#clojure – 2010-10-17

2010-10-17

00:17	LauJensen	Good morning guys
00:37	amalloy	hi LauJensen
04:05	notsonerdysunny	why should " (def (symbol "hello") 10) " work .. does anybody have an insight into this?
04:05	notsonerdysunny	why shouldn;t " (def (symbol "hello") 10) " work .. does anybody have an insight into this?
04:06	amalloy	notsonerdysunny: def requires a symbol literal
04:07	amalloy	,[(symbol "hello") 'hello]
04:07	clojurebot	[hello hello]
04:08	amalloy	(symbol "hello") is the same as 'hello, and it seems pretty reasonable that (def 'hello 10) shouldn't work, right?
04:10	LauJensen	user> `(def (symbol "hello") 5)
04:10	LauJensen	(def (clojure.core/symbol "hello") 5)
04:10	LauJensen	user> `(def 'hello 5)
04:10	LauJensen	(def (quote user/hello) 5)
04:10	LauJensen
04:11	amalloy	yes?
04:11	LauJensen	In either case its clear they wont work, since neither are a literal symbol
04:11	notsonerdysunny	amalloy: hmm...
04:11	notsonerdysunny	so we cannot programmatically generate the symbols..
04:12	amalloy	you can, but you have to do it in a macro
04:13	notsonerdysunny	amalloy: the other day you had suggested that I use apply-macro to make sure that the arguments of a macro get evaluatated before the macro expansion .. but I couldn't figure out how to use it
04:13	notsonerdysunny	have you used it before?
04:15	amalloy	i wouldn't go so far as to say i suggested it...i think you are trying to do things at compile time that you can't do until runtime, and apply-macro lets you get close to that sometimes
04:15	notsonerdysunny	ok sorry for misrepresenting you.. :)
04:15	amalloy	if you really want to create code based on the execution of code, why not just have your clojure program generate a new .clj file
04:15	notsonerdysunny	but I was not able to figure out how to use it in any case ..
04:15	amalloy	and ship that to the user
04:17	notsonerdysunny	:) good point .. but was hoping I needn't have to do that ..
04:18	notsonerdysunny	I think we are trying to make an argument like people using languages who don't have macros would .. :)
04:18	amalloy	if you want the user to be able to run the same program that you do, unmodified, then you have to ship them mathematica. i don't know how it can possibly work otherwise
04:18	notsonerdysunny	I would do to Ahead Of Time compilation
04:19	notsonerdysunny	so all the macros would be expanded in the resultant code ..
04:19	notsonerdysunny	well atleast that was the theory based on which I was working ...
04:20	amalloy	right, i always forget AOT is an option
04:26	amalloy	okay. here is an idea that i think i have tried a few times but i don't think i've really communicated it to you. one sec while i put together a gist
04:27	amalloy	notsonerdysunny: http://gist.github.com/630662
04:29	amalloy	here ask-mathematica is a function, which helps solve-pde generate code
04:31	notsonerdysunny	thanks amalloy .. Let me try to use your idea and come up with something..
04:31	notsonerdysunny	:)
04:32	amalloy	the general point here is you very rarely want to define a macro that uses another macro you define - the second level and below usually want to be functions
04:34	amalloy	fwiw, (solve-pde x y) really does macroexpand to (+ y (* x 10)) there - the reference to ask-mathematica has vanished
04:50	notsonerdysunny	:) ... but as a side-note.. It would be nice if we had a set of macros which would wrap the special forms and be able to handle non-literal values.. be it macro or just regular code .. just wishfull thinking .. I don't know enough even to evaluate the feasibility of such macros .. :)
04:57	amalloy	notsonerdysunny: well, for (def) you don't need such a macro
04:59	amalloy	(intern (symbol "test") 'size 10)
04:59	amalloy	can't do that in clojurebot obviously, but it should work at your repl
05:00	amalloy	anyway, i'm off to bed. good luck, i hope to see how it all turns out
05:36	_ulises	morning all
07:22	serabe	hi
07:23	raek	hello
08:02	serabe	just released rinzelight, an image library for clojure, http://github.com/Serabe/rinzelight
09:58	shanmu	Hi, What options would I have if I want to run a long running process/daemon in clojure (which has to do some periodic activity)?
09:59	shanmu	my thought is to have a compojure app, to which I do wget from cron...
10:00	shanmu	or another option is to write a quartz job but I am not sure how to integrate in clojure...
10:02	shanmu	or yet another option would be invoke the cljoure script/compiled class direct from shell script/cron but pay jvm startup cost everytime
10:13	LauJensen	shanmu: jvm startup cost is typically < 1 second with the -client option. It would be trivial to write a program which ran a future every x seconds though. Its a tradeoff between startup time and memory consumption, weigh it thinking about the frequency of the activity
10:15	shanmu	LauJensen: thanks for the suggestion - I had missed future... (i am still a n00b :))
10:28	Bahman	Hi all!
10:53	FUser	,(range 1000)
10:53	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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133
10:53	FUser	:)
11:16	FUser	,(counted? (seq [2 33]))
11:16	clojurebot	false
11:16	FUser	,(counted? (seq '(2 33)))
11:16	clojurebot	true
11:17	FUser	wat?
11:17	FUser	makes no sense
11:17	jarpiain	,(doc counted?)
11:17	clojurebot	"([coll]); Returns true if coll implements count in constant time"
11:17	Chousuke	FUser: they're different types
11:18	FUser	they are both sequences
11:18	Chousuke	,(let [x '(2 3)] (identical? x (seq x)))
11:18	clojurebot	true
11:18	Chousuke	that's the reason
11:18	FUser	look
11:18	FUser	,(counted? [2 32])
11:18	clojurebot	true
11:18	FUser	,(counted? '(23 2))
11:18	clojurebot	true
11:19	MayDaniel_	,[(type (seq [1 2])) (type (seq '(1 2))]
11:19	clojurebot	Unmatched delimiter: ]
11:19	FUser	,(counted? (seq [2 23]))
11:19	clojurebot	false
11:19	raek	seqs are not guaranteed to be counted
11:19	FUser	these are all expected
11:19	FUser	yeah
11:19	raek	but PersistentLists are their own seqs
11:19	FUser	but why are seqences made out lists counted
11:19	raek	so (seq a-list) happens to be counted
11:19	FUser	also maps
11:19	tensorpudding	where's a good place to get a tutorial to clojure?
