#clojure – 2009-09-05

2009-09-05

01:50	hiredman	~sandbox
01:50	clojurebot	sandbox is http://calumleslie.blogspot.com/2008/06/simple-jvm-sandboxing.html
03:45	mathias	I am trying to start swank separately to be able to connect to it from a separate emacs instance. I am able to start the swank server and even connect to it from emacs but I never get a lime buffer up. I can send things for evaluation from a normal emacs buffer but how do I get the REPL?
03:45	mathias	I start clojure like this: java -cp clojure/clojure.jar:swank-clojure clojure.lang.Repl
03:46	mathias	I then do: (require 'swank.swank)...
03:46	mathias	and lastly: (swank.swank/start-server "/tmp/slime02" :port 4005 :encoding "utf-8")
03:50	hiredman	clojurebot: how can you start swank seperately and connect to it from emacs?
03:50	clojurebot	with style and grace
03:51	hiredman	(sorry)
03:53	mathias	hiredman: ?
03:54	hiredman	not an emacser, so I don't know anything about swank
03:55	mathias	hiredman: ah
04:48	tomoj	aw he left
05:38	octe	can i refresh a package loaded with the use-method?
05:54	LauJensen	(use :reload-all ns)
05:56	LauJensen	@ octe
07:12	octe	LauJensen: thanks
07:14	LauJensen	np
10:34	slyrus_	rhickey: who was the author of the database book you mentioned near the end of the blip.tv talk? Andres Rieter? I think I'm spelling that wrong.
11:41	winterstream	I've been poking around clojure-contrib to see how test-is should be used. One thing I can't figure out is how to define a variable that can be used by all tests in a given namespace - that is, I only want to do the work of setting up the variable once.
11:41	winterstream	Does anyone know how to do this?
11:46	Chousuke	hmm
11:47	Chousuke	there's the custom test hook, and fixtures.
11:51	winterstream	Chousuke: Thanks for the tip. I think I'm just going to take a shortcut now. I want to load a fairly large XML file for my tests, but I think I'll trim it down and include it as a string in my file.
11:51	kiras	i've been reading through "Programming Clojure" and i have a question about one of the examples. i'm sure it's really simple, but i'm new to clojure/lisp and somewhat confused. is it ok to show code here or is there a pastebin site that people generally use?
11:52	LauJensen	~paste
11:52	clojurebot	lisppaste8, url
11:52	lisppaste8	To use the lisppaste bot, visit http://paste.lisp.org/new/clojure and enter your paste.
11:52	LauJensen	Go go go :)
11:52	kiras	ty :)
11:53	Chousuke	winterstream: you could also just use a function that loads it once and caches the result
11:54	winterstream	Chousuke: That would actually be the best way; I should learn how to memoize in Clojure anyway :).
11:54	Chousuke	hm
11:55	Chousuke	(doc memoize)
11:55	clojurebot	"([f]); Returns a memoized version of a referentially transparent function. The memoized version of the function keeps a cache of the mapping from arguments to results and, when calls with the same arguments are repeated often, has higher performance at the expense of higher memory use."
11:55	winterstream	So now I can be even lazier :). Thanks.
11:55	Chousuke	dunno if that works for functions with no arguments though :/
11:56	winterstream	If not, I'll just hack it by adding an unused parameter :)
11:56	kiras	http://paste.lisp.org/display/86614
11:57	kiras	my question has to do with line 3. in the book, it says "On line 3, the zero-argument body returns the concatenation of the basis values [0 1] and then calls the two-argument body to calculate the rest of the values."
11:58	LauJensen	So whats the question?
11:59	Chousuke	hm, the wording of the book is a bit weird.
11:59	kiras	so... does concat return values multiple times? or does it calculate the value once and then return it?
11:59	kiras	did i phrase that ok?
11:59	Chousuke	concat just concatenates two sequences
12:00	Chousuke	(lazy-seq-fibo 0 1) call produces a lazy sequence, so concat concatenates [0 1] with that sequence
12:00	LauJensen	,(concat [1 2] [3 4])
12:00	clojurebot	(1 2 3 4)
12:00	LauJensen	clojurebot: compute!
12:00	clojurebot	Pardon?
