#clojure – 2011-01-27

2011-01-27

01:19	no_mind	anyone has a working example of JSON being returned from a clojure program ? I am trying this https://github.com/sivajag/Address-Book but no luck
02:07	amalloy	LauJensen, you around?
03:00	LauJensen	amalloy: yea
03:00	LauJensen	Just got in
03:00	amalloy	nm, i decided to bug you in #clojureql instead
03:00	LauJensen	:)
04:12	mrSpec	Hello! I have problem with (eval (read)) in my program. I have: (defn -main [& args] (while (= 1 1) (print "REPL> ") (eval (read)))) but Instead of "REPL>" I have empty line. When user writes something, "REPL>" is printed and after that the result. What am I doing wrong? Can I force Clojure to print "REPL>" before evaling what user write?
04:14	opqdonut	try flushing out
04:14	zmila	did you tryed (println "") instead of (print )?
04:34	mrSpec	thanks! (flush) solve this problem :)
04:40	solar_sea	Is the joy of clojure book ready for reading or it's still a work in progress ? (I'm wondering wether I should get the "meap" subscription for it or wait for a final release)
04:44	clgv	solar_sea: I heard that it is already completed and you can buy it from Amazon or Manning
04:56	LauJensen	solar_sea: http://joyofclojure.com/buy
04:56	clgv	how do I tell leiningen about local dependencies? I have a java implementation in a jar file I want to use.
06:43	edoloughlin	For a simple cache, a map of maps keyed by id & shared by multiple threads: is it best to use an atom or ref?
06:44	companion_cube	i'd say a ref, because to keep the cache consistent you would modify the map in a transaction
06:47	ejackson	agreed - if more than one thread is writing to the cache, then you need to coordinate it
06:48	ejackson	if only one writes, and the others read, then I'd argue an atom is better as it doesn't need to be coordinated.
06:49	edoloughlin	If all I'm doing it adding (and periodically removing/cleaning) entries to a map, then can't an atom provide a sufficient level of locking? There's only one variable involved...
06:50	companion_cube	if two threads read the atom, add something to the cache, and put the new value in the atom, one of the changes will be lost
06:51	edoloughlin	What if the atom _is_ the cache? (def cache (atom {}))
06:51	clgv	there is a memoize implementation that uses only atoms: http://kotka.de/blog/2010/03/memoize_done_right.html
06:51	clgv	maybe that might help you
06:53	ejackson	it depends what you are storing in the cache. if every call needs to be recorded then i think you need a ref
06:53	companion_cube	oh, right, there is a kind of compare-and-swap in atoms
06:53	companion_cube	so you can spot concurrent updates
06:53	edoloughlin	clgv: Looks interesting. This is my first venture into concurrency in Clojure. I'm trying not to to the noob thing of throwing everything at it.
06:54	raek	atoms are fins as long as you don't need to coordinate the change with some other change
06:54	ejackson	companion_cube: i did not know that. seems that in that case you would want a ref though, to take care of the concurrency collisions, no ?
06:54	clgv	edoloughlin: I learned a lot reading that post. I think the decisions are well explained there.
06:54	edoloughlin	Will read. Thanks.
06:54	ejackson	raek is exactly right. coordination is the differentiator.
06:55	ejackson	and you need to coordinate if >1 thread can alter data, and depend on another's value. such as an incrementer, say. at least i think so :)
06:55	raek	in memoize, the atom never leaves the function, so you know it will never have to be coordinated with someone else's changes
06:56	ejackson	if you (swap! atom-on-int inc) from two threads together, you will be one count short.
06:56	clgv	but multiple threads can use the function and the cache shall be share like explained in the post
06:56	raek	no, atoms are thread safe
06:56	ejackson	bollocks :)
06:57	ejackson	oops, as in not you're talking bollocks, but bollocks i'm wrong again.
06:57	raek	but if you want to change multiple atoms or one atom in multiple steps, then you can't do it atomically
06:57	raek	that's what refs add
06:57	clojurebot	Titim gan éirí ort.
06:58	ejackson	ah yes... that's it.
06:59	raek	my rule of thumb is: if you are not sure someone else might want to use your atom/ref in a multi-step or multi-ref change, use refs
07:00	companion_cube	i think you can do it with atoms, but refs (with STM) are much simpler to keep state consistent
07:44	qoja	I'm trying to reify a java.io.FilenameFilter with: (reify FilenameFilter (accept [_ name] (.endsWith name ".sql"))) and I get Can't define method not in interfaces: accept--what am I doing wrong?
07:46	clgv	qoja: I am not 100% sure, but shouldnt you use proxy instead of reify since reify is only intended for protocols?
07:46	qoja	aren't protocols interfaces actually ?
07:48	clgv	well, they are clojure constructs compiled to java interfaces and you are trying to implement a java interface. so there is a semantic difference here.
07:49	companion_cube	it's a bit of static typing, right ?
07:49	clgv	if you got the method-signature right and reify doesnt work I would definitely try proxy instead
07:49	clgv	I used it for TreeCellRender...
07:52	companion_cube	the blog entry on memoization is really interesting
07:56	qoja	oh what a silly mistake. forgot this-variable as first parameter!
07:57	raek	for interfaces, both proxy and reify can be used
07:57	clgv	raek: ah, good to know
07:57	raek	but reify is designed to be clojure-ish
07:58	raek	so you can't do subclassing with it, for instance
08:02	ejackson	and i believe is much faster
08:03	raek	it is.
08:03	raek	proxy also allows you to update the methods at run time
08:03	raek	(which I haven't used myself...)
08:19	ejackson	raek: i've played that trick through deftype. Its fantastic from the repl.
08:48	zippy314	Hey, guys, I know this isn't strictly a clojure question, but since I'm hanging out here and using it for colure dev... It's an emacs customization question. I've used technomancy's starter emacs repo, and it calls for removing my .emacs file and using .emacs.d. Where in that directory should I be putting my own cusomizations?
08:50	arbscht	zippy314: ~/.emacs.d/init.el is a reasonable choice. it is documented in greater detail at http://www.gnu.org/software/emacs/manual/html_node/emacs/Init-File.html
08:52	zippy314	tx!
08:55	fliebel	Hrm, I wonder if Logos, and logic programming in general could solve blue eyes. I guess it can, if you define the problem correctly.
08:55	ejackson	a hesitate to ask..... what is blue eyes ?
08:55	fliebel	http://xkcd.com/blue_eyes.html
08:56	fliebel	I have set it as my home page until I solve it :)
08:57	zippy314	fliebel: nice1
08:58	zippy314	(s/1/!/) correction: fliebel: nice!
08:58	ejackson	eep
08:59	fliebel	I have made a few assertions, but solved nothing yet...
09:00	fliebel	zippy314: If you'd have typed that regex without the cruft, sexpbot would have xxx it for you ;)
09:00	fliebel	s/xxx/replaced
09:00	sexpbot	<fliebel> zippy314: If you'd have typed that regex without the cruft, sexpbot would have replaced it for you ;)
09:01	zippy314	text
09:01	zippy314	(s/x/s/) correction: test
09:01	zippy314	hmm what cruft?
09:01	zippy314	I just typed "s/x/s/"
09:01	ejackson	s/ee/EEEEEEEEEEEEEE/
09:01	sexpbot	<ejackson> EEEEEEEEEEEEEEp
09:02	zippy314	Could it be my client?
09:02	ejackson	no you typed "correction:....." after the regex
09:03	fliebel	probably.
09:03	zippy314	nope, I didn't I think my irc client must be doing that or something...
09:03	ejackson	haha, its like a macro. your client got it before sexpbot did :)
09:03	fliebel	$sed -zippy314 s/'t/'t,/
09:03	sexpbot	<zippy314> nope, I didn't, I think my irc client must be doing that or something...
09:04	zippy314	I'm using Colloquy (OS x)
09:04	fliebel	$sed -zippy314 s/x/X/
09:04	sexpbot	<zippy314> I'm using Colloquy (OS X)
09:04	zippy314	:-)
09:04	fliebel	Okay, I'll to annoying people...
09:05	fliebel	s/to/stop/
09:05	sexpbot	<fliebel> Okay, I'll stop annoying people...
09:05	zippy314	is anybody else here using technomancy's emacs starter kit?
09:07	raek	I do.
09:08	zippy314	raek: where do you add your own emacs customizations?
09:11	clgv	fliebel: seems to be a proof by induction ;)
09:12	incandenza	zippy314: I use ~/.emacs.d/username.el
09:12	fliebel	clgv: uh?
09:13	clgv	fliebel: the blue eyes puzzle ;)
09:13	fliebel	clgv: the proof by induction ;)
09:13	zippy314	Looks like ~/.emacs.d/custom.el is also autoloaded for this purpose...
09:14	incandenza	that is actually used for M-x customize
09:14	incandenza	you probably shouldn't put stuff in there
09:14	zippy314	BTW if any body else is on a Mac and is annoyed by the default scrolling behavior of Carbon emacs, here's something that fixes it: http://www.emacswiki.org/emacs/SmoothScrolling
09:14	neilcj	zippy314: i throw any code in custom.el and it works, donno if that's ok :P
09:15	clgv	fliebel: you dont know the term, or what?
