps your swank-clojure settings are looking at the wrong clojure, or you haven't rebuilt clojure.jar?
15:03	Victorr	ok, I'm pretty sure I just downloaded the latest clojure, but I'll try again
15:03	Victorr	I'll build from source
15:03	hiredman	the latest release is way old
15:04	hiredman	and I am sure the latest swank/slime/etc won't work with it
15:04	Victorr	hiredman, thanks a lot, indeed building clojure from source fixed it!
15:08	cooldude127	did PersistentQueue disappear or something?
15:09	kotarak	cooldude127: just used it today clojure.lang.PersistentQueue/EMPTY
15:10	cooldude127	shouldn't have, looks like it's in the source tree. so why does simply typing clojure.lang.PersistentQueue into the REPL spit an error
15:10	cooldude127	i have some code that was using it and worked fine, and now when i try to load that file it whines
15:10	jayfields	thanks shoover.
15:10	kotarak	cooldude127: well. it does not for me
15:11	kotarak	,clojure.lang.PersistentQueue
15:11	clojurebot	clojure.lang.PersistentQueue
15:13	cooldude127	ok, i did ant in the clojure dir, killed SLIME, brought it back and we're good
15:13	cemerick	it would be very handy if when-let could accept more than one binding, returning nil if any binding in the binding-form is nil
15:14	ayrnieu	cemerick - 'whereas' is a name for a syntax like this. It makes some legalistic sense :-)
15:14	cemerick	ayrnieu: is that what CL calls it?
15:15	ayrnieu	cemerick - CL doesn't have a name for it, but the name came up in a discussion on comp.lang.lisp
15:15	cemerick	ah
15:16	cemerick	it doesn't seem like another name should be at all necessary
15:16	cemerick	at least given when-let's general charter
15:18	ayrnieu	yeah, the limitation is arbitary. I'll post a patch.
15:18	taggart	,(abs -2)
15:18	clojurebot	java.lang.Exception: Unable to resolve symbol: abs in this context
15:18	jwinter	,(Math/abs -1)
15:18	clojurebot	1
15:19	taggart	aha interop
15:19	jwinter	:)
15:20	taggart	am I correct, that that is calling java.lang.Math.abs()?
15:20	taggart	or is it a namespaced clojure func?
15:20	ayrnieu	that is calling Math.abs()
15:20	cooldude127	taggart: you should look at clojure.contrib.math
15:21	taggart	ah yes thanks
15:21	cooldude127	it provides better definitions of certain math functions like abs and sqrt that work on all numeric types
15:21	cooldude127	including clojure's ratios
15:22	jwinter	clojure.contrib.math url:http://github.com/kevinoneill/clojure-contrib/blob/a19a1d1921618fe802e57fb835eb98c2164e51ba/src/clojure/contrib/math/tests.clj
15:26	tashafa	question, this might not even apply to clojure so please forgive me
15:27	cooldude127	go for it
15:27	tashafa	im using the lazy-xml from clojure.contrib to parse xml from a url
15:27	tashafa	im getting the sax parse error "Content in prolog"
15:28	tashafa	and i know the cause
15:28	Chouser	tashafa: what's the value of clojure.contrib.lazy-xml/has-pull
15:28	Chouser	oh.
15:29	Chouser	good, then. :-)
15:29	lisppaste8	ayrnieu pasted "n-binding when-let" at http://paste.lisp.org/display/76480
15:29	tashafa	the xml from the server doesnt start with <?xml ... ?>
15:30	tashafa	anyway i can modify the stream before i send it to parse-trim
15:31	lisppaste8	ayrnieu annotated #76480 "oh, right, we have zero? at this point" at http://paste.lisp.org/display/76480#1
15:32	tashafa	Chouser: has-pull is false
15:33	tashafa	The exact SaxParseException is "Content is not allowed in prolog."
15:33	Chouser	tashafa: ok, that just means you're a sax parser (instead of pull parser) running in its own thread using a blocking queue for communication with the main thread.
15:35	tashafa	ok... so i need to chage has-pull to true?
15:35	Chouser	nono, it's fine.
15:35	Chouser	but since you were getting an error from the parser, it might be useful to know which parser.