12:01	kiras	ok. the way the book put it, it seemed like it was saying that concat returned (0 1) first, then called (lazy-seq-fibo 0 1)
12:02	kiras	at least, that's how it seemed to me
12:02	Chousuke	well, the (lazy-seq-fibo 0 1) gets evaluated when it's passed to concat.
12:02	Chousuke	but the result of it is a lazy sequence, so no items of the sequence are actually produced
12:03	kiras	well, i think i understand the lazy sequence part
12:03	Chousuke	and since concat is lazy too, no additional items will get computed until you have consumed the first two (the [0 1]) from the resulting sequence.
12:04	kiras	i was mostly confused about how arguments to concat were evaluated
12:04	kiras	oh
12:04	kiras	lol
12:04	kiras	i think i'm confused again...
12:04	Chousuke	arguments to functions are always evaluated.
12:04	kiras	ok
12:05	Chousuke	but the thing is, when evaluated, (lazy-seq-fibo 0 1) produces a lazy sequence, and unless you actually try to access items from the sequence (which concat won't do until it's needed), then no items will be computed
12:06	kiras	ok... so basically, if it wasn't lazy, it would produce the whole set of fibonacci numbers?
12:06	Chousuke	yes.
12:06	kiras	ok
12:07	LauJensen	kiras: I dont know if you already got this, but if not this might help: http://paste.lisp.org/display/86614#1
12:08	kiras	laujensen: ty, i'll look at that
12:12	kiras	LauJensen: i understand your example. at the moment, i think i'm ok with the recursion that i've seen, just the lazy sequence part is something i don't remember coming across before.
12:12	LauJensen	Just the simplest for of dispatching on argument count
12:12	LauJensen	Alright
12:12	kiras	ty
12:13	LauJensen	But you understand now?
12:14	kiras	well... sort of... i guess to really understand, i might have to look at how lazy-seq actually works with take and similar functions
12:17	LauJensen	,(doc lazy-seq)
12:17	LauJensen	You can use that in your emacs repl when new functions come up
12:17	clojurebot	"([& body]); Takes a body of expressions that returns an ISeq or nil, and yields a Seqable object that will invoke the body only the first time seq is called, and will cache the result and return it on all subsequent seq calls."
12:18	Chousuke	hmm
12:18	Chousuke	,(lazy-seq 5)
12:18	clojurebot	Don't know how to create ISeq from: Integer
12:18	kiras	ty, i was actually reading that just now lol
12:18	Chousuke	ah, right.
12:27	kiras	so a lazy-seq is "a Seqable object that will invoke the body only the first time seq is called..."
12:27	kiras	?
12:28	LauJensen	Yes sir
12:28	kiras	so is it actually a sequence or does it just implement the ISeq interface? if that makes any sense?
12:30	Chousuke	,(ancestors (lazy-seq [1 2]))
12:30	clojurebot	nil
12:30	Chousuke	huh.
12:30	Chousuke	,(ancestors (class (lazy-seq [1 2])))
12:30	clojurebot	#{java.lang.Iterable clojure.lang.ISeq java.util.Collection clojure.lang.Seqable java.util.List clojure.lang.IMeta java.lang.Object clojure.lang.IObj clojure.lang.Obj clojure.lang.Sequential :clojure.contrib.generic/any clojure.lang.IPersistentCollection}
12:30	LauJensen	As I understood it in plain english, its a sequence of items where each expression is calculated when you eval it. Ex (take 5 (iterate inc 1)) will return (1 2 3 4 5), the next item is 6, but its not calculated, and 1-5 is cached for future calls.
12:31	Chousuke	take itself produces a lazy seq though
12:31	Chousuke	at least, IIRC :)
12:32	LauJensen	true
12:32	Chousuke	,(take 2 (take 5 (iterate (fn [x] (do (println x) (inc x))))))
12:32	clojurebot	java.lang.IllegalArgumentException: Wrong number of args passed to: core$iterate
12:32	Chousuke	oops
12:32	Chousuke	,(take 2 (take 5 (iterate (fn [x] (do (println x) (inc x))) 1)))
12:32	LauJensen	You forgot the offset
12:32	clojurebot	(1 2)
12:33	Chousuke	weird, nothing printed :P
12:33	LauJensen	its in the inferior lisp buffer :)
12:36	kiras	when you say "1-5 is cached for future calls" how long are those values cached?