09:15	fliebel	clgv: I'm reading wikipedia now ;)
09:15	zippy314	incandenza: hmm, what's M-x customize for?
09:15	clgv	fliebel: ok^^
09:15	raek	zippy314: ~/.emacs.d/raek.el and ~/.emacs.d/raek/<oher files>.el
09:15	incandenza	zippy314: it's a way of setting configuration variables with a menu system
09:17	raek	"First, the user-specific-config file is the file named after your user with the extension ".el". In addition, if a directory named after your user exists, it will be added to the load-path, and any elisp files in it will be loaded."
09:17	raek	from https://github.com/technomancy/emacs-starter-kit
09:17	incandenza	^ yeah, follow that :)
09:17	neilcj	yeah, just moved stuff I'd added from custom.el to neilcj.el and everything worked: do that, zippy
09:42	markskilbeck	Anybody know of a trick in emacs to mimic aquamacs's ability to click a previously entered expression in the REPL to have it evaluated again?
09:44	neilcj	middle clicking an expression inserts it after the REPL prompt, if that helps
09:54	raek	hitting <ret> after a previous expression copies it to the bottom, iirc
09:54	raek	M-p can be useful too
09:58	markskilbeck	neilcj, raek: thanks
10:14	zippy314	raek: what do you do with M-p?
10:16	semperos_	go up one entry in REPL history in Emacs
10:17	zippy314	oh sweet! I need the slime repl cheat sheet.
10:18	zippy314	looks like there is one: http://www.pchristensen.com/slimecommands.pdf
10:25	raek	the slime commands listed here are really useful: https://github.com/technomancy/swank-clojure
10:31	hoggarth	do you have an easy way to convert a string to a list of characters?
10:33	ejackson	,(map identity "Hello, world!")
10:33	clojurebot	(\H \e \l \l \o \, \space \w \o \r ...)
10:33	hoggarth	nice
10:33	Chousuke	,(seq "hello world")
10:33	clojurebot	(\h \e \l \l \o \space \w \o \r \l ...)
10:33	Chousuke	use that :P
10:33	ejackson	nicer !
10:34	Chousuke	actually, since most seq-expecting things call seq on their argument anyway, you can often just pass in plain strings
10:34	hoggarth	i'll keep that in mind
10:35	Chousuke	,(first "foo") for example
10:35	clojurebot	\f
10:35	hoggarth	thanks guys
10:51	Kruppe	Is there any way of finding out what slots are in a class (short of using inspect or reading the source)?
10:52	ejackson	Kruppe: I usually use show from repl-utils
10:53	Kruppe	ejackson: I'll take a look at it, thanks :)
10:53	ejackson	although you should be careful with my advice, I haven't been correct yet today ;)
10:53	Kruppe	ejackson: these things happen
11:20	Luyt	My solution for Euler #1, http://clojure.pastebin.com/rw9zgSyt , I guess I could use some range that counts to 1000 and filter it on multiplep(3) OR multiplep(5). Would that be more elegant?
11:24	chouser	I like what you've got.
11:25	chouser	hm, essentially identical, though I started my ranges as 0
11:25	chouser	at 0
11:27	Luyt	Somehow expanding two lists, concatenating them, then making the numbers in it unique using 'set', seems like a detour to me.
11:28	Luyt	I'd rather have one seq which counts from 0 to 1000, selects only the multiples of 3 and 5 from them, and sums those.
11:29	incandenza	Luyt: you could do it like this: (reduce + (filter #(or (= 0 (mod % 3)) (= 0 (mod % 5))) (range 3 1000)))
11:29	Luyt	Aha, that looks more like it
11:31	chouser	Luyt: well, that means calling mod at least once on every number under 1000
11:32	incandenza	yes, it is slower, I noticed
11:32	chouser	what you've got now only touches numbers that will be included in the set
11:32	Luyt	Indeed ;-)
11:33	chouser	and remember both ranges and concat are lazy, so it's not really expanding two lists before building the set
11:33	Luyt	Oh? I didn't know that.
11:35	chouser	the set is what forces the realization of the lazy seqs in that expression, so really only one chunk of input numbers exists at any one time while the set is being built up
11:37	symuyn	Hey, is reverse on a lazy seq linear or constant?
11:38	chouser	linear. reverse itself isn't lazy and realizes the whole input seq
11:38	chouser	as it of course must
11:38	Luyt	Hmmm when I run it on a range of 3000000 Incandenza's seems to be a lot quicker. How could you easily measure it?
11:39	incandenza	I just used 'time'
11:39	semperos_	$doc time
11:39	semperos_	guess that only works for source...
11:39	Luyt	Ah, I'll try.
11:39	symuyn	chouser: Is it the same for reversing vectors?
11:39	Luyt	,(doc time)
11:39	clojurebot	"([expr]); Evaluates expr and prints the time it took. Returns the value of expr."
11:40	chouser	be careful doing micro benchmarks -- the JVM is complex and will give widely varying times, especially the first several times you run a particular chunk of code
11:40	chouser	symuyn: yes, but don't use reverse on a vector, use rseq
11:40	symuyn	Oh dear
11:40	symuyn	Thanks
11:41	Luyt	Incandenza's code is quicker, at least on my machine. 588 msec vs. 3700 msec.
11:42	incandenza	probably the mod is cheaper than the set membership test, in the long run
11:42	Luyt	Yeah, because with smaller sets my code is slightly quicker
11:43	incandenza	and I guess the set itself can't be lazy
11:43	Luyt	Well, that's an interesting side effect ;-) On to problem 2...
11:47	chouser	set can't be lazy, but two sorted lazy seqs could be merged lazily, and adjacent dups could be removed lazily
12:16	chouser	Luyt: I added a solution to problem 1 at the bottom of http://clojure-euler.wikispaces.com/Problem+001 if you're interested.
12:18	Luyt	Interesting to see all these solutions
12:18	Luyt	1000 ways to skin a cat
12:20	Luyt	I see many people used the reduce, set, concat thing.
12:21	chouser	Luyt: they all either used a set, distinct (which uses a set internally), or subtracted the duplicate sum afterward
12:22	chouser	that's what I thought it might be interesting to have the merge solution there -- it's the only one that skips the dups instead of subtracting them, but without building a full set
12:23	Luyt	Right now I noticed by homebrewn fibonacci routine is dog slow, and my jaw drops when looking at http://en.wikibooks.org/wiki/Clojure_Programming/Examples/Lazy_Fibonacci ;-)
12:23	chouser	I'm not saying it's better, just different. Ends up being a lot of code for such a simple problem. :-P
12:25	symuyn	If I want to make an empty collection as a base case for a sequence-returning function, should I use the empty list or the empty vector?
12:26	chouser	symuyn: well, are you building a vector, a list, or a lazy-seq?
12:27	symuyn	A lazy seq
12:27	chouser	you don't want nil for the base case/
12:27	chouser	?
12:27	pdk	ok dumb q
12:28	symuyn	chouser: Oh, of course. Silly me
12:28	pdk	what out of the box function would i use to append an item to a seq
12:28	pdk	i see there's cons to prepend but i figured i could append in one step too
12:29	chouser	pdk: lazily without forcing it? a single item, or will it end up being many?
12:29	pdk	one item at a time
12:29	pdk	it can be lazy
12:30	pdk	let me guess it involves wrapping the item in () then doing lazy-cat :p
12:30	chouser	well, concat then, but if you're doing it a lot that's rather inefficient
12:30	pdk	what would you suggest to optimize it or change the strategy
12:30	mattmitchell	could someone recommend a lib for dealing with command line args?
12:31	chouser	pdk: use a vector instead of a lazy seq -- you can efficiently add to the right
12:32	chouser	pdk: or use lazy-seq directly
12:35	pdk	yeah i was thinking a vector though i didnt find anything right away to append to them
12:35	semperos_	conj
12:35	chouser	conj appends to vectors. very very fast
12:35	pdk	could always use lazy-concat in place of concat too
12:35	pdk	hm here i was using conj on sets
12:35	pdk	er lazy-cat
12:36	semperos_	conj is the cons of vectors
12:36	semperos_	and vectors efficiently add items to the right
12:36	semperos_	while lists to the left (head)
12:36	chouser	conj works on everything: sets, maps, vectors, lazy seqs, cons, ...
12:39	Luyt	Is there a pastebin/codepad somewhere which will execute clojure code?
12:39	pkinney	is it ok to paste a one-liner piece of code here?
12:39	tonyl	pkinney, one-liners are fine
12:39	pkinney	tonyl: thanks
12:40	pkinney	(map #(bit-xor (int (get %1 %2)) (int (get %1 (- 1 %2)))) lock (iterate inc 0))
12:40	semperos_	put & at the beginning
12:40	tonyl	Luyt maybe try-clojure but don't think it would return a paste url
12:40	Luyt	,(map #(bit-xor (int (get %1 %2)) (int (get %1 (- 1 %2)))) lock (iterate inc 0))
12:40	clojurebot	java.lang.Exception: Unable to resolve symbol: lock in this context
12:40	pkinney	lock is a string
12:40	Luyt	tonyl: I'll check.