11:19	Chousuke	FUser: because lists don't need to be made into sequences.
11:20	Chousuke	FUser: they are sequences already
11:20	raek	(let [l (list 1 2 3)] (identical? l (seq l)))
11:20	raek	,(let [l (list 1 2 3)] (identical? l (seq l)))
11:20	clojurebot	true
11:20	FUser	,(counted? (seq {3 22}))
11:20	clojurebot	true
11:20	Chousuke	I suppose it might be an oversight, though.
11:21	FUser	whats the diff between seq function and sequence function?
11:22	FUser	,(seq nil)
11:22	clojurebot	nil
11:22	FUser	,(sequence nil)
11:22	clojurebot	()
11:22	FUser	oh
11:23	raek	tensorpudding: this covers the basics http://en.wikibooks.org/wiki/Learning_Clojure
11:24	FUser	sequence is also faster it seems
11:25	FUser	(time (do (sequence (range 1000000000)) nil))
11:25	raek	I have seen 'sequence' before...
11:25	FUser	,(time (do (sequence (range 100000000)) nil))
11:25	clojurebot	"Elapsed time: 0.135 msecs"
11:26	FUser	,(time (do (seq (range 100000000)) nil))
11:26	clojurebot	"Elapsed time: 0.212 msecs"
11:26	raek	seq is the one that is usually used to turn a collection into a sequence
11:27	_ulises	,(doc sequence)
11:27	clojurebot	"([coll]); Coerces coll to a (possibly empty) sequence, if it is not already one. Will not force a lazy seq. (sequence nil) yields ()"
11:28	tensorpudding	what does clojure call lambda?
11:28	Kruppe	fn
11:28	Kruppe	(fn [] (print "hi"))
11:28	tensorpudding	okay
11:29	tensorpudding	clojure looks pretty good for being built on java, i suppose
11:30	Kruppe	tensorpudding: It's pretty nice, id say its more built on the jvm than built on java though.
11:32	raek	there are plans to reimplement the compiler (which is now written in Java) in Clojure
11:32	FUser	:D
11:33	FUser	the only problem is that clojure uses a shitton of reflection and allocated billions of objects per second
11:34	raek	reflection is only used of java methods calls
11:35	raek	and can be avoided completely, if the type is known (e.g. through type hints)
11:35	Raynes	tensorpudding: Ooh, you're the second prominent #haskell user in here in the last week.
11:35	tensorpudding	heh
11:35	Raynes	kmc is here too. <3
11:36	FUser	raek: so if I am using clojure Fns, I am not using reflection?
11:36	raek	FUser: exactly.
11:37	raek	clojure fns are always take Objects and return Objects
11:37	FUser	allocations are still a problem though
11:37	raek	yes
11:37	FUser	they did some testing
11:37	Raynes	raek: All your clojure fns are belong to us.
11:37	raek	but the HotSpot JVM is pretty good at it
11:38	FUser	compared ArrayList and LinkedList in java, inserting in the middle of the list, which is O(n) for array list and O(1) for linked list
11:38	raek	but having immutable data structures kinda requires much allocation
11:38	FUser	array list still faster unless the list size is over a million
11:39	FUser	simply the allocations for the linked list objects
11:39	FUser	are too much
11:39	raek	I guess that the mutable data structures are simpler and faster, but they still have the problems Clojure is trying to solve
11:39	FUser	and that's java mutable linked list, I can only imagine how badly the clojures quasiconstant time inserts must be
11:40	Chousuke	Don't imagine, measure :)
11:40	FUser	I'm just sayin'...O(n) notation doesn't say shit :P
11:40	raek	neither PersistentVector nor PersistentList is the correct data structure for doing fast inserts to the middle
11:40	raek	I have heard that chouser is working on some data structure that has good performance for that
11:41	Chousuke	Finger trees
11:41	wakeupsticky	best clojure tutorial for the non-java programmer (i'm familiar with lisp and haskell)?
11:42	FUser	transient vectors are probably fast
11:43	Chousuke	They'll still be slow for inserts in the middle
11:43	FUser	if they are actually java ArrayList then they are fast
11:43	raek	wakeupsticky: these were what I got started with. YMMV. http://clojure.blip.tv/posts?view=archive&nsfw=dc
11:43	Chousuke	they aren't.
11:43	Chousuke	they can't be. :/
11:44	Chousuke	because transient->non-transient and vice versa must be a constant time operation
11:44	FUser	ok then I'll just use (.java.util.ArrayList) :P
11:45	Chousuke	if you're really doing lots of inserts, maybe :P
11:45	FUser	fastest access too
11:45	Chousuke	but realise that you're giving up safety and ease of use
11:45	FUser	I know
11:46	FUser	I can't cope with having 300 times slower apps
11:46	Chousuke	I'm not sure about that.
11:47	Chousuke	For a whole bunch of apps Clojure's vectors won't be a bottleneck
11:47	Chousuke	Don't give up on them until you encounter a real performance problem.
11:48	Chousuke	Fast programs are nice, but working ones are nicer :P
11:48	FUser	clojure has a bunch of weird performance issues
11:48	Chousuke	Like?
11:49	FUser	(if (> its 0)
11:49	FUser
11:49	FUser	compared to
11:49	Chousuke	I can't think of anything that's particularly weird. There is room for improvement though.