12:38	Chousuke	kiras: until they're no longer referenced.
12:39	Chousuke	and they're referenced as long as you hold on to the head of a sequence that contains them.
12:39	kiras	Chousuke: so if they aren't stored in a variable or the such, they would be calculated again, right?
12:39	Chousuke	well
12:40	Chousuke	if you do (let [x (iterate inc 1)] [(take 5 x) (take 10 x)]) then on the first take call, the values get calculated, but for the second the cached ones would be used for the first 5 and then again calculated
12:41	kiras	that makes sense
12:41	LauJensen	kiras: but short answer is yes
12:42	Chousuke	the caching is why you don't want to do (def fibos (lazy-fibo))
12:42	kiras	hehe
12:42	Chousuke	because then if you did (take 10000 fibos) then it'd hold on to the 10000 fibos until the program ends (or until fibos is redef'd)
12:43	kiras	right
12:43	Chousuke	accidentally holding on to the head of lazy sequences is the main cause of out of memory errors in programs btw :)
12:43	kiras	lol
12:44	Chousuke	it's sometimes difficult to see if you're holding a reference to something or not.
12:44	kiras	i was wondering if that was a very common mistake or not
12:44	LauJensen	kiras: Chousuke does it all the time
12:44	Chousuke	especially because the clojure compiler currently does not nullify references after last use.
12:44	Chousuke	only at tail calls
12:48	kiras	i never really thought about that. do you mean that, if you were to do (def fibos (lazy-fibo)) and store a lot of numbers in there, used them once and then never used them again, it might be possible for the compiler to know the fibos was never going to be used again and set it to null?
12:48	Chousuke	nah, not that.
12:48	Chousuke	I meant local references
12:49	LauJensen	kiras: As it is now, that reference would be kept until program termination
12:50	Chousuke	eg. (fn [x] (let [foo (do-something (drop 100000 x))] "the reference to x still exists here even though it's not needed")
12:51	Chousuke	if the compiler were smarted it could figure out x is not used after the do-something call and nullify the reference, but currently it doesn't do that.
12:51	kiras	LauJensen: seems like it would. i thought maybe Chousuke meant that someday the compiler might be able to tell and set it to null. would think that would be very difficult to do though.
12:51	Chousuke	smarter*
12:52	kiras	Chousuke: i see...
12:52	kiras	interesting
12:55	Chousuke	to get around that you need to do (fn [x] (letfn [(work [y] (let [foo (do-something y)] "no ref to x here"))] (work (drop 100000 x))))
12:56	Chousuke	because the (work ...) call is at the tail position the compiler knows x can't be referenced after the call and nullifies the reference
12:56	Chousuke	of course, the example is rather silly like that, but imagine that the strings returned were something more complex (possibly utilising foo, but not x)
12:57	LauJensen	kiras: With dynamic code (macros) I think its near impossible.
12:59	kiras	so do lazy sequences keep track of references to themselves and the parts of the sequence each reference is holding onto (pointing to?)? then when a variable that is using part of the sequence is deleted, the lazy sequence drops that part of the sequence from its cache? or is that handled outside the lazy sequence somehow?
13:00	Chousuke	kiras: the lazy sequence is like a linked list. it's made of smaller subsequences. :)
13:01	Chousuke	kiras: and each subsequence just caches its own value after it's computed.
13:01	Chousuke	the head of each subsequence that is.
13:02	tomoj	so if you're holding the head, the entire sequence computed so far is not gcable?
13:02	Chousuke	yes.
13:02	Chousuke	because there's a reference from the head to the second item, from the second item to the third, etc. until the last actually computed item
13:03	Chousuke	this might be oversimplified though, because you can have seqs that are just "views" to something that already exists fully in memory, like a vector.
15:01	fps	hello, could someone give me a hint on how to use the clj-record library? http://github.com/duelinmarkers/clj-record/tree/master/
15:03	fps	I already compiled it, but I cannot find a jar to add to the classpath
15:09	Bercilak	for those of you using emacs clojure-mode, have you had any success using the clojure-update command?