12:41	Luyt	Argh... "The service is not available. Please try again later." http://tryclj.licenser.net/
12:41	tonyl	try this http://www.try-clojure.org/
12:41	pkinney	,(map #(bit-xor (int (get %1 %2)) (int (get %1 (- 1 %2)))) "testing" (iterate inc 0))
12:41	clojurebot	java.lang.NullPointerException
12:42	pkinney	I think its throwing the error because the int is getting a nil on the last iteration
12:42	tonyl	there is no last iteration
12:42	tonyl	is a lazy sequence
12:42	pkinney	(sorry the 1st)
12:42	Luyt	tonyl: Thanks, that site is alive, however it won't accept opening parentheses. Could be a Konqueror issue ;-)
12:42	tonyl	&(take 100 (iterate inc 0))
12:42	sexpbot	⟹ (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)
12:43	Luyt	quick! What't the sum of that? (Gauss)
12:43	tonyl	the map function to apply to the items only takes 1 argument though
12:44	tonyl	Luyt: crap forgot that, 5015?
12:44	Luyt	N=100, N/2 * 100 - N
12:44	tonyl	&(get 0 1)
12:44	sexpbot	⟹ nil
12:44	chouser	&(get 0 1 :lunchtime)
12:44	pkinney	tonyl: I thought if u gave the map more args they run in parallel
12:44	sexpbot	⟹ :lunchtime
12:45	Luyt	Gauss solved that in a few seconds when he was a little boy (his schoolmaster gave him the silly assignment of adding the numbers 1 through 100 just to keep him busy)
12:45	tonyl	pkinney: if you give it more then 1 collection ##(map vector (range 0 10) (range 11 20))
12:45	sexpbot	⟹ ([0 11] [1 12] [2 13] [3 14] [4 15] [5 16] [6 17] [7 18] [8 19])
12:48	pkinney	yeah
12:48	tonyl	Luyt: found this in my delicious http://ideone.com/ I knew I saw a service like that somewhere
12:48	tonyl	pkinney: my bad, I forgot about the string there
12:48	pkinney	I thought I could use the other coll (from iterate) as an index
12:49	Luyt	tonyl: Looks very good, thanks.
12:49	tonyl	it seems that the error might be in 1 of those gets
12:49	tonyl	&(map vector "testing" (range))
12:49	sexpbot	⟹ ([\t 0] [\e 1] [\s 2] [\t 3] [\i 4] [\n 5] [\g 6])
12:50	pkinney	oh, does range work that way too, lol
12:50	pkinney	thats cool
12:51	tonyl	yeah i use it for shortcut
12:51	tomoj	&(map-indexed vector "testing")
12:51	sexpbot	⟹ ([0 \t] [1 \e] [2 \s] [3 \t] [4 \i] [5 \n] [6 \g])
12:51	amalloy	do you really want (- 1 %2)? in infix that's 1 - %2
12:51	pkinney	yeah
12:52	tonyl	(dec %2)
12:52	sexpbot	⟹ -1
12:52	tonyl	&(dec %2)
12:52	sexpbot	java.lang.Exception: Unable to resolve symbol: %2 in this context
12:52	amalloy	lol
12:52	tonyl	lol
12:52	pkinney	its xoring the letter together
12:52	amalloy	tonyl: you activated the karma plugin
12:52	tonyl	yeah
12:52	tonyl	wonder who is %2
12:52	amalloy	tonyl: nobody
12:52	tonyl	&(map (fn [[x y]] [(int (get x y)) (int (get x (dec y))]) "testing" (range))
12:52	sexpbot	java.lang.Exception: Unmatched delimiter: ]
12:52	amalloy	but just in case he signs on
12:53	pkinney	eg. t ^ e = 3... 3 ^ s... etc (3 is made-up)
12:54	amalloy	pkinney: and you have negative numbers as keys in the maps?
12:54	pkinney	oh, i'm wrong sorry... t ^ e, e ^ s...
12:54	tonyl	&(map (fn [[x y]] [(int (get x y)) (int (get x (dec y)))]) "testing" (range))
12:54	sexpbot	java.lang.IllegalArgumentException: Wrong number of args (2) passed to: sandbox17750$eval19737$fn
12:55	pkinney	amalloy: what do you mean?
12:55	amalloy	&(- 1 5)
12:55	sexpbot	⟹ -4
12:56	amalloy	my point is that (- 1 x) will be negative for any x larger than 1
12:56	Luyt	Well, my solution for Euler #2. It contains some suckage...
12:56	amalloy	and you're using the result of as a key in your maps
12:56	Luyt	http://ideone.com/iCCOS
12:56	pkinney	aaah
12:56	tonyl	&(int (get \t 0)
12:56	sexpbot	java.lang.Exception: EOF while reading
12:56	tonyl	&(int (get \t 0))
12:56	sexpbot	java.lang.NullPointerException
12:57	tonyl	&(int (get \t 1))
12:57	sexpbot	java.lang.NullPointerException
12:57	tonyl	pkinney: that might be your problem
12:57	pkinney	yeah, thats it
12:58	pkinney	(need to find a way outa that
12:58	amalloy	pkinney: ##(dec 5)
12:58	sexpbot	⟹ 4
12:59	pkinney	It'll strill try and get the -1 index though
13:00	tonyl	i am guessing because of (dec 0)
13:00	pkinney	,(map #(bit-xor (int (get %1 %2)) (int (get %1 (dec %2)))) "testing" (iterate inc 0))
13:00	clojurebot	java.lang.NullPointerException
13:00	pkinney	yup
13:00	Luyt	I think these Project Euler problems focus too much on math and not on real world problems
13:01	pkinney	is there a quick way to get it to give a 0 if the index doesn't exist
13:01	tonyl	that is why is called Euler :P
13:01	qoja	is there some short macro, to merge metadata (and not replece, like with-meta does) ?
13:01	Luyt	I mean, to write a web app you don;t have to solve the sum of the first 10001 primes who are amicable
13:01	amalloy	Luyt: so don't solve them
13:01	tonyl	qoja metadata is map and there is a merge fn for maps
13:02	amalloy	pkinney: ##(get {} -1 0)
13:02	sexpbot	⟹ 0
13:02	Luyt	amalloy: Excellent idea!
13:02	tonyl	Luyt: it even says you can do them in paper
13:02	qoja	tonyl: but what with (with-meta (fn .... ) (merge ??? { :my 'metadata }))
13:02	tonyl	the not-found arg I always forget it too
13:02	qoja	i could of course write a fun for that, but isn't there one standard already?
13:03	Luyt	tonyl: My dad studied math and he solved some problems by just analysing them... while I pull out my fav language and go bruteforce it ;-)
13:03	pkinney	thanks
13:03	tonyl	Luyt: I wish I could do that like your dad, but I like bruteforcing every now and then
13:04	tonyl	qoja: as far as I know there isn't
13:04	raek	qoja: vary-meta
13:04	raek	,(doc meta)
13:04	Luyt	tonyl: Yup. But not in all cases bruteforcing is even remotely possible.
13:04	clojurebot	"([obj]); Returns the metadata of obj, returns nil if there is no metadata."
13:04	raek	,(doc vary-meta)
13:04	clojurebot	"([obj f & args]); Returns an object of the same type and value as obj, with (apply f (meta obj) args) as its metadata."
13:04	qoja	tonyl: ok thanks, i thought it maybe is present int contrib. it seems to be a common scenario
13:05	tonyl	yeah and raek's suggestion might help
13:05	qoja	raek: that's what i need
13:05	qoja	thx
13:06	amalloy	&(vary-meta #'first (comp merge meta) #'comment)
13:06	sexpbot	java.lang.IllegalArgumentException: Wrong number of args (2) passed to: core$meta
13:06	amalloy	oh duh
13:07	amalloy	&(vary-meta #'first merge (meta #'comment))
13:07	sexpbot	java.lang.ClassCastException: clojure.lang.Var cannot be cast to clojure.lang.IObj
13:07	tonyl	what's up with var's private meta? what is it used for?
13:09	raek	it stops people from making functions that use those vars outside the namespace they live in
13:10	tonyl	ok thanks raek
13:11	raek	,clojure.core/array
13:11	clojurebot	java.lang.IllegalStateException: var: clojure.core/array is not public
13:11	raek	...or from even using it from the outside of the namespace
13:11	raek	,@#'clojure.core/array ; with exceptions... :-)
13:11	clojurebot	#<core$array clojure.core$array@1479a6b>
13:12	symuyn	&(seq? nil)
13:12	sexpbot	⟹ false
13:13	pdk	,(Integer/parseInt "3")
13:13	clojurebot	3
13:13	pdk	,(Integer/parseInt 3)
13:13	clojurebot	java.lang.ClassCastException: java.lang.Integer cannot be cast to java.lang.String
13:26	mattmitchell	how can i check if a variable exists?
13:28	mattmitchell	i guess i mean, how can i check if a var exists?
13:28	raek	,(doc resolve)
13:28	clojurebot	"([sym]); same as (ns-resolve ns symbol)"
13:28	kumarshantanu	mattmitchell: the compiler checks that for you :-p
13:28	raek	maybe this
13:28	mattmitchell	interesting
13:29	raek	you very rarely need to use it though...
13:29	mattmitchell	makes me think i might be handling this bit of code incorrectly then :)
13:38	bendlas	key folks!