15:36	Chouser	if you don't specify one, it'll use the pull parser if available, otherwise the builtin sax parser
15:37	tashafa	has-pull is def'ed false
15:37	cemerick	ayrnieu: looks good, but it should probably assert (pos? (count bindings)) too
15:38	tashafa	so i dont see how it would use parse-seq-pull which in turn usess the pull parser
15:39	ayrnieu	cemerick - (let [] ...) also works. The user doesn't need to be protected here, and it helps with macros over when-let [including when-let itself, here].
15:40	cemerick	ah, I didn't realize that (let [] ) worked.
15:40	tashafa	i do have XPP in my calss path
15:40	tashafa	classpath*
15:42	lisppaste8	ayrnieu annotated #76480 "grammar-fix in assert string" at http://paste.lisp.org/display/76480#2
15:43	tashafa	unless i change lazy_xml/has-pull to true in the source
15:44	tashafa	which i dont want to do
15:44	tashafa	:-/
15:47	ayrnieu	(def x 1) (.bindRoot #'x 2) x => 2
15:49	tashafa	anyways back to the original question
15:50	tashafa	can a stream be modified before being used?
15:50	tashafa	if so, how?
15:51	pjstadig	so after some tests with a simple java app...
15:52	pjstadig	it appears that I shouldn't expect Boolean.TRUE to be == across VMs
15:52	pjstadig	I could make Boolean.TRUE a root so that it is the same instance across VMs, but I get an error when I try to do that
15:53	pjstadig	so they either need to be compared using equals or .booleanValue == .booleanValue
15:53	pjstadig	or i could create another static field make it a root and make sure that field is always used instead of Boolean.TRUE and Boolean.FALSE
15:54	pjstadig	we almost have that with RT.T and RT.F, but there are some cases where Boolean.TRUE and Boolean.FALSE are used directly
15:55	tashafa	let me rephrase...
15:56	tashafa	all i need to do is parse some xml, the xml is valid but has no prolog and i have no way to modify the stream
15:57	tashafa	to add the prolog
15:58	tashafa	actually its the xml declaration
15:58	tashafa	and i think the saxparser chokes on this
15:58	Chouser	tashafa: (load "lazy_xml/with_pull") is supposed to re-def has-pull if it successfully loads xpp
15:59	tashafa	Chouser: but (load "lazy_xml/with_pull") is only called when has-pull is true
16:00	tashafa	([s] (if has-pull (parse-seq-pull s) (parse-seq s startparse-sax)))
16:00	tashafa	(defn- parse-seq-pull [& _]) (try (load "lazy_xml/with_pull") ...
16:01	dcnstrct	Clojure + HtmlUnit FTW!! it's the ultimate web-interaction testing system. It totally blows away any of the Ruby libs for this kind of thing.
16:01	tashafa	or i'm i missing somethign
16:02	tashafa	brb... meeting :(
16:03	Chouser	tashafa: no, that code is unusual and confusing. It may be buggy, but not in the ways you've mentioned yet.
16:03	Chouser	tashafa: feel free to ping me when you get back.
16:03	dcnstrct	infact I'm finding so many java libraries that are better than most of the Ruby stuff I'm used to dealing with.. it's just that they were so painful to use before.. but now...
16:07	jwinter	what are you comparing HTMLUnit to?
16:13	Hooke	hi
16:18	Raynes	http://groups.google.com/group/LSharp/browse_frm/thread/f966f9a809cc702b Read what Rob said. (Yes this is relevant.)
16:22	tashafa	Im back
16:22	Chouser	Hooke: hi
16:23	Hooke	hi Chouser :)
16:23	Chouser	tashafa: the 'try' is not part of the definition of parse-seq-pull. It could probably do with a blank line in there.
16:24	tashafa	ah
16:24	tashafa	i see
16:24	Chouser	if you really want to try using xpp, which may or may not solve your real problem, you could try commenting out the try and catch parts
16:25	Chouser	that should give you a stack trace of what's actually failing in with_pull
16:29	tashafa	ok, so there isnt a way to add the xml declaration to an xml stream
16:31	Chouser	I'm surprised you need it. let me try some things..
16:32	Chouser	ok, the xpp parser doesn't complain when given a correct xml document with no <?xml?> header.