15:10	Bercilak	I'm running into an error where it tries to run ant in the wrong directory, and I think it's a bug in clojure-mode.
15:11	cschreiner	I get an error with Git
15:11	cschreiner	but I've had no success using the clojure-update command
15:12	Bercilak	out of curiosity, what's the error?
15:15	cschreiner	just that Git cannot be found
15:16	cschreiner	I just left it, and continued with what I was doing before
15:16	cschreiner	there are a couple of things that don't work right in emacs, but I'll manage
15:18	cschreiner	like; code-completion, java-lookup, some problems with elisp formatting from swank etc
15:19	cschreiner	especially when doing (println "whatever") statements
15:20	Bercilak	hmm. I'm still in the process of switching to emacs from textmate, so I'm useless for those...
15:20	cschreiner	error in process filter: slime-dispatch-event: Elisp destructure-case failed: (:write-string "Whatever")
15:21	cschreiner	yeah, textmate, never seen much of that since going back to emacs
15:22	cschreiner	I tend to spend more time browsing elisp code than actually clojouring
15:23	cschreiner	I still need to understand the big picture though (re. emacs)
15:27	LauJensen	For those of you who are interested I'm launching a blog which pivots on Clojure but from a business relevant angle. You'll find it at http://bestinclass.wordpress.com
15:27	cschreiner	sure :-)
15:31	Bercilak	textmate is very approachable and snippets are simple to understand and damn useful
15:31	Bercilak	slime is so worth the switch though, and eventually I'll get the hang of emacs
15:31	clojurebot	slime is icky
15:31	Bercilak	heh
15:31	LauJensen	~forget slime
15:31	clojurebot	I forgot slime
15:31	Chousuke	in the end emacs is more powerful than any other editor
15:31	Chousuke	it just requires some effort :/
15:32	cschreiner	I messed with the ergoemacs keybindings, but I returned to a blank slate
15:32	Bercilak	I always knew that I was going to switch to emacs or vim someday, so recently I set up a linux dev box and I've been doing all my work on that
15:32	cschreiner	better to build the system bit by bit with functionality I am really going to use
15:33	cschreiner	on another hand, the original keybindings are really hard on my old hands
15:33	Bercilak	ergoemacs looked interesting, it's nice to hear from someone who tried it out
15:33	Chousuke	maybe you just need a better keyboard :/
15:33	cschreiner	maybe
15:34	cschreiner	I need an original symbolics thing
15:34	Bercilak	hehe. I type in dvorak which makes a lot of these layouts kind of infuriating. They make sense in qwerty but not dvorak
15:34	Chousuke	I get by with caps lock as meta and cmd as control but still I think a better keyboard would help a lot :P
15:34	cschreiner	Well, I have three different apple keyboards here, and all of them makes my hand ache
15:35	cschreiner	I need to think more about typing less
15:35	Bercilak	maybe it's blasphemy, but I might switch to using vimpulse/viper for basic text movement and manipulation
15:35	Bercilak	elisp is where it's at for configuration though
15:35	cschreiner	care to share?
15:36	cschreiner	vimpulse, is that a vim thing?
15:36	Chousuke	viper.el is a vi implementation for emacs.
15:36	Chousuke	and vimpulse adds some vim things missing from viper.
15:36	Chousuke	I used viper but it didn't play well with paredit so now I'm trying to get used to emacs controls
15:37	cschreiner	any torough docs/intros?
15:37	cschreiner	I think the incremental search is the most important thing, and sexp navigation of course
15:37	Chousuke	paredit is awesome.
15:38	cschreiner	so few modifiers, so many modes
15:39	Chousuke	I don't even do much besides some barfing and slurping and sexp-killing with paredit but it's still an enlightening experience.
15:39	Bercilak	I just googled it... the paredit koolaid looks delicious.
15:40	Chousuke	I suppose if I learned to use the sexp navigation commands properly I would be one step closer to lisper nirvana
15:41	cschreiner	examining the pareditcheatsheet now http://www.emacswiki.org/emacs/PareditCheatsheet
15:43	Chousuke	one of the more common operations I do with paredit is deleting a let, and it's really very easy :/
15:44	Chousuke	or the reverse, wrapping some code in a let.