13:38	bendlas	/s/key/hey/
13:39	bendlas	fun clojure quirk of the day
13:39	bendlas	,(loop [x 5]
13:39	bendlas	(when (pos? x)
13:39	bendlas	(try
13:39	clojurebot	EOF while reading
13:39	bendlas	(recur (dec x))
13:39	bendlas	(finally (print x)))))
13:39	bendlas	,(loop [x 5] (when (pos? x) (try (recur (dec x)) (finally (print x)))))
13:39	clojurebot	bendlas: No entiendo
13:40	chouser	the bots don't like you using try/catch
13:40	bendlas	OIC
13:41	chouser	they could be used to prevent killing of runaway exprs and such.
13:41	bendlas	guessed that
13:41	bendlas	anyways, the expression compiles, but the finally is silently discarded
13:41	bendlas	due to the recur
13:42	hiredman	bendlas: thats a bug, the latest clojure won't compile the expression
13:42	raek	,((fn f [x] (when (pos? x) (try (f (dec x)) (finally (print x))))) 5)
13:42	clojurebot	raek: Huh?
13:42	hiredman	recur is not allowed across try/binding
13:42	raek	right. the catch form.
13:43	bendlas	hiredman: so it's on the radar already, good
13:45	hiredman	bendlas: recuring across a dynamic context like try or binding is a buggy way of thinking, the compiler was just loose in allowing you to shoot your foot off
13:49	bendlas	actually, I need a variation of (binding ) that works on #'var s, with each binding expression in a try
13:49	bendlas	but using recur there smelled fishy right away, so i checked first
13:49	bendlas	luckily
13:51	hiredman	sounds icky
13:52	bendlas	it does
13:53	bendlas	and of course there is a simpler way, which i've just realized
13:55	pkinney	is there an easy way to add an item to the beginning of a coll? like a reversed conj?
13:55	bendlas	(cons item collection)
13:55	chouser	pkinney: conj on a list adds items to the left
13:55	hiredman	conj add items in the most effecient way
13:55	hiredman	,(conj [1] 2)
13:55	clojurebot	[1 2]
13:55	chouser	,(conj '(b c d) a)
13:55	clojurebot	java.lang.Exception: Unable to resolve symbol: a in this context
13:55	chouser	,(conj '(b c d) 'a)
13:55	hiredman	,(conj '(1) 2)
13:55	clojurebot	(a b c d)
13:55	chouser	sigh
13:55	clojurebot	(2 1)
13:55	chouser	sorry
13:56	pkinney	chouser: ah rite, thanks :)
13:58	pkinney	(how didn't I notice that? heh)
14:12	pkinney	gaaah, I'm still stuck in the mold (of OOP)
14:14	amalloy	pkinney: OOP backwards is POO. keep that in mind
14:15	pkinney	hehe
14:15	pkinney	I just think the overall design of my functions are wrong
14:17	pkinney	https://gist.github.com/799022
14:17	amalloy	could be. everything's a learning experience
14:17	pkinney	my ugly function
14:17	pkinney	I'm abusing let
14:18	amalloy	pkinney: fwiw (list* s) should be the same as just plain s, unless you want to coerce some other sequence to be specifically a list
14:19	pkinney	Oh
14:19	amalloy	and there's nothing wrong with using let to label intermediate results, although you might want to pick names that mean something :P
14:20	pkinney	heh, yeah a to d isn't good for variable names
14:20	amalloy	if you just want to reduce the nesting factor of your intermediate results, you can use -> or ->> instead
14:21	pkinney	I've seen other code snippets about using -> and ->>, thought they were interesting, but haven't got my head round them enough to using them
14:21	amalloy	(->> (count lock) (reduce some-fn) (conj (map other-fn other-data)) (map nibble-swap) (map char) (apply str))
14:21	pkinney	does it just push the result onto the next...
14:21	pkinney	aah
14:22	amalloy	&(macroexpand-all '(->> (range) (filter even?) (drop 100) (take 100) (reduce +)))
14:22	sexpbot	⟹ (reduce + (take 100 (drop 100 (filter even? (range)))))
14:22	pkinney	could take out the let altogether then?
14:22	zippy314	I'm trying to remove an item from a ref to a vector. I can't figure out why this doesn't work: ##(def x (ref [1 2 3 4])) (dosync (alter x remove [#{2}]))
14:22	sexpbot	java.lang.SecurityException: You tripped the alarm! def is bad!
14:22	zippy314	oops
14:23	amalloy	zippy314: #{2}, not [#{2}]
14:23	chouser	zippy314: x bevomes the first arg to remove, which is probably not what you want
14:23	amalloy	oh, or chouser is so right
14:24	pkinney	amalloy: thanks, thats gave me some food for thought
14:24	amalloy	pkinney: welcome, and enjoy it. i'm off
14:24	pkinney	me too ^_^ thanks again
14:25	zippy314	amalloy: that doesn't work either: ##(let [x (ref [1 2 3 4])] (dosync (alter x remove #{2})) @x)
14:25	sexpbot	⟹ ()
14:27	zippy314	so, I'm trying to figure out what's the proper way to remove an item from a ref to a vector?
14:28	amalloy_	zippy314: chouse [as usual] has the right answer
14:28	amalloy_	(alter x #(remove #{2} %))
14:29	chouser	zippy314: removing items from a vector is suspect in the first place. you're sure you don't want a set instead?
14:29	tonyl	the way alter is calling is like (remove @x #{2})
14:29	shortlord	is directly using an (immutable) var in a function without passing the var as an argument considered to be impure?
14:29	tonyl	yes, but you can still do it
14:29	zippy314	Well, the thing is I care about the order.
14:29	hiredman	:(
14:29	zippy314	is there and ordered set?
14:29	chouser	there's a sorted set
14:29	hiredman	(alter x (partial remove #{2}))
14:31	zippy314	yah baby! that's what I really want, sorted-set. Cool.
14:31	jweiss_	can someone point me to a typical dir structure for a project with java and clojure source? (/src/java, /src/clj)? (java/src, clojure/src)? etc
14:31	chouser	note that remove will return a lazy seq, not a vector, and because it's lazy has the potential to stack calls when later realized
14:32	zippy314	but how do you define the sort algorithm?
14:32	shortlord	tonyl: but using other functions directly would not be considered impure, right?
14:32	chouser	zippy314: see sorted-set-by, but be aware that if your fn says two things are equal, only one will be stored in the set
14:32	jweiss_	clojure itself uses src/clj and src/jvm, oddly
14:33	chouser	jweiss_: the maven standard is something like src/main/clojure and src/main/java
14:33	tonyl	shortlord: what do you mean directly
14:33	technomancy	jweiss_: consider why you feel that Java code needs to be in a separate dir
14:33	chouser	...and there's the non-maven voice. :-)
14:34	jweiss_	technomancy: i suppose that was a false assumption that it DOES have to be separate?
14:34	chouser	jweiss_: it'll depend on your tools.
14:34	shortlord	tonyl: well, where's the difference between using vars or functions in another function without passing them as arguments? using built-in functions like map is not considered to make a function impure, so why is using an immutable var?
14:34	technomancy	jweiss_: maybe you have a good reason, but I don't think most people do
14:34	technomancy	worth some thought =)
14:34	jweiss_	i'm using leiningen to build, so i assume if the creator of lein says its ok then lein won't complain either
14:34	jweiss_	:)
14:35	jweiss_	i suppose that if it doesn't work out, separating the two types of files won't be hard.
14:38	tonyl	the functions you pass to map won't change anything, it would just create a new value and map would save those values in a seq
14:38	tonyl	vars are global objects
14:38	technomancy	jweiss: for the record I haven't actually tried that; I have noticed that it hasn't been a problem to keep JS files in src.
14:38	tonyl	I think I might be missing something here
14:39	chouser	does lein know how to compile .java files?
14:39	technomancy	chouser: yeah, added in the latest version
14:39	jweiss_	technomancy: as long as lein will still know to compile java, and put the class files in the jar, along with the clj files, i'll be happy
14:39	shortlord	tonyl: no, I mean using 'map' inside one of your functions without passing the map function as an argument. (quite theoretical, I know). Just using clojure's core functions is conceptually not different to using your own functions without passing them as arguments
14:40	shortlord	tonyl: and since vars could be functions or anything else but are in any case immutable, I don't see why just using them in a function without passing them as an argument makes the function impure
14:41	tonyl	shortlord: oh that is a good question about the purity or functional langs. in the strictest of meanings that is impure
14:43	tonyl	shortlord: but as you say you would need some impurity to use the language functionalities. I would say AFAIK in clojure you can use global functions inside your functions, but when it comes to other values in vars, depend less on those.
14:43	tonyl	but that is my opinion
14:43	shortlord	I ask about that quite theoretical thing, because I still have problems determining how pure any particular function in my code should be. For small problems it's quite easy to stay very pure, but in a larger project all the functions seem to require more an more arguments to stay pure
14:45	tonyl	good question, but if you want to stay really pure then you can only use the special forms that clojure comes with, which I think are about 12.