11:49	FUser	(if (> its (int 0)))
11:49	FUser	guess the speed difference
11:49	Chousuke	FUser: in the former case, there is boxing
11:49	FUser	not only that
11:49	Chousuke	so of course it's slower
11:50	FUser	but there's also lookup in > function
11:50	FUser	for type
11:50	Chousuke	yeah, because it's polymorphic :p
11:50	FUser	second version runs 9 times faster
11:50	Chousuke	(numerical performance is kind of the focus of 1.3 btw)
11:51	Chousuke	So expect things to improve
11:51	FUser	that kind of performance differences for such a simple change in code is unacceptable
11:51	Chousuke	No they are not
11:51	Chousuke	it's not a simple change,
11:51	Chousuke	it does a lot.
11:51	FUser	I'll spend time saved in development doing profiling afterwards
11:51	FUser	well technically yes
11:52	FUser	but I can't see why I'd have to cast a constant 0, why can't compiler do that for me
11:52	Chousuke	Because the compiler is not very good yet.
11:52	Kruppe	FUser: its a new language, 1.3 will improve that particular problem
11:52	FUser	:)
11:52	Chousuke	It doesn't try to be, either; there are bigger things to do first.
11:52	Kruppe	FUser: for now its not perfect, thats for sure.
11:54	FUser	sorry I'm mad, but after doing SQL optimizations, where there's a bunch of hidden rules how things get evaluated and slightly changed but semantically same sentences produce vastly different performance I'm really disgusted by these kinds of things
11:55	FUser	it all becomes just a whole bunch of arcana
11:55	FUser	and code gets ugly fast
11:56	Kruppe	FUser: I sympathize, I wrote a fairly complicated simulation in clojure and it was a real pain to work out all the performance kinks.
11:56	Kruppe	FUser: the simulation was easy to write though.
11:59	FUser	yeah but you see what I mean....you write a simple sql query, optimizer screws it up in ways you know nothing of, then you scour the forums for ages to find some kind of hack (adding dumb shit like "or false" to queries, that makes the optimizer do things differently)
12:00	FUser	and that simple query becomes ugly unreadable and ununderstandeable to any other programmer, unless they are in the know of these hacks
12:01	Kruppe	FUser: Its performance definitely isn't a non issue, if its not good enough right now you could always use something else.
12:02	Kruppe	FUser: Or write the bottlenecks in java since you're already on the jvm.
12:02	FUser	well, I am not using it professionally so it's not much of an issue
12:02	Kruppe	FUser: oh, well in that case. Readability isn't such a huuuuuuuge hit.
12:02	Chousuke	Or you could go and improve the compiler, and be everyone's hero :)
12:02	FUser	:D
12:02	Kruppe	FUser: if you don't need concurrency semantics you could always use common lisp.
12:03	Kruppe	FUser: I find it easier to write fast code in CL atm, but i suspect that will change soon enough.
12:03	FUser	CL?
12:03	Kruppe	FUser: common lisp
12:03	FUser	oh
12:03	FUser	never used lisp
12:03	FUser	mostly java
12:03	Chousuke	In Clojure's case the compiler is destined to be rewritten in Clojure someday though.
12:04	FUser	my main java issues are already solved by C#
12:04	FUser	so that would be my perfect language
12:04	Kruppe	Chousuke: yeah I have confidence that rich is very concerned with performance.
12:04	FUser	but I have job requirement to use java so I'm stuck with it
12:04	Chousuke	Yeah.
12:04	FUser	I'm taking up clojure purely for the learning experience
12:05	FUser	since it's my first dynamically typed functional language
12:06	FUser	I'm fresh out of ideas what to make in it
12:06	Kruppe	FUser: It's a general purpose language, could make anything you want in it.
12:06	FUser	that's my main problem, I was learning Android, J2Me, Clojure, but I have no inspiration about what to make in them so I end up not practicing enough
12:06	Kruppe	FUser: performance issues aside I suppose.
12:07	FUser	I know :)
12:07	Kruppe	FUser: could just pick an open source project and contribute
12:07	FUser	hm....
12:07	FUser	not a bad idea
12:07	Kruppe	FUser: dunno if there are any clojure open source projects you would want to contribute to. But you could look.
12:09	FUser	it's funny how I'm the only guy at the office who programs and reads about programming in his spare time
12:09	FUser	people call me an academic
12:11	Kruppe	FUser: those people are probably just lazy and are surprised at your initiative
12:11	Kruppe	FUser: probably burnt out on coding java
12:12	FUser	I'm amazed. They have no drive to better themselves at their chosen profession
12:12	FUser	even though they are terrible at java
12:12	FUser	I don't need to improve my java, I'm one of the best in the firm
12:13	Kruppe	FUser: working at large companies can ruin you desire for self improvement
12:13	FUser	I see shit daily that would bring tears to your eyes
12:13	FUser	40 employees
12:13	Kruppe	FUser: oh
12:13	FUser	every non-member function in the project is in a class named Utils
12:14	FUser	every constant in the project is in a class named Constants
12:14	FUser	shit like that
12:14	Kruppe	FUser: that sounds pretty tame, but this is also a little offtopic now
12:14	FUser	class structure is 1 level deep at most, there are no interfaces defined
12:14	FUser	ok
12:15	FUser	basically cobol programmers trying to do OO
12:15	FUser	OOP
12:15	Kruppe	FUser: it sounds like that structure was just decided by a manager
12:15	FUser	nope, by actual programmers with 10 year more experience than me
12:16	Kruppe	FUser: I guess their minds have been somewhat tainted by working in cobol all the time.
12:16	FUser	no kidding :D
12:16	Kruppe	FUser: teach em how it's done then, you then will really be the academic.
12:16	FUser	that's why I'm learning clojure, not to get completely trapped in Java mindset
12:17	Kruppe	FUser: if learning is your goal i recommend haskell as well. Lots of interesting programming language concepts are available in it to play around with.