16:35	tashafa	cool
16:36	tashafa	aight let me see whats up
16:36	tashafa	IT keeps deleting my .emacs file
16:36	tashafa	they think its some sort of virus file
16:37	hiredman	makes sense to me
16:40	shoover	Hooke: I'm happy for you and sad that the conquest merits such elation
16:40	shoover	tashafa: THAT is a new one
16:41	shoover	+1 for emacs, -1 for emacs. dead even yet again
16:41	Hooke	shoover: thanks, .. I guess :D
16:43	Chouser	tashafa: I also get no error with the standard SAX parser. I think your problem must be somethine else?
16:44	tashafa	Chouser: for real
16:44	tashafa	shoover: if it doesnt look like a windows file and its sitting in root they delete it
16:45	Chouser	(clojure.contrib.lazy-xml/parse-trim (java.io.StringReader. "<foo/>"))
16:45	shoover	tashafa: that's understandable. can you put it in your user profile directory?
16:46	tashafa	yeah thats what i need to do
16:49	tashafa	chouser: if i give you the url to the xml could you try it on your end?
16:49	Chouser	sure
16:50	tashafa	pm'ed
16:56	cemerick	if-let has the same limitation of a single binding form as when-let. That probably makes more sense, though (especially since one wouldn't know what bindings are false in the else branch).
17:02	ayrnieu	cemerick - 'none of them' is wise.
17:05	cemerick	ayrnieu: what made me think of it is a situation where I needed to test whether any of a set of bindings were true, so I don't know about the friendliness/obviousness of that.
17:10	Raynes	(doc defonce)
17:10	clojurebot	defs name to have the root value of the expr iff the named var has no root value, else expr is unevaluated; arglists ([name expr])
17:10	Raynes	Lot's of def's.
17:23	replaca	tashafa: re: .emacs on Windows, you can set your HOME environment var and then put your .emacs file there instead of in c:\
17:25	tashafa	Chouser: figured it out
17:25	Chouser	ah, great, what was it?
17:26	tashafa	the respponse from the server is compressed
17:26	tashafa	gzip
17:26	Chouser	ah, ok.
17:26	tashafa	now to figure out how to umcompress
17:27	Chouser	there's a java library for that! TM
17:27	tashafa	un*
17:28	Chouser	java.util.zip.GZIPInputStream
17:29	tashafa	thanks
17:39	cmvkk	it's not safe to use \| as a symbol right?
17:40	cmvkk	or rather, it's possible/likely that it will be used for something in the future?
17:47	tashafa	chouser: golden! thanks for all your help man!
17:47	tashafa	Chouser: golden! thanks for all your help man!
17:49	kotarak	cmvkk: the allowed characters for a symbol are detailed on the clojure site: http://clojure.org/reader I would only rely on those.
17:51	cmvkk	kotarak, yeah, that's why I asked, actually. Just wondering if there was any updated information along those lines.
17:51	cmvkk	thanks though
18:01	tashafa	ftw, add an item to the end of a list?
18:02	jakewins	Hi! I'm really wanting to try clojure out, but I am something of a newbie in Java-environments, and as such am having trouble getting my first clojure project to run.
18:02	jakewins	Could someone point me to a hello-world example or something?
18:02	tashafa	,"hello world"
18:02	clojurebot	"hello world"
18:02	jakewins	lol
18:03	jakewins	Thanks. I need it to build together with a java application though,
18:03	tashafa	jakewins: do you already know your way around the repl?
18:03	jakewins	Tashafa: I know my way around nothing, I am afraid
18:04	tashafa	what exactly are you trying to build
18:05	tashafa	or from what programming language are you coming from
18:05	tashafa	?
18:05	tashafa	web, desktop, etc..