15:45	cschreiner	working with a non-us keyboard layout, everthing is fucked from the beginning
15:45	cschreiner	the intention of the original layout does not count
15:46	cschreiner	or, is non-existant
15:47	Chousuke	I find it usable on a Finnish layout.
15:47	Chousuke	though I think if I got used to a US one it would be better
15:47	cschreiner	Suomi?
15:47	Chousuke	it's not like I need ä ör ö when coding :/
15:47	Chousuke	yeah. :P
15:48	cschreiner	I would never have guessed it
15:48	cschreiner	Norway here
15:52	cschreiner	barfage and slurpage seems reasonable enough, never touched paredit- before
15:53	cschreiner	guess I've loaded it, but not bound any of the functions
16:21	quidnunc	cgrand: Is it possible to select a node whose child satisfies some predicate in enlive? I'm trying to do (select foo [[:font (has [:a (attr-contains :href "blah")])]]) but it doesn't seem to be working (the selector on the child works on its own.
17:19	cgrand	quidnunc: (select foo [[:font (has [[:a (attr-contains :href "blah")]])]])
17:20	cgrand	'has takes a whole selector not a single selector-step
17:23	quidnunc	cgrand: Thanks.
17:26	quidnunc	cgrand: There is no way to get a parent of a node once I from the value returned by select right?
17:27	cgrand	not by select itself
17:30	cgrand	look at zip-select: something like (zip-select (map xml/xml-zip nodes) [your selector here]) -- where xml is net.cgrand.xml
17:30	quidnunc	cgrand: One last question. Is it possible to use zip-filter.xml functions on the nodes returned by select?
17:32	quidnunc	cgrand: I wasn't clear about the difference between zip-select and select. What does zip-select do differently than select?
17:32	cgrand	zip-filter.xml may barf on xml comments (other than whitespace handling it's the main difference between clojure.xml and net.cgrand.xml)
17:34	cgrand	'select returns nodes, 'zip-select returns locs (zips) to these nodes... and you can get the parent of a loc
17:35	cgrand	once you have a loc you can freely use clojure.zip or zip-filters
17:37	quidnunc	Thanks cgrand.
17:37	cgrand	welcome
17:41	cgrand	quidnunc: bedtime for me
18:11	cemerick	holy crap, I just grokked what subseq is all about
18:12	cemerick	I had never read the docs -- thought it was over Indexed colls!
18:35	cemerick	it'd be great to see PTM become extensible (so as to support building stuff like interval trees on top of it, etc)
18:44	emacsen	Why doesn't this do what I expect (knowing I clearly misunderstand): (reduce merge [{:a 10} {:b 20} {:c 30}])
18:45	Chousuke	,(reduce merge [{:a 10} {:b 20} {:c 30}])
18:45	clojurebot	{:c 30, :b 20, :a 10}
18:46	Chousuke	isn't that what you want it to do?
18:46	emacsen	I didn't get that... let me try again
18:46	Chousuke	though you could also do (apply merge ...)
18:46	Chousuke	which probably internally uses reduce :)
18:46	Chousuke	but might not!
18:46	emacsen	lol, okay this was a slime problem with my display
18:46	Chousuke	~def merge
18:46	emacsen	I was like "I must be going crazy... this should work"
18:47	Chousuke	hm, merge could be changed to use transients I suppose
18:48	emacsen	Chouske, okay here's a harder one... given another vector in mynode's content, this is invalid:
18:48	emacsen	(map #({(:k (:attrs %)) (:v (:attrs %))}) (:content mynode))
18:48	emacsen	I've tried that function on a single argument and it works, and (:content mynode) gives me back a persistent vector
18:49	Chousuke	#() is not fn
18:49	Chousuke	#({}) means calling {} with no parameters
18:49	Chousuke	I assume you meant to return the map
18:49	emacsen	isn't #() a macro for (fn ..)
18:49	Chousuke	no
18:49	Chousuke	#(foo) = (fn [] (foo))
18:50	emacsen	oh good point
18:50	emacsen	right... okay this all makes more sense
18:50	hiredman	> ((x: Any, y: Any) => (f: Function2[Any, Any, Any]) => f(x, y))(1, 2)((x: Any, y: Any) => x)
18:50	clojurebot	Any = 1
18:50	hiredman	:D
18:50	Chousuke	what.