12:17	Kruppe	FUser: don't get your heart set on writing haskell for money or anything though.
12:18	FUser	true
12:18	Kruppe	FUser: clojure is a nice intermediate step though, you get to play around with macros and FP.
12:18	FUser	yeah
12:18	Kruppe	FUser: not to mention really slick concurrency.
12:18	FUser	hell I'm hoping for C# someday
12:18	FUser	at least it has function pointers
12:19	Kruppe	FUser: not a huge fan, but I don't enjoy programming for windows so.
12:19	raek	there is Clojure CLR...
12:19	FUser	implementing interfaces for a simple filter function in java is maddening
12:19	FUser	C# delegates neatly cover a lot of this
12:20	FUser	raek I'm not pining for CLR
12:20	FUser	just having delegates (function pointers actually) and proper templates (C# ones not java ones) would solve a lot of my day to day work problems with java
12:21	Kruppe	FUser: I don't doubt it.
12:21	FUser	clojure has all of it also, of course
12:21	FUser	but it's less commercially popular, so I'm not gonna get a job with it
12:21	Kruppe	FUser: Hey boss, can I use this java library called clojure?
12:22	FUser	hehe, you know
12:22	FUser	I thought about it :D
12:22	Kruppe	FUser: actually that would never fly.
12:22	Kruppe	FUser: and it would be pretty irrisponsible of you to decieve your boss like that.
12:22	FUser	I know I know :D
12:22	Kruppe	FUser: and your code would be unmaintainable by anyone else there sadly
12:22	Kruppe	FUser: if only it were that easy
12:22	FUser	yes that is the main problem
12:23	FUser	I leave company now they can't fix anything in that code,because nobody else knows clojure
12:23	Kruppe	FUser: I tried convincing my boss to use svn or git for our software projects
12:23	FUser	svn ? version control system?
12:23	Kruppe	FUser: even that was difficult
12:23	Kruppe	FUser: yeah, we weren't using any version control for the longest time
12:23	Kruppe	FUser: it was pretty depressing
12:23	FUser	wow
12:24	FUser	that's pretty bad
12:24	Kruppe	FUser: I don't work there anymore.
12:24	FUser	I'm trying to tell the boss that we need design documentation
12:25	Kruppe	FUser: really? That I find really unusual.
12:25	FUser	we are making none since the customers only pay for requirements documentation
12:25	Kruppe	FUser: i need to practically document my bathroom schedule
12:25	FUser	so bascially each programmer is free to design his parts of the program however he wants
12:25	FUser	you only have architectural restrictions
12:26	FUser	you have to use this DB, that XML Serializer library, that web service etc
12:26	FUser	so project parts vary wildly in quality depending on who from our firm is working on them
12:27	Kruppe	FUser: that's unfortunate.
12:27	FUser	only for our customers :DDDDDD
12:27	FUser	see this is how it kills you...I stopped giving a fuc
12:29	Kruppe	FUser: Yeah, that will happen if you're not careful.
12:29	FUser	I told my boss multiple times that better programmers should be doing class design for whole project and worse programmers should be implementing closely specified indivdual methods
12:29	FUser	can't be arsed to do it again
12:30	Kruppe	FUser: that sucks.
12:31	kryft	C# templates? Are those anything like C++ templates?
12:32	FUser	not sure
12:32	FUser	I just know that Java Generics are much like C# templates
12:32	Chousuke	Nothing like C++ templates then :/
12:32	FUser	but C# templates are proper and Java Generics are just stupid compiler hack with limited functionality due to havign to be backwards compatible
12:33	Kruppe	FUser: they work alright though. It could be worse.
12:33	FUser	I've run into their limitations several times before
12:33	Kruppe	FUser: language warts like that happen over time.
12:34	Kruppe	FUser: yeahhhh. Sometimes java just makes things difficult.
12:34	FUser	It could be better if Sun wasn't being so stubborn with their compatiblity
12:34	raek	I find it interesting that the C++ template system is Turing complete...
12:35	FUser	you can't run java 5 classes on java 1.4 anyway so why bother
12:35	Kruppe	FUser: its not stubbornness necessarily.
12:35	Kruppe	FUser: They don't want to scare anyone away.
12:35	Kruppe	FUser: in general they have done a real good job keeping java THE language for business.
12:36	FUser	here an example: public void <E> someMethod(E param) {
12:36	Kruppe	FUser: as a result it may not be the nicest language to work with but its still extremely popular.
12:36	FUser	E.class doesn't work, E.getClass() doesn't work
12:36	Kruppe	FUser: yeah thats one of those warts right there.
12:36	SergeyD	Hi, is there a way to get colorful Clojure cheat sheet? It was in the zip a year ago, but now that link is broken. (In the bottom of http://clojure.org/cheatsheet)
12:37	FUser	http://clojuredocs.org/quickref/Clojure%20Core
12:40	raek	SergeyD: I have the .tex file. Can you compile that?
12:40	SergeyD	FUser: thanks for another source
12:41	SergeyD	raek: yes, I'll figure out
12:41	FUser	what's a monad
12:42	raek	SergeyD: http://raek.se/cheat-sheet/
12:42	kryft	FUser: http://intensivesystems.net/tutorials/monads_101.html
12:42	kryft	FUser: (I don't really know myself, but I plan to read that at some point and find out. ;)
12:43	SergeyD	raek: thanks!
12:43	Kruppe	kryft: I've used haskell for a few personal projects, and I still can't really describe what monads are in words.
12:52	raek	compiled PDFs are now up too http://raek.se/cheat-sheet/
12:53	SergeyD	raek: thanks :)
13:03	_ulises	can anybody recommend a good library for performing 2d vector operations?
13:03	_ulises	I've sort of handcrafted my own but I'm sure that there are better options out there (google wasn't my friend :(
13:04	esj	_ulises: I think Incanter wraps Colt
13:04	_ulises	oh, right
13:04	esj	unsure how far they went
13:04	_ulises	I am looking for something a bit more lightweight though :)
13:05	_ulises	,(doc max)
13:05	clojurebot	"([x] [x y] [x y & more]); Returns the greatest of the nums."