18:05	jakewins	Php mostly, some c
18:06	jakewins	And I know java as a language, ie, I can write pretty decent applications
18:06	jakewins	So I can get a normal java app to run without any trouble
18:06	jakewins	What I want to do is go into "clojure-land" from my java class
18:07	jakewins	One sec
18:08	tashafa	ftw, add an item to the end of a list? cons
18:08	tashafa	god, im losing it today
18:08	jakewins	From my main()-method in java, I run something in the likes of
18:08	jakewins	clojure.lang.RT.var("com.example.clotest.test", "main").invoke(args);
18:08	jakewins	Which indeed launches clojure
18:09	jakewins	But it launches it into the interactive-loop
18:09	jakewins	or whatever it's referred to as
18:09	jakewins	ie. I don't think it even runs the clojure function "main"
18:10	tashafa	nah...clojure doesnt work anything like java
18:10	Hooke	jakewins, I'm new too, but I'd say that the first you have to decide is what IDE to use. if you want netbeans, there the "enclojure" plugin. if you like emacs, there's clojure-mode
18:11	jakewins	Tashafa: Which is what I'm after :) My companies clients demand stuff that runs on java-platforms, but I'm not a fan of java, which is why i'd like to try this out
18:11	jakewins	Hooke: I use eclipse
18:11	jakewins	I found a clojure plugin for it, I haven't been able to figure out if it's any good though
18:12	jakewins	*company's, sorry bout the spelling
18:12	tashafa	i tried the eclipse plugin for clojure a while back and it was a bit immature, dont know how well it fairs now
18:12	Hooke	ah, yeah, I installed that one too. didnt try it much though
18:13	tashafa	fares*
18:13	jakewins	Hooke, tashafa: I dunno, I haven't been able to really try it yet :)
18:14	Hooke	now I configured emacs according to http://riddell.us/clojure/
18:14	Hooke	if you do emacs, i think thats best
18:14	tashafa	i'm also coming from eclipse... i would suggest emacs
18:14	Hooke	especially the example, ants
18:14	Hooke	ants.clj
18:14	jakewins	Hmm.. I'll definately looked into that
18:14	Hooke	very impressive
18:14	tashafa	its a step learning curve, but worth every minute once you are comfortable
18:15	Hooke	if you try emacs, try the paredit mode
18:15	tashafa	ants.clj is pretty much on the advanced side
18:15	jakewins	Tashafa: I gave emacs a shot a while back, and I can see that once you've got the basic keyboard-shortcuts down it will rock your world
18:16	tashafa	what is the paredit mode anyawas
18:16	cp2	ants.clj is my favorite :)
18:17	Hooke	tashafa: it's for never having unballanced brackets
18:17	jakewins	Oh well, I'll dig deeper into this at work tomorrow, I've gotta get up in 7 hrs
18:17	jakewins	Thanks for your help guys
18:18	tashafa	im i just turn on the Paren Hihglight matching
18:18	tashafa	jakewins: come back on tomorrow
18:18	tashafa	we'll be here
18:18	Hooke	all commands do things to sexprs, like splice, nest, and so on, without ever deleting a bracket on one side only
18:19	tashafa	hmm i have to try it out
18:20	Hooke	http://www.emacswiki.org/cgi-bin/wiki/ParEdit
18:20	tashafa	ftw, add an item to the end of a list? cons -> false
18:22	Hooke	http://mumble.net/~campbell/emacs/paredit.html
18:24	Hooke	tashafa: sorry, I dont get what you mean
18:25	tashafa	Hooke: sorry i was asking how you would add an item to the end of a list
18:26	Hooke	ah, yeah, I understand that, but .. do you mean bringing an item from outside or writing it .. I mean, write it you always can
18:26	tashafa	from outside
18:26	Hooke	bringing it from after the list would be the command paredit-forward-slurp-sexp
18:26	tashafa	i guess
18:27	tashafa	nah not talking ide anymore
18:27	Hooke	C-), C-<right>
18:28	Hooke	ahm, sorry, what are you talking about now?
18:28	tashafa	,(cons '( 1 2 3) 4)
18:28	clojurebot	java.lang.IllegalArgumentException: Don't know how to create ISeq from: Integer
18:29	tashafa	,(conj '( 1 2 3) 4)
18:29	clojurebot	(4 1 2 3)
18:29	schoppenhauer	,blah
18:29	clojurebot	java.lang.Exception: Unable to resolve symbol: blah in this context
18:29	tashafa	,(cons 4 '( 1 2 3))
18:29	clojurebot	(4 1 2 3)
18:29	schoppenhauer	,(lambda () ())
18:29	clojurebot	java.lang.Exception: Unable to resolve symbol: lambda in this context
18:29	schoppenhauer	,(fn () ())
18:29	clojurebot	java.lang.NullPointerException
18:29	schoppenhauer	,(fn [] )
18:29	clojurebot	#<sandbox$eval__7396$fn__7398 sandbox$eval__7396$fn__7398@3813c>
18:30	schoppenhauer	,(do (fn []) (fn []))
18:30	clojurebot	#<sandbox$eval__7402$fn__7407 sandbox$eval__7402$fn__7407@12421db>
18:30	schoppenhauer	scnr
18:31	tashafa	is it even possible to add to the end of a list cause of lazy-seq?