14:46	tonyl	that would be a tedious language to use, instead of using of the available library functions like the seq, maps, vectors, etc ... functions
14:47	shortlord	tonyl: I am not very interested in staying pure for the sake of staying pure, but I have the feeling that I'm often still doing things too imperatively. Whenever I try to actively write pure code, it's very hard to determine when to pass which things as arguments and when to use something directly
14:48	shortlord	is there any good resource on software development using functional programming that is largely language agnostic and goes beyond explaining the functional concepts in terms of isolated 5 - 50 lines examples?
14:49	tonyl	shortlord: I battle with that thinking sometimes. but I haven't stumble upon a source like that. I have just started reading the SCIP book and it seems quite good in functional concepts, although it uses scheme to illustrate
14:50	tonyl	I have just started with functional programming with clojure
14:50	technomancy	scheme is as close to language-neutral as you can get
14:51	tonyl	technomancy: yeah, I can understand about 80% of what the code means. maybe because I have some clojure exposure
14:51	shortlord	I've already heard quite a few good things about SICP, I guess I should really put it on my to-read list
14:52	shortlord	thx for the suggestion :)
14:53	edw	shortlord: And watch the H-P lectures from the '80s by Sussman & Abelson.
14:53	benreesman	what is the idiomatic way in clojure to assoc a key that's nested in a map, perhaps several levels in?
14:53	semperos	assoc-in, I think
14:54	semperos	,(doc assoc-in)
14:54	clojurebot	"([m [k & ks] v]); Associates a value in a nested associative structure, where ks is a sequence of keys and v is the new value and returns a new nested structure. If any levels do not exist, hash-map...
14:54	shortlord	benreesman: if you simply want to assoc it, use assoc-in, otherwise update-in can operate on it using a function
14:54	benreesman	thank you!
14:57	benreesman	is there a way to have it add the keys if not present as maps?
14:58	semperos	part of the docs for assoc-in
14:58	semperos	"If any levels do not exist, hash-maps will be created"
14:59	semperos	that what you meant?
15:01	benreesman	definitely
15:02	benreesman	brilliant, works like a charm
15:16	sritchie	hey all -- I'm looking to work with a java function that takes a byte array as a parameter and fills it with data. this sort of goes against the clojure way, as data structures are supposed to be immutable -- also, when I pass in the symbol for my array (defined with (def array (byte-array (* 1200 1200)))), it doesn't change
15:17	sritchie	what's the most idiomatic way to deal with a java method like this?
15:19	tonyl	i don't think that java fn would know how to change the value of a binding to the var that is holding the array
15:20	brehaut	tonyl: by the byte array itself is a mutable block of memory
15:20	tonyl	mmm... I haven't done anything like this, but maybe somebody here hve
15:20	brehaut	s/by/but/
15:20	sexpbot	<brehaut> tonyl: but the butte array itself is a mutable block of memory
15:20	sritchie	got it, so it's modifying the array, but can't switch the binding over
15:21	brehaut	sexpbot, that wasnt a general regexp. why did you apply it twice?
15:21	brehaut	sritchie: no
15:21	tonyl	thankr brehaut, I clearly need more knowledge about this then :)
15:22	sritchie	the other option is a method that returns a DirectByteBuffer
15:22	brehaut	as i understand it, all array types are mutable, and largely exist for interop (and speed)
15:22	tonyl	I can see that in the code now
15:23	brehaut	i think something else is going wrong to cause sritchie's array to not be modified
15:24	brehaut	sritchie: also idiomatically, i think you would try to isolate the code that worries about mutation as much as reasonably possible.
15:24	brehaut	ie keep it internal to one function
15:25	amalloy_	brehaut: sexpbot does a global replace. he's not perl, awk, or sed, he's just looking for s/foo/bar (optional trailing slash)
15:25	sritchie	brehaut: I'm reading MODIS tiles from nasa, enclosed in these dataset objects -- reading this byte array is the last step before I can free myself from mutable arrays
15:26	brehaut	sritchie: thats fine then. java interop is 'first class' code in clojure
15:26	brehaut	amalloy: apparently so :P
15:27	tonyl	sritchie: you can use java structures in clojure functions too
15:28	sritchie	well, looks like my first step is figuring out why that byte array isn't being modified
15:29	tonyl	maybe the java function is not actually changing it?
15:30	brehaut	sritchie: does you java function have some api docs (or better source) on the internet?
15:30	brehaut	and can you paste your snippit somewhere?
15:30	sritchie	yeah, here's the doc -- http://gdal.org/java/org/gdal/gdal/Band.html#ReadRaster_Direct(int, int, int, int)
15:30	sritchie	brehaut: pardon, here it is -- http://gdal.org/java/org/gdal/gdal/Band.html
15:32	sritchie	https://gist.github.com/799187
15:32	sritchie	that's how I was calling it
15:33	sritchie	I'm going to go verify that the Band has good data in it
15:33	brehaut	sritchie: thats a good idea.
15:33	tonyl	that method needs a ByteBuffer object
15:34	brehaut	tonyl: its heavily overloaded
15:34	sritchie	tonyl: yeah, check the example at the very top, it just needs a byte array
15:34	tonyl	yeah I scrolled further more
15:34	tonyl	i see the method signature now
15:35	tonyl	maybe type hinting would help resolve which method is being choose to operate
15:36	sritchie	tonyl: looks like the method is returning CE_None, which shows that all went well -- I'm going to try this with a different dataset, and see what happens
15:36	tonyl	ok
15:36	sritchie	but the fact that the method can modify a byte[] created with byte-array is great
15:36	sritchie	that simplifies this application quite a bit
15:37	brehaut	sritchie: are any exceptions being raised about finding methods?
15:38	brehaut	im presuming not, otherwise you probably would have mentioned them before now?
15:38	sritchie	brehaut: no, the application finds the method, and returns 0, which is CE_None (meaning no failure), but the byte array returned the same value from the REPL
15:39	sritchie	I assumed that it hadn't been modified, since (take 100 gdal-bytes) resulted in 0s, but based on these answers and the CE_None, I'm worried that I chose an empty band
15:42	brehaut	sritchie: use (some (complement zero?) gdal-bytes)
15:42	brehaut	it will return true if there is a non zero value and nil otherwise
15:43	brehaut	instead of take 100
15:44	sritchie	yeah, I do get true
15:44	brehaut	100 bytes out of 1400000 is not a very reliable sample
15:44	amalloy	or (not-every? zero? bytes)
15:44	brehaut	amalloy cleans up my mess again :P
15:44	amalloy	i think that exists. maybe not. ##(doc not-every?)
15:44	sexpbot	⟹ "([pred coll]); Returns false if (pred x) is logical true for every x in coll, else true."
15:44	amalloy	whew
15:45	brehaut	amalloy: and it works right too
15:45	tonyl	lol not-every is just (comp not every?) oh i love it, so simple
15:46	sritchie	hey, great, this --
15:46	sritchie	(some #(if (or (zero? %) (neg? %)) nil %) my-bytes)
15:46	sritchie	returns 30
15:46	tonyl	so there are values
15:46	sritchie	yup, looks like all's well
15:46	tonyl	:) I learned from this one too
15:47	sritchie	so! the lessons of today are -- byte arrays are mutable, and don't use (take 100 array) to investigate a 1200 x 1200 array
15:47	brehaut	tonyl: i learnt from tutoring at university that helping other people is always the best way to learn
15:48	tonyl	brehaut: so true
15:48	brehaut	sritchie: you next lesson is to take that test case, and write a deftest for it
15:49	sritchie	brehaut: sounds like a plan, that's a great idea
15:49	brehaut	which reminds me that i want to look into midje more
15:49	sritchie	thank you guys for the help, the clojure community's been fantastic
15:57	brehaut	it is definately time for a pot of tea
15:57	amalloy	brehaut: make some for sritchie too
15:58	amalloy	sritchie: btw that #(if ...) thing you did earlier is a complicated version of when-not
16:00	brehaut	i would pay money for a good A3 cheatsheat/quickref poster
16:00	amalloy	&(when-not (zero? 0) 10)
16:00	sexpbot	⟹ nil
16:01	sritchie	amalloy: does when-not return the value that it stops at?
16:01	amalloy	no, when-not is just a tweaked version of if, not of some
16:01	sritchie	gotcha
16:01	amalloy	you might also look at (first (filter #(when-not ...)))
16:01	brehaut	sritchie: there is a large family of if and when functions
16:02	amalloy	or i guess (first (filter pos? coll))
16:02	amalloy	now that i've actually read your or expression :P
16:03	sritchie	hey, I like that better
16:03	amalloy	if you didn't, i'd suggest getting your head examined :)
16:05	sritchie	especially since my test was sort of meaningless, just meant to reassure me that this damned data exists, I can (take 10 (filter pos? coll)) to my heart's content
16:06	amalloy	&(->> (range) (filter even?) (filter odd?) first)
16:07	amalloy	my favoritest infinite loop
16:07	sexpbot	Execution Timed Out!
16:07	brehaut	huh. so it is
16:07	amalloy	find me the first odd even number, plz
16:08	sritchie	very nice
16:09	tonyl	wouldn't that be just 0
16:12	brehaut	,[(odd? 0) (even? 0)]
16:12	clojurebot	[false true]
16:13	amalloy	one that's a little sneakier is (->> (range) (map (partial * 2)) (filter odd?))