13:05	Kruppe	Anyone else having trouble getting incanter and cake to work together?
13:05	esj	_ulises: oh, in that case the seq functions are your friend
13:06	_ulises	seq? how? kind of like (apply map + points) where points is ([x y] [x y]...) ?
13:06	raek	,(merge-with + {:x 1, :y 2} {:x 0, :y -3})
13:06	clojurebot	{:x 1, :y -1}
13:06	_ulises	ooh
13:06	_ulises	nice!
13:06	raek	not a seq function, but...
13:06	_ulises	no no, that is awesome
13:07	dnolen	_ulises: cantor is pretty good if you have simple needs, http://github.com/ztellman/cantor
13:07	_ulises	now I have to figure out how to use the same style for euclidean distance and the other fns...
13:08	esj	dnolen: thanks for the pointer - that looks pretty useful
13:08	_ulises	dnolen: that looks good! thanks
13:08	Kruppe	(require '(incanter core stats charts)) gives me a java.lang.ExceptionInInitializerError using cake
13:08	Kruppe
13:08	raek	,(map + [1 2] [0 -3])
13:08	clojurebot	(1 -1)
13:08	dnolen	_ulises: using maps or vectors is ridiculously slow if you're doing interactive grahis and the like
13:08	_ulises	dnolen: indeed!
13:08	dnolen	s/grahis/graphics
13:09	dnolen	_ulises: cantor has good perf, identical to using javax.vecmath in an immutable manner.
13:09	_ulises	awesome, will give it a whirl
13:10	_ulises	I can see how this is somewhat targeted to a physics engine
13:10	dnolen	_ulises: yeah it's integrated into penumbra, the clojure opengl library
13:10	_ulises	cool
14:10	kmc	wow a bunch of haskellers showed up :D
14:10	rlb	Is the best way to access a static nested java class to just import Foo$Bar?
14:13	kmc	Chousuke, finger trees in Clojure? sweet
14:14	raek	#clojure - #haskell cross-pollination?
14:14	kmc	seems so
14:14	kmc	FUser, Kruppe, i can try to explain monads if you like, or at least give some relevant meta-advice
14:14	tensorpudding	i had been meaning to try clojure for a little while
14:15	_ulises	kmc: I wouldn't mind that :)
14:15	tensorpudding	it's not a lot like haskell so far as i can tell though
14:15	kmc	me too
14:15	kmc	tensorpudding, i think it's philosophically close to Haskell, and technologically quite different
14:16	kmc	the barrier to me getting started was setting up all the java classpath apache maven 80 mb of code to do nothing jvm that tries to take down your whole system bs
14:17	tensorpudding	i had tried setting it up before
14:17	kmc	so i had a couple false starts due to that
14:17	kmc	but eventually i got lein set up ok
14:17	tensorpudding	i was inspired to try again while cleaning up the mess that was my .emacs
14:17	kmc	and the actual clojure part of clojure is nice
14:18	tensorpudding	clojure seems to have improved a lot on the packaging front
14:19	tensorpudding	for that matter, ELPA too
14:19	raek	Clojure and Haskell seems to have the preference for immutability in common.. (well, maybe preference is a bit of an understatement in the case of Haskell :) )
14:20	raek	I remember the time before lein...
14:20	tensorpudding	haskell likes immutability, but also code segregation through typing, which clojure doesn't appear to have
14:20	raek	yes, Clojure tries to do as much as possible without types
14:21	Kruppe	kmc: I wouldn't mind a nice description of monads. Thanks for the offer.
14:21	raek	data representation, dynamic dispatch and hierarchies doesn't have to do anything with types
14:22	_ulises	+1 on the monads explanation :)
14:22	kmc	raek, well Haskell has mutable data structures too
14:22	raek	inc
14:22	kmc	and in fact there's some very popular haskell libs that use mutation behind the scenes, but present a pure interface
14:22	_ulises	s/+1/inc/
14:22	kmc	writing such a thing is tricky
14:22	raek	reminds me of transients
14:22	kmc	yes
14:23	kmc	Haskell's ST monad is a lot like transients -- you get to use mutation locally, and expose the result as pure value
14:23	kmc	in Haskell the type system proves that your mutation doesn't leak out
14:23	kmc	and so there's no runtime overhead, compared to less safe approaches
14:23	raek	Haskell is very high on my to-learn list
14:24	raek	probably #1
14:24	tensorpudding	the only way to get performance on some things is mutation behind the scenes
14:24	Kruppe	raek: real world haskell did a good job for me. Pretty awesome book.
14:25	FUser	how's performance of Haskell
14:25	raek	ah, yes. I have borrowed that book from a friend. have barely started with it.
14:25	Kruppe	FUser: very good.
14:26	Kruppe	FUser: immutable data structures cause similar performance problems but you can use mutable data structures if necessary.
14:26	raek	ever since I was introduced to SML in the Programming Theory course I'm attending, I have been incresingly interested in learning Haskell
14:26	kmc	GHC is one of the smartest native-code compilers i know of
14:26	kmc	for any language
14:27	Chousuke	kmc: I don't know the status of Chouser's finger tree implementation though. It seemed sort of half-ready the last time I checked.
14:27	Kruppe	kmc: I miss macros when im in haskell though, template haskell feels clunky :(
14:27	kmc	yes
14:27	kmc	Template Haskell is incredibly clunky
14:27	kmc	that's one of my main complaints with Haskell these days, and it's one of my main reasons for learning Clojure
14:27	FUser	I miss types in clojure or even plain old C style structs
14:27	tensorpudding	TH is an dark, dingy failed attempt masquerading as a code extension.