18:32	tashafa	,(conj [1 2 3] 4)
18:32	clojurebot	[1 2 3 4]
18:32	hiredman	,(concat '(a b)
18:32	clojurebot	EOF while reading
18:32	hiredman	,(concat '(a b) '(c))
18:32	clojurebot	(a b c)
18:32	hiredman	but don't do that
18:32	tashafa	ok...
18:32	hiredman	lists are for fast adding to the head
18:32	tashafa	change to a vec and conj that
18:33	hiredman	ugh
18:33	tashafa	vec the list and then add it to the end
18:34	hiredman	I guess, but it would be to either always use a vector or always use a list
18:34	tashafa	turn the list into a vector and conjoin the item?
18:35	hiredman	you just said the same thing three times in a row
18:35	hiredman	and I said "ugh" the first time
18:35	tashafa	sorry i thought you didnt understand what i said
18:35	hiredman	,(time (vec (range 1000)))
18:35	clojurebot	[0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 13
18:35	clojurebot	"Elapsed time: 0.999 msecs"
18:36	Hooke	,(+ 2 2)
18:36	clojurebot	4
18:36	hiredman	,(time (do (vec (range 100000)) nil))
18:36	clojurebot	"Elapsed time: 66.523 msecs"
18:36	hiredman	I guess that could be acceptable
18:39	tashafa	no p i was applying a list to a partial struct ...
18:39	tashafa	so i jest changed the structure of the list
18:39	tashafa	so i just changed the structure of the defstruct*
18:39	tashafa	if that makes any sense
18:40	tashafa	i need to take a break...
18:40	tashafa	my typing is awful
18:40	tashafa	goodnight folks
18:40	Hooke	gnite taggart
18:40	Hooke	sorry
18:40	Hooke	gnite tashafa
19:14	stuhood	,(class (vec (range 10)))
19:14	clojurebot	clojure.lang.LazilyPersistentVector
19:14	stuhood	hiredman: i thought 66ms looked suspicious... heh
19:15	stuhood	,(class (doall (vec (range 10))))
19:15	clojurebot	clojure.lang.LazilyPersistentVector
19:16	stuhood	~source vec
19:17	stuhood	,(class (to-array (range 10)))
19:17	clojurebot	[Ljava.lang.Object;
19:18	stuhood	nevermind... it wasn't actually being lazy. it's just that fast
19:48	yonatan__	newbie question: any way to get a stack trace on an exception?
19:49	yonatan__	i'm getting "java.lang.IndexOutOfBoundsException (bitmap.clj:0)"
19:49	yonatan__	how do you debug this?
19:52	Chouser	yonatan__: you're at a regular repl?
19:52	yonatan__	yeah
19:52	Chouser	(.printStackTrace *e)
19:52	yonatan__	thanks!
19:54	Chouser	np
19:55	Drakeson	is this: (import '(org.postgresql Driver)) supposed to work, when I have obtained http://jdbc.postgresql.org/download/postgresql-8.3-604.jdbc4.jar ?
19:57	Chouser	yes. That .jar must be in your classpath.
19:58	Chouser	or must be in a directory named by -Djava.ext.dirs=
20:02	Drakeson	Chouser: thanks
20:02	jhawk28	related to Issue 34: is there any good way to access the value of a Var in the dynamic scope on the Java side?
20:07	jhawk28	I can look up the var using the find method, but it has the global value instead of the dynamic scope
20:09	Chouser	jhawk28: you'd have to be within dynamic scope, of course, but then you should be able to call the .get() method
20:33	jhawk28	chouser: I think my issue is the dynamic scope
20:42	jhawk28	Chouser: is there a reason why the eval reader does not have visibility of the myVar in: (binding [myVar false] #=(eval (def x 3)))
20:43	Chouser	forms are processed in several phases. string->data which is called "read", and that's when the #=() form is evaluated.
20:43	Chouser	then compilation, then eval.
20:44	jhawk28	so, even though it is in the scope from a lexical perspective, it is not actually in the dynamic scope?
20:45	ayrnieu	jhawk, #=() is also not in the form's lexical scopes.
20:46	hiredman	the answer is, don't use #=
20:47	ayrnieu	jhawk - it's as if you were copying some text from one page to another by hand, and you resolved to always replace lowercase symbols with uppercase symbols, to always replace 'foo with (quote foo) , etc. So you do this -- and this is when #=() happens -- and then turn to what you've written and say: should I expand any macros, here?