18:51	Chousuke	Clojurebot does Scala now?
18:51	Chousuke	what blasphemy is this.
18:51	emacsen	oh that's scala?
18:51	hiredman	yep
18:52	hiredman	via the simplyscala online repl
20:42	AWizzArd	Anyone awake who knows well about snapshots in Clojure? When I made a snapshot of a ref, and then the ref is mutated, will it use copy on write, as in forks?
20:46	hiredman	uh
20:46	hiredman	what?
20:48	AWizzArd	when we make a snapshot of a map and have another thread changing the map, will then Clojure have to copy everything from the root to that leaf that gets mutated?
20:48	hiredman	mutating a ref is just assigning it a new value, te copy-onwrite or not depends on what the value is
20:48	AWizzArd	imagine a fork.. when the parent or child mutates data then the os will first copy the page, because everything is marked as read-only
20:49	hiredman	that has nothing to do with refs
20:49	AWizzArd	but mutating something in a tree may require to copy data from the root to that mutated leaf
20:50	hiredman	refs do not care, and don't know anything about what they wrap
20:50	hiredman	and copy on write, etc, is the responsibility of the object wrapped by the ref
20:50	AWizzArd	i snapshot a ref (i call deref on it) and have a long running function that works on the object
20:51	hiredman	sure
20:51	AWizzArd	and now another thread mutates this object..
20:51	hiredman	then you are screwed
20:51	AWizzArd	then COW is happening under the hood?
20:51	hiredman	that is not how refs work
20:51	hiredman	the value a ref holds should never mutate
20:51	AWizzArd	the ref is already forgotten after a snapshotted/derefd
20:51	hiredman	the ref can mutate and point to different values
20:52	AWizzArd	yes, let's forget the ref
20:52	hiredman	we cannot just forget the ref
20:53	hiredman	if you wrap a mutable object like stringbuilder in a ref, the ref is useless
20:53	AWizzArd	yes
20:53	AWizzArd	but say it is a clojure map
20:53	AWizzArd	we call (foo @our-map)
20:53	hiredman	ok, clojure maps are implemented using structural sharing
20:54	AWizzArd	and while foo is running some other function bar mutates our-map
20:54	hiredman	aand Immutable
20:54	hiredman	AWizzArd: you cannot
20:54	hiredman	Immutable
20:54	AWizzArd	our-map sits in a ref
20:55	hiredman	doesn't matter
20:55	AWizzArd	and as you said, the structure is shared
20:55	AWizzArd	so, it seems to us that we are working on a completely new object, while under the hood the same thing is mutated
20:55	hiredman	no!
20:55	AWizzArd	if our-map is 600 mb in size, then not each time those 600mb need to be copied
20:55	hiredman	no!
20:55	hiredman	it is a new object
20:56	hiredman	it just contains a pointer to another object that it shares some structure with
20:56	hiredman	no mutation
20:56	AWizzArd	I know, they are not identical
20:56	Chouser	only the parts that are identical are shared
20:57	AWizzArd	when we call (foo @our-map) then at those point foo sees an immutable map, and all threads can see the same thing
20:57	AWizzArd	similar to a fork
20:57	Chouser	the COW, if you want to think of it that way, happens the moment you do (assoc ...), regardless of whether a ref is involved or not.