14:27	kmc	metaprogramming is so powerful and so underappreciated, and Lisp is absolutely king there
14:28	Kruppe	FUser: clojure has structs
14:28	kmc	i wouldn't say TH is "failed" -- it is used in an essential role by a wide variety of useful projects. it's just a lot more pain than necessary
14:28	Chousuke	records :P
14:28	kmc	anyway, monads!
14:28	FUser	if they are implemented as maps then they aren't stucts even if they are named as such
14:28	Kruppe	kmc: rock on
14:28	kmc	my advice on learning about monads
14:29	Chousuke	FUser: records are the closest you get to C-style structs
14:29	Kruppe	FUser: they are quite fast as well
14:29	kmc	one thing to keep in mind is that "monad" is just the name of a generic API
14:29	kmc	public class List<T> implements Monad ...
14:29	kmc	and it's implemented by a lot of different unrelated things
14:29	Chousuke	They provide a maplike interface for convenience and are immutable.
14:30	kmc	people often look for a deep intuitive connection between all the monad instances
14:30	kmc	when they wouldn't expect this of just any generic API
14:30	Chousuke	(deftype is there if you need custom behaviour, but that's not recommended)
14:31	Chousuke	kmc: It helped me a lot to think of a monad as an "execution strategy" kind of thing :P
14:31	kmc	also, before you read (or write!) a monad tutorial, you should read: http://byorgey.wordpress.com/2009/01/12/abstraction-intuition-and-the-monad-tutorial-fallacy/
14:31	FUser	(Kruppe if they are maps then they aren't fast TBH
14:31	kmc	the Haskell community has moved pretty strongly against all these analogy-based tutorials
14:31	Kruppe	FUser: i don't think they are maps under the hood.
14:31	kmc	the common wisdom now is to get comfortable with a few concrete monad types
14:31	raek	FUser: (btw, transients in clojure, a way to temporarily use a data structure as a mutable one: http://clojure.org/transients)
14:31	kmc	you don't have to care as much about what the generic API means, if you focus on one type at a time
14:32	kmc	and also, to learn some of the less powerful (meaning more general) APIs first
14:32	kmc	functor and applicative functor
14:32	Kruppe	kmc: the one that really confuses me is ST, and the type errors i get when using it.
14:32	kmc	yeah, ST is pretty special
14:32	kmc	its specialness comes not from being a monad, but from the primitives that are provided for ST specifically
14:33	kmc	same for IO, STM, etc.
14:33	kmc	it's a common fallacy to learn IO first and then generalize it to all monads
14:33	kmc	when really IO is a strange atypical monad
14:33	kryft	kmc: Right, familiarize yourself with a few concrete examples and then let your brain do the generalization from there.
14:33	kmc	Kruppe, over in #haskell we'd be happy to help you with ST type errors
14:33	Kruppe	kmc: yeah that seems to always be the first monad to be described, maybe because its used all the time
14:34	kryft	kmc: I have a tendency to want to understand everything as generally as possible from the get-go, and that usually doesn't turn out so well. :)
14:34	Kruppe	kmc: thanks hehe, but its been a while since my last use of the ST monad. My projects right now are tied with C libs and python.
14:34	kmc	i like the approach of some materials (LYAH?) where they present IO early on, but give you the concrete explanation and avoid the m-word until much later
14:34	kmc	some of the people who "don't get monads" are actually not getting the idea of building IO-descriptions for later execution
14:35	kmc	which is an orthogonal idea
14:36	kryft	kmc: I take it you use both clojure and haskell?
14:36	kmc	yup
14:36	kmc	i've been using Haskell for about 5 years
14:36	kmc	just learning Clojure
14:37	awkorama	what do you guys use closure for?
14:37	Kruppe	awkorama: closures? or clojure?
14:38	awkorama	sorry, i meant clojure
14:38	awkorama	the language for jvm
14:38	Kruppe	awkorama: just making sure. I've used it for simulations mostly. Dunno about everyone else.
14:38	tensorpudding	hmm, i have the clojure-contrib installed, but it seems (use 'clojure.contrib.monads) doesn't work
14:38	Kruppe	awkorama: its a general purpose language though
14:39	awkorama	Kruppe: sure, but what about traditional 3 layer enterprise web app?
14:39	awkorama	can it handle these?
14:39	tensorpudding	and now i've got it talking about classpaths
14:39	awkorama	what about annotations and stuff? i found no info on these in clojure
14:39	Kruppe	awkorama: I don't see why not, but I haven't tried it myself.
14:40	Kruppe	awkorama: you have access to all of javas libs and can basically write java in clojure if you want.
14:40	zakwilson	Clojure has the ability to talk to databases and render templates or generate HTML.
14:40	raek	tensorpudding: how are you starting clojure? lein?
14:40	awkorama	zakwilson: including transaction management?
14:41	tensorpudding	I installed clojure from the Ubuntu repositories
14:41	tensorpudding	/usr/bin/clojure -r
14:41	rlb	FUser: I belive records aren't "maps" at all, i.e. they're compiled classes where the members are fields.
14:41	tensorpudding	it's a script
14:41	rlb	(and can have optional type annotations)
14:41	FUser	that's cool
14:41	tensorpudding	does some environment mashing
14:41	FUser	I'll use precompiled version then
14:41	raek	tensorpudding: I wouldn't recommend starting clojure that way. it makes classpath management very hard.
14:42	raek	tensorpudding: for just "give me a repl" use http://github.com/liebke/cljr
14:42	FUser	I use clojurebox
14:42	tensorpudding	eeeenh
14:42	zakwilson	awkorama: ClojureQL appears to handle transactions. I have not used it though.
14:42	tensorpudding	i really was wanting to not have to use much in the way of hand-compiled stuff
14:42	raek	tensorpudding: for projects (i.e. code in more than one source file) use http://github.com/technomancy/leiningen
14:44	raek	the thee major clojure tools (lein, cake, irclj) have all been released since ~november 2009..