16:16	sritchie	One quick question -- for that particular method we discussed, is it preferable to convert the java array into a vector, after the mutation?
16:16	tonyl	I guess I meant mathematically speaking
16:17	tonyl	sritchie: to use it in clojure world?
16:17	sritchie	yeah, with incanter -- I need to cut the array up and form 2d time series arrays
16:17	brehaut	tonyl: ok. how about this. Oddness is defined in terms of evenness. Evenness is defined as devisible by 2 with no remainder
16:17	brehaut	oddness is the complement of evenness
16:18	tonyl	so 0 can't be use as a value if it is the pivot used to test evenness
16:19	raek	sritchie: the functions in clojure's sequence library will call seq on the argument, so to use arrays with map, filter, reduce, et.c. you donät have to do anything
16:19	tonyl	but we need practicality in programming so that would suffice :P
16:19	raek	*don't
16:19	brehaut	tonyl: i dont understand?
16:19	sritchie	raek: but is it preferable to convert to vector to get immutability?
16:20	brehaut	sritchie: someone has to go out of their way to mutate the array
16:20	brehaut	using (frinstance) aset
16:20	brehaut	or passing it to some java
16:20	sritchie	brehaut: got it, and since I'm writing the code, (vector ... ) would just be an extra operation
16:20	brehaut	yes, although you'd use vec
16:21	tonyl	I mean we are using 0 in the implementation of even? to test if 0 is even or not
16:22	raek	if you can't know whether the other code in your progam (or the libraries it uses) will respect your wish for it to remain constant, then I guess you need to create a immutable version of it
16:22	brehaut	tonyl: in what sense? (defn even? [n] (= 0 (mod n 2))) would be my first guess at the imp
16:23	raek	if it never escapes your function in array form, there's no need to worry
16:23	sritchie	raek: got it, makes sense
16:23	raek	you could just call seq on it to hide the fact that the data is in a vector
16:23	tonyl	brehaut: but you are trying to find the evenness of 0 by testing an operation (mod) against 0
16:24	brehaut	thats fine
16:24	brehaut	surely
16:24	tonyl	yeah it is
16:25	tonyl	I am just saying that I was thought in math classes that 0 can be even or odd and positive or negative
16:25	brehaut	im not sure i understand how that invalidates the 0 as a part of the series of even numbers though
16:26	hiredman	http://en.wikipedia.org/wiki/Parity_of_zero
16:26	hiredman	tonyl: that is not true
16:27	brehaut	hiredman: huh, scales make the meaning even particular intuitive
16:28	tonyl	hiredman brehaut: thanks for the info. I am going to have to send a very angry letter to my math professors now
16:29	brehaut	that would be a funny letter :P
16:29	hiredman	please do, be sure to also complain if you were taught that multiplication is repeated addition
16:30	tonyl	a quote from the wiki page 'For them the question exemplifies "common knowledge ... that any well-educated adult should have"'
16:31	tonyl	i guess i am not an adult yet then
16:31	tonyl	hiredman: yeah they did that too
16:31	pdk	,(Math/pow 2 50)
16:31	clojurebot	1.125899906842624E15
16:31	pdk	great
16:31	hiredman	http://scienceblogs.com/goodmath/2008/07/teaching_multiplication_is_it.php
16:32	brehaut	obligatory xkcd http://xkcd.com/123/
16:33	hiredman	that is a good one, I'd not seen it before
16:34	brehaut	it was footnoted in the most recent one
16:34	brehaut	i had forgotten about it
16:34	brehaut	(being bad at both physics and bond films)
16:40	pdk	,(Math/pow 2 25)
16:40	clojurebot	3.3554432E7
16:58	sritchie	hey guys, is it preferable to typehint on java interop as a matter of course (for a third party API, say), or is it best to do it only on performance bottlenecks, as the clojure documentation suggests?
17:03	tonyl	performance bottlenecks except for swing, swing doesn't like reflection
17:03	Raynes	Swing doesn't like humanity.
17:07	joly	,(doc gcd)
17:08	clojurebot	Gabh mo leithscéal?
17:08	ossareh	is there a tool that does the reverse of hiccup? i.e. reads some html and gives me a bunch of vectors?
17:08	brehaut	ossareh: enlive
17:08	ossareh	brehaut: of course!
17:09	bendlas	only it gives you {:tag :html :attrs {} :body []}
17:09	raek	enlive<->hiccup https://gist.github.com/633049 :-)
17:10	brehaut	raek: nice :)
17:10	raek	there are other wrapping libs than Enlive that use Tagsoup for parsing
17:11	raek	some of them output data in hiccup style
17:11	ossareh	That is awesome
17:12	ossareh	So, while we're on this - my app has the need to create html - which I was hoping to later on postwalk to replace certain things in the tree - i.e. if it see <r:user:name> it knows to replace it with (:name user). Is there such a thing already?
17:13	raek	enlive can match and replace elements
17:13	ossareh	awesome
17:13	raek	it is often used as a html templating engine
17:13	raek	that's the thing it does besides the parsing
17:14	bendlas	indeed and it's the winningest html templating engine i know of, btw
17:14	konr	I've got a package "foo" with a global bar and many different function sets operating on it so that there are namespaces foo.fun1, foo.fun2 etc. Where should I put bar? In a new namespace like foo.variables? Or I can :use foo.core on foo.fun1 etc and vice versa?
17:14	raek	example for a guestbook: https://github.com/raek/lcug-guestbook/blob/master/src/se/raek/lcug/guestbook/view.clj
17:15	ossareh	huh - I've been using hiccup for the past 2 or so months. Is enlive significantly better suited to this type of thing? (i.e. is it worth the time to convert)
17:15	raek	running server: (beware of retro 90's html, your eyes might hurt a bit) http://services.raek.se/guestbook/
17:16	ossareh	It is very likely that much more of my application will have the html created in something like TinyMCE and data just "slotted in".
17:16	raek	AFAIC, hiccup only does clojure data to html conversion
17:16	tonyl	konr: I am guessing you are refering to namespace foo not package "foo"
17:16	konr	tonyl: yes, indeed
17:16	tonyl	konr: usually if it is a global var it goes in the core
17:16	tonyl	like foo.core/bar
17:17	tonyl	or just the base like foo/bar
17:17	raek	konr: namespaces cannot have cyclic dependencies
17:17	ossareh	raek: in that example deftemplate is reading that html, throwing it into a var and then what are the forms after that? Perhaps replacing the content of those entries with the output of the form?
17:18	raek	yes, those forms describe what html elements to match and what to do with them
17:18	raek	you use CSS-style selectors
17:19	ossareh	right
17:19	ossareh	I do like that. A lot in fact.
17:19	ossareh	deftemplate seems to be reading from the filesystem - that could come from anywhere I hope?
17:19	raek	the right hand side is a function that takes the subtree and returns a (potentially) altered one
17:20	bendlas	ossareh: enlive is especially worth it, when you have html from other sources. generating new html based on existing html is the thing enlive does.
17:20	raek	yes, it's just one of the convenience macros, iirc
17:20	raek	LauJensen has written about enlive in his blog
17:20	LauJensen	many times :)
17:21	ossareh	raek, bendlas: awesome - thanks. I'm going to spend some quality time with it.
17:22	pdk	,(apply +)
17:22	clojurebot	java.lang.IllegalArgumentException: Wrong number of args (1) passed to: core$apply
17:22	pdk	hmm
17:22	brehaut	,(apply + [])
17:22	clojurebot	0
17:23	brehaut	apply needs a fn and a seq minimum
17:24	pdk	i was looking for an equivalent to funcall but hey
17:24	pdk	guess wrapping the function i generate in extra parens will work
17:24	brehaut	,(+)
17:24	clojurebot	0
17:24	pdk	,(Math/pow 2 13)
17:24	clojurebot	8192.0
17:24	raek	yes, the first thing in the list just has to evaluate to something that implements IFn. no funcall is needed.
17:25	bendlas	,(.invoke + 1 2)
17:25	raek	like in Scheme
17:25	clojurebot	3
17:25	bendlas	,(.invoke +)
17:25	clojurebot	0
17:26	raek	,(* (+ () ()) (+ () () ()) (+ () () () () () () ()))
17:26	clojurebot	42
17:27	brehaut	haha awesome
17:27	bendlas	wow
17:27	bendlas	just wow
17:27	brehaut	real simple
17:27	brehaut	the identity of multiplication is 1
17:28	ossareh	LauJensen: so bestinclass is built with the manly sounding CMS you eluded to close to new year?
17:28	brehaut	if you think of * as reducing its arguments, you'd have (reduce binary_mult 1 seq)
17:29	brehaut	so you can use (*) as weird-hand for 1, and the rest is basic math
17:34	brehaut	huh
17:35	brehaut	the implementation of + is actually really close to that, jsut with some explicity arities and an inline
17:35	brehaut	s/+/*/
17:36	tonyl	but + with zero args returns 0
17:36	bendlas	yeah, rich knows his algebra
17:37	brehaut	tonyl: yes, 0 is the identity for addition
17:37	tonyl	yup
17:38	mattmitchell	how can i test a files existence in clojure?
17:39	bendlas	see java.io.File
17:40	bendlas	with convenience fns from clojure.java.io
17:40	mattmitchell	bendlas: excellent thanks!