20:47	ayrnieu	the answer is: understand what you're doing.
20:49	jhawk28	hehe
20:50	gnuvince_	Anybody knows git here?
20:50	ayrnieu	gnuvince - ask a question about it.
20:54	gnuvince_	I think I found it
20:55	gnuvince_	I had committed and pushed a file with a DB password
20:55	gnuvince_	I wanted to rewrite the history
20:55	gnuvince_	I think I found it
21:37	yangsx	what is the best way to get the numeric value of a char?
21:37	ayrnieu	,(int \a)
21:37	clojurebot	97
21:39	yangsx	ayrnieu: thanks, I thought that's for numbers :)
22:02	hiredman	~translate to ru: they just chug half-and-half in the morning
22:02	clojurebot	??? ?????? chug ??????? ? ?????? ???????? ???
22:06	drewr	~translate to es: clojurebot is so crazy
22:06	clojurebot	clojurebot es tan loco
22:12	jhawk28	clojurebot: where are you?
22:12	clojurebot	http://github.com/hiredman/clojurebot/tree/master
22:56	felzix	I find that if I define function A before function B then A cannot call B. Is there a way around this?
22:57	hiredman	,(doc declare)
22:57	clojurebot	"([& names]); defs the supplied var names with no bindings, useful for making forward declarations."
22:57	felzix	ah, thank you :)
22:58	cooldude127	is there any difference between def and declare in that case except that declare accepts multiple names?
22:58	cooldude127	i.e. (def x) == (declare x) ?
22:59	felzix	cooldude127: no
22:59	felzix	it just maps over what you supply, def'ing it
23:00	felzix	but you can do lots at once :)
23:00	cooldude127	felzix: so for the single case, it is the same?
23:00	cooldude127	that's what i was asking
23:00	felzix	yes
23:00	cooldude127	k, cuz i had been using def for that purpose
23:00	cooldude127	didn't even know declare existed
23:02	felzix	me either. I'm still fuzzy on how globals work in clojure so I wasn't sure I could def things like that
23:04	cooldude127	we don't call them globals :)
23:04	cooldude127	do we?
23:04	hiredman	def once, run anywhere
23:05	cooldude127	lol
23:05	felzix	ah
23:06	Chouser	they do the same thing, but they imply different things to the human reader.
23:06	Chouser	use 'declare' for forward declarations that will be re-def'ed.
23:06	cooldude127	yeah i figured that was the ideal way
23:07	Chouser	use 'def' with no initial value for things that you will thread-local bind later.
23:27	dreish	~translate to ru: I can eat glas, it does not hurt me.
23:27	clojurebot	? ???? ?????? Glas, ??? ?? ?????? ???.
23:27	cooldude127	where did that sentence come from?
23:27	dreish	If I'd spelled it correctly, it would have come from http://www.geocities.com/nodotus/hbglass.html
23:27	dulanov	~trabsltae to ru: It makes a lot more sense when we start adding multiple domains, users and aliases to our setup.
23:28	clojurebot	excusez-moi
23:28	dreish	Which is in turn devoted to a long-extinct page from I'll say 1995.
23:28	dulanov	~translate to ru: It makes a lot more sense when we start adding multiple domains, users and aliases to our setup.
23:28	clojurebot	??? ?????? ?????? ?????? ????????, ????? ?? ???????? ??????? ????????? ???????, ????????????? ? ?????????? ??? ????? ?????????.
23:50	cooldude127	can clojurebot translate to other languages?
23:51	cooldude127	~translate to es: your mother
23:51	clojurebot	tu madre
23:51	cooldude127	that answers that
23:53	slashus2	su madre?
23:53	cooldude127	tu madre is the informal version
23:54	cooldude127	~translate to es: I'm on a boat mother fucker, don't you ever forget
23:54	clojurebot	Estoy en un barco de puta madre, �nunca se olvide
23:54	cooldude127	lol
23:56	cooldude127	~translate to es: I'm on a boat, mother fucker, don't you ever forget
23:56	clojurebot	Estoy en un barco, hijo de puta, no te olvides nunca
23:58	cooldude127	~translate to ro: I'm on a boat, mother fucker, don't you ever forget
23:58	clojurebot	Sunt pe o barca, mama fraiere, nu te mai uita