20:57	AWizzArd	yes
20:57	hiredman	,(let [a '(1 2 3) b (cons 4 a)] [a b]) ;shared structure, no mutation
20:57	clojurebot	[(1 2 3) (4 1 2 3)]
20:57	AWizzArd	this is what i want to know
20:57	AWizzArd	does a COW happen, and how efficient will it be
20:58	AWizzArd	some other thread runs the function bar which is doing transactions and alters our-map
20:58	hiredman	AWizzArd: it happens everytime you cons, conj, assoc, etc
20:59	hiredman	AWizzArd: you cannot alter the map
20:59	hiredman	cannot
20:59	Chouser	AWizzArd: http://blog.higher-order.net/2009/02/01/understanding-clojures-persistentvector-implementation/
20:59	hiredman	you can create a new map that shares structure with the old map, and place the new map into the ref
20:59	AWizzArd	yes
20:59	hiredman	but that in no way effects our-map
21:00	AWizzArd	Chouser: thanks for the link
21:00	AWizzArd	yes hiredman I know, but I want to know how efficient it will be under the hood
21:00	AWizzArd	consing is no problem
21:00	AWizzArd	but changes in the middle might be
21:01	hiredman	,(time (reduce conj [] (range 10000)))
21:01	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
21:01	clojurebot	"Elapsed time: 10.073 msecs"
21:01	Chouser	AWizzArd: changes on the end and in the middle are handled identically (for both vectors and maps)
21:01	clojurebot	for is not used enough
21:03	hiredman	AWizzArd: I recomend you read http://clojure.org/refs and http://clojure.org/state
21:04	AWizzArd	those pages don't explain much of how structure sharing is done
21:04	hiredman	correct, I am more concerned with you understanding of refs
21:04	AWizzArd	that is no problem at all
21:05	hiredman	…
21:05	AWizzArd	only that i was not talking about them, and fortunately Chouser understood this and gave me an interesting link
21:06	mabes	any vimclojure users here? When I try to run ant in the plugin dir the build fails:
21:06	hiredman	are you following the directions in the readme?
21:06	mabes	http://pastie.org/607372
21:07	mabes	hiredman: I was following a screen cast.. I'll double check the readme to see if something has been updated
21:08	hiredman	you need to follow the build instructions for contrib
21:08	mabes	it is as if it can't find my clojure-contrib.jar but I defined it in the local.properites
21:08	mabes	hiredman: ok.. so I just don't run "ant"?
21:09	hiredman	mabes: run ant and check for a "Warning ..." and follow the instructions
21:09	mabes	hiredman: k, thanks for the help
21:10	AWizzArd	Chouser: when there is a big map and in a snapshot we have the k/v pair [:a 100] and some other thread dissocs it, then the snapshot still sees it. Will that operation require very much work under the hood, or is that still very efficient?
21:18	Chouser	AWizzArd: that has nothing to do with threads or refs. 'dissoc' will do what it does
21:19	hiredman	,(time (let [v (vec (range 1e6))] (nil? (reduce #(assoc % %2 (* %2 3)) v v))))
21:19	clojurebot	"Elapsed time: 2788.926 msecs"
21:20	hiredman	1e6 new vectors
21:22	AWizzArd	I will definitly need to look into that. It seems interesting to see how dissoc can remove elements in such a way that no snapshots will see it.
21:22	AWizzArd	Oki, thanks you two and good night
21:22	hiredman	No!
21:23	hiredman	dissoc doesn't effect any "snapshots"
21:23	hiredman	the snapsots are immutable!
21:23	AWizzArd	exactly
21:23	AWizzArd	they won't see the change
21:23	AWizzArd	but when the structure is shared then some copying must happen, so that the snapshot is not a full copy but still can see the dissoc'd element
21:24	AWizzArd	n8 n8
21:24	Chouser	yes, clojure's persistent structures are amazing.
22:04	emacsen	n8?
22:04	emacsen	oh, night
22:29	emacsen	hey technomancy
22:30	emacsen	technomancy, how mature is your http abtraction library?
22:39	technomancy	emacsen: in terms of age/usage, not very
22:39	technomancy	I wrote it in three days and haven't touched it yet
22:39	technomancy	I've got like three patches contributed that I've been sitting on.</shame>
22:40	emacsen	I don't need any http library yet but soon will
22:40	emacsen	and looking for something other than the java one, because I don't know Java, I find it distracting to go back and forth between documentation and programming styles
22:41	emacsen	unless Clojure comes with another one that I've missed
22:44	technomancy	the java one certainly requires a lot of ceremony
22:45	technomancy	emacsen: there's one in contrib too; uses agents
22:45	emacsen	that's a lot of overkill for what I'm doing
22:47	technomancy	if you're working with a restful API then I'd recommend the "resourcefully" namespace in clojure-http-client
23:05	nathanmarz	any monad masters around?
23:38	emacsen	technomancy, I am indeed working with a restul api
23:53	prospero_	how do I reset the state of the repl?
23:58	hiredman	what state?