17:43	bendlas	clojurebot: do you do javadoc?
17:43	clojurebot	Pardon?
17:43	bendlas	,(javadoc java.io.File)
17:43	clojurebot	java.lang.Exception: Unable to resolve symbol: javadoc in this context
17:43	bendlas	clojurebot: google java.io.File
17:43	clojurebot	First, out of 535000 results is:
17:43	clojurebot	File (Java 2 Platform SE v1.4.2)
17:43	clojurebot	http://download.oracle.com/javase/1.4.2/docs/api/java/io/File.html
17:50	zanes_	dnolen: Why?
17:50	Raynes	mattmitchell: (.exists (java.io.File. "/home/matt/somefile"))
17:51	mattmitchell	Raynes: That's exactly what I'm looking for thanks!
17:51	raek	(also consider using clojure.java.io/file for joining path segments in a platform independent way)
17:54	dnolen	zanes: seems like a fantastic way to annotate code w/o cluttering it up with comments while under heavy development. also seems like a great way to annotate other people's code while you're figuring it out.
17:54	raek	interesting
17:55	raek	especially the second usage
17:55	mattmitchell	I have a condition setup, where if the test is true, I'd like to print a message and then call (System/exit 0)
17:55	mattmitchell	but the (System/exit 0) part is never executed.
17:55	mattmitchell	only the printing of the message
17:56	semperos	you capturing the actions in a do block?
17:56	tonyl	(when true (println "message) (System/exit 0))
17:56	tonyl	&(when true (println "message) (System/exit 0))
17:56	sexpbot	java.lang.Exception: EOF while reading string
17:56	mattmitchell	no
17:56	tonyl	&(when true (println "message") (System/exit 0))
17:56	sexpbot	java.security.AccessControlException: access denied (java.lang.RuntimePermission exitVM.0)
17:56	semperos	if you're using an if expression, the second form is the 'else' statement
17:57	semperos	i.e. the second form after the condition
17:57	mattmitchell	semperos: ok i see
17:57	tonyl	&(if true (do (prn "message") (System/exit 0)))
17:57	sexpbot	java.security.AccessControlException: access denied (java.lang.RuntimePermission exitVM.0)
17:57	semperos	yep
17:57	mattmitchell	tonyl semperos thanks!
17:58	brehaut	mattmitchell: do look at the flow control section of http://clojuredocs.org/quickref/Clojure%20Core
17:58	brehaut	there are a lot of short hand forms that will make your life easier
18:05	mattmitchell	brehaut: awesome thanks
18:06	mattmitchell	brehaut: very helpful indeed
18:06	pdk	,(Math/pow 2 20)
18:06	clojurebot	1048576.0
18:11	jweiss_	i know this is only tangentially related to clojure, but somebody here probably knows the answer: when i run 'lein deps', the deps that are SNAPSHOT versions do not get updated from the remote repository, it just copies the existing snapshot from the local repo. is there a way to change this behavior?
18:13	dakrone	jweiss_: lein will check once a day for an update to snapshots
18:13	jweiss_	dakrone: is there a way to get it to check every time?
18:13	technomancy	if you need more than that you should probably use checkout dependencies
18:14	dakrone	jweiss_: you can remove it from your ~/.m2 dir
18:14	technomancy	jweiss_: I attempted to implement that, but the Maven API call to force it to check every time seemed to be a no-op.
18:14	technomancy	could be I was doing something wrong, but checkout deps are generally a better solution for most cases
18:14	jweiss_	hm. maybe i can run mvn -U ?
18:14	technomancy	that works too
18:14	jweiss_	i suppose rm -rf will work just as well
18:15	jweiss_	ok thanks
18:23	ossareh	so, with enlive, when I deftemplate name file - what are the lookup semantics for where that file is? I'm getting an NPE with all sorts of different path attempts
18:25	brehaut	ossareh: your classpath i believe
18:26	brehaut	oh, maybe source path
18:28	ossareh	brehaut: cool - looks like you're right (just stepping through the enlive code now) clojure.lang.RT/baseLoader
18:28	brehaut	ossareh: ok cool, i'm making pigs out of the sausage of my own sites implementation
18:29	brehaut	in lein i have :source-path defined as "src", and inside src i have templates/….html and my deftemplates are qualified with respect to src
18:31	ossareh	aha - when you package up with lein war is everything kept sane?
18:31	brehaut	haha no i use ductape and string :P
18:31	brehaut	im sure most sane people will chastise me for my deployment strategy
18:32	ossareh	brehaut: I think we're in the same camp - so not chastening from me =)
18:32	brehaut	i git pull on the live site, lein uberjar, kill the old process, java -server -jar ... & disown
18:32	ossareh	heh - pretty much similar except I scp the uberjar from my local machine
18:33	brehaut	ah right. i have all my media resources served by nginx on a subdomain
18:33	brehaut	andi just point it at the git repo
18:33	technomancy	as long as it's not production code...
18:33	brehaut	technomancy: what do you mean by production ?
18:33	technomancy	live users; your job on the line, etc
18:33	brehaut	oh right
18:33	brehaut	nope
18:34	technomancy	ok, carry on =)
18:34	brehaut	if it was my job to run my blog, i'd have a lot more time to get the deployment sorted, and i'd earn a lot less :P
18:35	ossareh	technomancy: if I specify resources-path is that imported into a jar / uberjar ?
18:35	technomancy	ossareh: yeah. you don't have to specify it; if you have a resources/ directory, then it will be included in the jar
18:35	ossareh	oh, OK.
18:35	ossareh	thanks :)
18:36	technomancy	sure
18:37	brehaut	ossareh: theres a bunch of different lein plugins for different automated deployment tools/strategies these days.
18:37	brehaut	realistically i should put aside half a day to learn one
18:38	ossareh	brehaut: blog about it :)
18:38	brehaut	haha
18:38	brehaut	no doubt
18:39	__name__	brehaut: would you mind giving me the url to your blog?
18:39	brehaut	http://brehaut.net/
18:39	__name__	does not load for me
18:39	brehaut	comments about the colors can be redirected to /dev/null/
18:40	brehaut	;)
18:40	__name__	brehaut: i do not get any meaningful page with chromium
18:40	__name__	echo "my eyes!" > /dev/null
18:41	brehaut	hmm. it looks fine in both safari and webkit for me
18:41	brehaut	and http://www.downforeveryoneorjustme.com/brehaut.net says its up
18:41	__name__	brehaut: i do get /something/.
18:42	ossareh	;)
18:42	brehaut	ossareh: OMG monads WTF is reasonable
18:42	ossareh	technomancy & brehaut - thanks - I now have my deftemplates working =)
18:43	brehaut	no worries
19:04	joshua__	I'm using duck-streams write-lines method to write a file on a Linux OS which is than read by a C++ program on a Window's OS.
19:05	joshua__	The program doesn't run as expected. When you open up the file in Notepad there is an invisible char directly before line breaks.
19:05	joshua__	I get that this is some sort of new lines are different based on operating system thing, but how do I fix this?
19:06	brehaut	the basic newline differences would go the other way
19:06	brehaut	unix uses \n and windows uses \n\r (i think that way round)
19:07	brehaut	start with opening the file in a hex editor or something
19:07	joshua__	So what is my actual problem?
19:07	brehaut	you need to look at the files byte contents, rather than a text editors interpretation
19:07	brehaut	text file encoding is a broad and subtle field for erros
19:08	joshua__	brehaut, how do I view the byte contents?
19:08	brehaut	hex editor or something
19:09	brehaut	joshua__: http://ubuntuforums.org/showpost.php?p=277111&postcount=2 for linux
19:09	joshua__	thanks
19:09	brehaut	a good hex editor wil lshow you line number, bytes as hex pairs, then the ascii output
19:10	brehaut	http://stackoverflow.com/questions/10426/what-is-a-good-windows-hex-editor-viewer for windows
19:11	brehaut	also, check the very first two bytes in the file
19:12	brehaut	thats where a byte order mark appears if it has been included (common for UCS and some UTF encodings)
19:12	brehaut	http://en.wikipedia.org/wiki/Byte_order_mark
19:20	joshua__	brehaut, thanks for pointing me in the right direction. If I dig into hex to fix a bug I'm allowed to call myself Mel, right?
19:21	brehaut	only if you rewrite your duckstream to take into account of drum timings to control program flow.
19:21	brehaut	using a hex editor to fix a bug is just unlucky
19:40	ossareh	OK. enlive rocks.
19:41	brehaut	totally eh
19:42	brehaut	one really nice property is that you can work on the design by just editing the template without any other webserver or anything in between.
21:00	pdk	(doc bindings)
21:00	clojurebot	Excuse me?
21:00	pdk	(doc binding)
21:00	clojurebot	"([bindings & body]); binding => var-symbol init-expr Creates new bindings for the (already-existing) vars, with the supplied initial values, executes the exprs in an implicit do, then re-establishes the bindings that existed before. The new bindings are made in parallel (unlike let); all init-exprs are evaluated before the vars are bound to their new values."