14:44	tensorpudding	hmm
14:45	kmc	i'm using leiningen even for tiny 1-file projects
14:45	kmc	it's pretty easy
14:45	tensorpudding	leiningen is like cabal?
14:45	kmc	tensorpudding, don't be that guy ;)
14:45	tensorpudding	what do you mean?
14:45	kmc	"hi #haskell, please tell me which Haskell function is like GetXW3byF97() from PHP thanks"
14:45	tensorpudding	i was asking you
14:46	kmc	thait said, leiningen is like cabal ;)
14:46	djahandarie	GetXW3byF97 only works in PHP kmc, Haskell isn't powerful enough to support it
14:46	kmc	but it can also do stuff like launch a repl
14:46	raek	it manages dependencies and the class path, among other things
14:46	SergeyD	tensorpudding: ubuntu's clojure is quite old - version 1.0.*
14:46	kmc	and it can build a jar with all your clojure libs baked in, so you can distribute it to Java zombies
14:47	tensorpudding	i've got 1.1.0 here
14:47	raek	basic workflow for a new project: lein new foo; cd foo; <editor> project.clj; lein deps; lein repl
14:47	tensorpudding	on maverick
14:47	tensorpudding	it sounds like cabal but even better
14:47	tensorpudding	hmm
14:47	raek	you declare dependencies in project.clj, and leiningen takes care of downloading it and putting it on the classpath
14:48	raek	cljr is like lein, but with one global project
14:48	tensorpudding	how is the classpath handled? env mangling?
14:48	raek	it constructs it from what's in project.clj
14:49	tensorpudding	okay
14:49	tensorpudding	does lein work with clojure 1.1.0?
14:49	raek	every project has its own lib/ directory with its versions of its dependencies
14:49	tensorpudding	oh hey, it's named after that story about the ants
14:49	raek	yes, just change [clojure "1.2.0"] to [clojure "1.1.0"] in project.clj
14:49	tensorpudding	i read that in middle school
14:49	kmc	tensorpudding, when i set up lein i did it in a way that downloaded the newest clojure automatically
14:49	rlb	I'm still not happy with what appears to be the java (and apparently inherited in clojure) practice of "just run this cmd and don't worry about where all the code comes from, who wrote it, or whether or not it's been vetted". I hope eventually that changes (i.e. wrt maven, clojars, etc.).
14:50	tensorpudding	kmc: downloaded it where?
14:50	kmc	tensorpudding, http://github.com/technomancy/leiningen
14:50	raek	into ~/.m2
14:50	kmc	yeah
14:50	tensorpudding	one of the things i strongly dislike about cabal is how it defaults to putting things in the home directory
14:50	kmc	i had to apt-get install maven first
14:50	raek	it uses maven for the dependency stuff
14:50	Once	If anyone is familiar with fnparse please take a peek at http://paste.lisp.org/+2H88 Very simple example, and I'm not sure what the best solution is. Problem is exampled in the paste.
14:50	tensorpudding	so i don't know where things are
14:50	kmc	which in turn downloaded about 80 MB of Apache Very Serious Enterprise Java libraries
14:51	Once	problem is explained*
14:51	tensorpudding	though i guess clojure basically is just a couple jarfiles isn't it?
14:51	raek	it is.
14:51	kmc	tensorpudding, yep, there's a jar for clojure and another for clojure-contrib
14:51	kmc	and another for each 3rd-party lib you install
14:51	tensorpudding	why ~/.m2?
14:51	raek	that's why you basically have to start it with some tool
14:52	raek	maven 2, I think
14:52	tensorpudding	what's maven?
14:52	tensorpudding	is that java's build tool?
14:52	raek	a java build tool
14:52	raek	yes.
14:52	tensorpudding	okay
14:52	raek	leiningen uses the dependency parts of it
14:52	tensorpudding	where are the scripts installed?
14:52	raek	which scripts?
14:53	tensorpudding	do i have to edit my $PATH?
14:53	tensorpudding	the lein script
14:53	raek	yes
14:53	raek	the lein script is the only script
14:56	ihodes	hey all! it's been a while. anyone here familiar with Compojure?
14:59	zakwilson	ihodes: I am a little bit.
14:59	amalloy	zakwilson: if booleans said that they would be adorable
15:00	ihodes	zakwilson: well i'm using it and Enlive to serve up some static pages on my local machine. problem is, the CSS will not load. I've tried relative and absolute paths, and opening the page myself (with the relative path) as a static file in safari displays the CSS just fine. any ideas?
15:00	zakwilson	Instead of loading the CSS, what does it do?
15:01	ihodes	the raw HTML displays, as if unstyled
15:03	LauJensen	ihodes: paste your app ?
15:03	raek	ihodes: what happens if you enter the stylesheet url in the browser?
15:04	tensorpudding	okay, so i have lein setup
15:04	tensorpudding	how do i get emacs to use it?
15:04	raek	tensorpudding: install ELPA in emacs first
15:04	raek	if you don't already have it
15:04	tensorpudding	i already did
15:05	tensorpudding	to install clojure-mode
15:05	raek	ok, then install packages slime, slime-repl and clojure-mode
15:05	ihodes	(html/deftemplate index "main.html" [] )
15:05	ihodes	(defroutes example
15:05	ihodes	(GET "/" [] (render (index)))
15:05	raek	slime is for interacting with the clojure process
15:05	ihodes	(route/not-found "Page not found"))
15:05	LauJensen	ihodes: You also need to tell your app to serve sttic files
15:05	LauJensen	static
15:06	ihodes	LauJensen: what do you mean?