21:10	pdk	,(let [myfn (fn [[arg]] arg)] (myfn))
21:10	clojurebot	java.lang.IllegalArgumentException: Wrong number of args (0) passed to: sandbox$eval930$myfn
21:10	pdk	,(let [myfn (fn [& [arg]] arg)] (myfn))
21:10	clojurebot	nil
21:13	tonyl	&((fn [& [x y & more]] [x y more]) 1 2 3 4 5 6)
21:13	sexpbot	⟹ [1 2 (3 4 5 6)]
21:16	brehaut	$source destructure
21:16	sexpbot	destructure is http://is.gd/ub7pAw
21:17	brehaut	well, thats a bunch longer than i was expecting
21:18	tonyl	it's magically longer
21:20	hiredman	maniacally
21:21	brehaut	pretty sure a lot of tea was involved
21:23	tonyl	aaah, i need to extend my nested closure tolerance
21:26	brehaut	tonyl: have you played about with it? it helps to know what its doing
21:27	pdk	i am officially concurrent programming wizard
21:27	brehaut	offically?
21:27	pdk	apparently trying to optimize an AI tree search algo by making it spawn a new thread for each branch in the tree unchecked is not wise
21:27	pdk	MUCH TO MY CHAGRIN
21:28	brehaut	tonyl: its rewriting a destructuring bind into a bunch of calls to nth etc
21:29	brehaut	,(destructure ['a 1])
21:29	clojurebot	[a 1]
21:29	tonyl	brehaut: I figure it creates new symbols based on the destructuring norms that are passed
21:29	brehaut	,(destructure ['[a b] [1 2]])
21:29	clojurebot	[vec__948 [1 2] a (clojure.core/nth vec__948 0 nil) b (clojure.core/nth vec__948 1 nil)]
21:30	tonyl	yup
21:30	hiredman	that is the tamest sort of destructuring
21:30	brehaut	the output is surprisingly simple compared to the work that goes into producing it
21:31	hiredman	,(destructure ['{{[x y z] :b} :a} {:a {:b [1 2 3}}])
21:31	clojurebot	Unmatched delimiter: }
21:31	hiredman	,(destructure ['{{[x y z] :b} :a} {:a {:b [1 2 3]}}])
21:31	clojurebot	[map__951 {:a {:b [1 2 3]}} map__951 (if (clojure.core/seq? map__951) (clojure.core/apply clojure.core/hash-map map__951) map__951) map__952 (clojure.core/get map__951 :a) map__952 (if (clojure.core/...
21:31	pdk	} truly is the best in the land clojurebot
21:33	hiredman	[{:keys [bytes} {:keys [x y z] :as args} :foo :as everything}] ;;ran into something like this in production code today
21:33	danlarkin	a travesty
21:34	brehaut	oh my
21:34	hiredman	danlarkin: right, well joe and I took some steps
21:34	tonyl	[bytes doesn't look like a valid symbol
21:34	danlarkin	good, abominations like that should not stand
21:35	hiredman	the [bytes} should be [bytes]
21:36	hiredman	then he complained about something like (let [a … b …] (.delete a) b)
21:36	tonyl	haha
21:36	tonyl	sounds like somebody had code amnesia
21:37	hiredman	so I suggested (let [a …] ((fn [b] (.delete a) b) …)) so he yelled at me for a bit
21:37	danlarkin	oh god
21:37	hiredman	yeah, just like that
21:38	tonyl	wow, did you show him his own code?
21:38	hiredman	oh the destructuring thing was code I had written too
21:42	brehaut	the frightening thing about those snippets is that they are pretty decent in the context of the wider industry
21:43	hiredman	what is the difference between (let [a 1] a) and ((fn [a] a) 1) ?
21:43	no_mind	what is the difference between defn and defn- ?
21:44	hiredman	let can bind primitives to names and fn creates a recur point
21:44	danlarkin	no_mind: defn is ok to use and defn- is wrong
21:44	brehaut	no_mind: defn creates a ns private var
21:44	no_mind	danlarkin: amny clojure libs are using defn-
21:44	brehaut	no_mind: many projects are using C++; doesnt make it right
21:44	danlarkin	if they jumped off a bridge would you too?
21:45	danlarkin	BAM! Mothered
21:45	hiredman	danlarkin is something of a luddite, he complains if I use letfn
21:45	no_mind	well i want to know what is defn- and why not to use it ?
21:45	brehaut	hiredman: its not his fault, he came from python
21:45	hiredman	no_mind: have you read the doc string?
21:45	no_mind	no
21:45	tonyl	&(doc defn-)
21:45	sexpbot	⟹ "Macro ([name & decls]); same as defn, yielding non-public def"
21:45	brehaut	no_mind: you dont want to use because someone will come along later and curse you that the fn they need is private and now tey need to copy and waste
21:46	hiredman	it is infact very easy to get access to a private var
21:46	brehaut	hiredman: oh real? that would be handy
22:04	technomancy	private is more about reducing the surface area of your API than actually hiding anything
22:04	brehaut	interesting
22:04	technomancy	why should foo.mylib.<TAB> complete to a bunch of internal junk you don't ever need to call?
22:04	brehaut	sure that makes sense
22:05	brehaut	time for the stupid question: emacs clojure mode supports that?
22:05	technomancy	no, you need swank
22:05	brehaut	ah
22:06	brehaut	pass on that for a while then
22:07	technomancy	suit yrself
22:08	brehaut	yup :) while it would be handy i am still pretty bad at emacs
22:12	Raynes	technomancy: Have you ever done any gnus stuff?
22:17	zippy314	Hi, I'm wanting to regex replace on strings and a google search yields re-gsub in clojure-contrib.str-utils. and also re-replace in clojure.contrib.strings What's the difference and which should I use?
22:17	hiredman	you should check the .replaceAll method on String and see if it doesn't do what you want
22:18	zippy314	I also what lower-casing and capitalization. Should I just use java methods?
22:18	hiredman	yes
22:20	zippy314	,(.replaceAll "fish bonk" #"\w")
22:20	clojurebot	java.lang.IllegalArgumentException: No matching method found: replaceAll for class java.lang.String
22:21	zippy314	?
22:21	hiredman	zippy314: when I said check I meant "read the javadoc"
22:21	zippy314	it's there.
22:21	hiredman	read the javadoc
22:21	zippy314	http://download.oracle.com/javase/1.4.2/docs/api/java/lang/String.html#replaceAll(java.lang.String,%20java.lang.String)
22:21	zippy314	oh (blush)
22:21	zippy314	,(.replaceAll "fish bonk" #"\w" _)
22:21	clojurebot	java.lang.Exception: Unable to resolve symbol: _ in this context
22:22	zippy314	,(.replaceAll "fish bonk" #"\w" "_")
22:22	clojurebot	java.lang.ClassCastException: java.util.regex.Pattern cannot be cast to java.lang.String
22:22	hiredman	you still haven't read the doc
22:22	hiredman	infact you haven't even read the link you just posted
22:22	zippy314	yah. sorry.
22:24	zippy314	ok. but please explain why I should be using this instead of the contrib libs.
22:24	brehaut	zippy314: its right there in the javadoc
22:25	hiredman	because dependence on contrib, given the dubious quality of lots of contrib is bad
22:25	zippy314	(.replaceAll "fish bonk" "\\W" "_")
22:25	zippy314	,(.replaceAll "fish bonk" "\\W" "_")
22:25	clojurebot	"fish_bonk"
22:26	zippy314	Hmmm. I'm not so keen about regexs encoded as strings. It's a pain to have to escape the escapes and so on..
22:27	brehaut	hiredman: in this instance, the fns in question have moved from contrib to clojure itself havent they?
22:27	jimbostrudel	you could store them in a file and read them in
22:27	brehaut	http://clojure.github.com/clojure/clojure.string-api.html
22:28	hiredman	brehaut: shrug
22:44	zippy314	brehaut: from that url it says to use: (:require '[clojure.string :as str]) But when I do that, I get "java.lang.Exception: lib names inside prefix lists must not contain periods" any idea what gives?
22:45	brehaut	is that part of your ns definition?
22:45	tonyl	(:require doesn't need the apostrophe
22:46	brehaut	zippy314: ie do you have (ns my.ns (:require …))
22:46	technomancy	Raynes: it's been a while since I had a really tuned gnus
22:46	Raynes	technomancy: It was fairly easy getting it to fetch my mail. The problem is that there are no mail groups. INBOX doesn't exist.
22:47	Raynes	It's bizarre.
22:47	Raynes	<3 thunderbird
22:47	scottj	you mean you can't subscribe to inbox?
22:49	Raynes	scottj: Yes. It claims that it doesn't exist.
22:50	scottj	but you can see the maildir/whatever file(s)?
22:51	Raynes	The only thing it's actually showing me is the default help group.
22:51	zippy314	brehaut: yes I was. The trick was what tonyl said.
22:51	brehaut	zippy314: ok cool.
22:51	zippy314	tx
22:52	scottj	Raynes: have you consulted the manual?
22:52	Raynes	Yes, and at least 4 different step-by-step tutorials. It doesn't look like anybody else has ever encountered this problem.
22:52	Raynes	It isn't a huge deal. I'll ask in #emacs eventually.
22:53	scottj	what select method are you using?
22:53	scottj	anyway, email clients that only show one message at a time are lame.
22:54	Raynes	nnml