15:06	ihodes	LauJeson: it's serving the static HTML just fine; just without the CSS
15:06	ihodes	main.html is in resources/
15:06	tensorpudding	i've used slime before
15:06	LauJensen	ihodes: Nothing in what you pasted is telling Compojure to serve CSS files right? So when you template looks for /css/main.css, it wont find anything
15:06	tensorpudding	with CL
15:06	raek	ihodes: you need to tell your app that GETting /foo.css means serving the file resources/foo.clj
15:06	LauJensen	ihodes: What it does, is read main.html from disk, and serve it via a template
15:07	raek	tensorpudding: then add [swank-clojure "1.2.1"] to your project as a :dev-dependency
15:07	LauJensen	ihodes: if all your static files are in resources/, try adding the middleware (wrap-file "resources")
15:07	zakwilson	ihodes: paste the whole thing to a pastebin?
15:07	raek	do lein deps, and then lein swank
15:07	ihodes	LauJeson: do i need to include that middleware somewhere? that sounds like what I want to do
15:07	ihodes	LauJenson
15:08	raek	you need both slime and slime-repl
15:08	raek	tensorpudding: example project.clj http://gist.github.com/555485
15:09	amalloy	ihodes: try LauJensen :)
15:09	raek	lein swank should start a swank server on port 4005, which you can connect to with emacs with M-x slime-connect
15:09	LauJensen	ihodes: Yea you do. I havent used Compojure in a long time, but I think its (defroutes ex (wrap-file "resources") (GET "/" ....
15:10	LauJensen	Also make sure you import it in your namespace declaration ring.middleware file
15:10	LauJensen	(thanks amalloy)
15:10	tensorpudding	oh lame
15:11	tensorpudding	got an exception
15:11	zakwilson	Is there a currently accepted "best" set of tools to use for web stuff in Clojure?
15:12	tensorpudding	oh wait, right, need to change versions
15:12	LauJensen	zakwilson: Depends on who you ask. There's primarily Moustache and Compojure. I like Moustache a lot and its very similar to Enlive (by the same author).
15:12	tensorpudding	no, that didn't fix it
15:12	LauJensen	I also like Compojure. But not enough to use it
15:13	ihodes	LauJensen: what are the main differences? i'm just starting a project using Enlive; should I give Moustache a look first?
15:13	raek	tensorpudding: swank-clojure 1.2.1 requires clojure 1.2 I think
15:13	LauJensen	ihodes: I think you should yea. Its a very elegantly designed micro-webframework. Check out the readme on its github page.
15:13	zakwilson	I have been using Compojure, though I haven't touched the web part of the app that uses it in a while.
15:13	raek	I'm currently very happy with the Ring + Moustache + Enlive combo
15:14	tensorpudding	raek: i thought leiningen grabbed the right version of clojure for you?
15:14	ihodes	i'll give Moustache a go, then. do you know if Chris will be updating Enlive or Moustache anytime soon?
15:14	tensorpudding	the exception is related to ant, it seems
15:15	raek	leiningen picks the clojure version you tell it to
15:15	tensorpudding	it's grabbing 1.2.0, supposedly
15:15	LauJensen	ihodes: AFAIK I know there are no plans in the immediate future
15:15	raek	tensorpudding: could you paste the stack trace?
15:16	zakwilson	Moustache looks like it might be good. I hate the name though.
15:16	raek	if you mix [clojure "1.2.0"] with [clojure-contrib "1.1.0"] for example, it will not work
15:17	raek	since clojure-contrib is ahead-of-time compiled for 1.1
15:17	raek	and clojure doesn't guarantee binary compatibility between verisons
15:17	freakazoid	I implemented base62 in Clojure - does anyone care if you have to convert, say, a sha1 hash to a BigInteger first?
15:18	freakazoid	the semantics are slightly different than base64.
15:18	tensorpudding	http://hpaste.org/40648/lein_dep_fails
15:18	freakazoid	I could make it the same as base64 by treating a byte stream as a base 256 number after the decimal point.
15:18	tensorpudding	my project.clj is the same as the one you gisted
15:18	raek	tensorpudding: have you executed "lein self-install"?
15:19	tensorpudding	yes
15:20	raek	which version of lein? the one linked on the project page?
15:20	tensorpudding	from git
15:20	tensorpudding	1.4.0-SNAPSHOT, apparently
15:20	raek	ah
15:20	raek	you have to build the bleeding edge version with leiningen...
15:21	tensorpudding	bleeding edge version of which?
15:21	raek	of leiningen itself
15:21	raek	http://github.com/technomancy/leiningen/raw/stable/bin/lein
15:22	raek	that one is 1.3.1
15:22	tensorpudding	is there a stable branch?
15:22	raek	yes
15:23	raek	the installation instructions likns to that one
15:25	tensorpudding	dammit, i hate git
15:26	kryft	tensorpudding: Why?
15:26	tensorpudding	because it's like mercurial, but more confusing
15:26	tensorpudding	but everyone uses it, and no one uses mercurial
15:28	tensorpudding	trying to remember how to manage remote branches is one thing which i don't fully get
15:29	raek	git has a steep learning curve
15:29	tensorpudding	okay, there, i've checked out stable
15:30	raek	you only need the file I linked
15:31	raek	everything else is .jars that it downloads automatically when you do lein self-install
15:31	tensorpudding	i know
15:32	tensorpudding	but since i bothered checking out the lein source, and made a /usr/local/bin symlink from it for the lein script, i didn't want to hardcode my lein script
15:33	tensorpudding	it's working now
15:34	tensorpudding	ugh, it even created a .gitignore
15:34	tensorpudding	i suppose clojure is a heavy git community
15:35	raek	yes, I suppose so...
15:35	tensorpudding	i'll have to edit the lein scripts to make an .hgignore instead
15:35	raek	source on github and jars on clojars seems to be very common
15:37	ihodes	LauJensen: i'm having trouble rending my template using Moustache now; http://pastebin.org/263049
15:37	ihodes	make that rendering
15:37	rplevy	what is a reasonable test of pmap vs map that would demonstrate speedup on a 4 core machine?
15:39	rlb	I'm working with maildirs from clojure -- does anyone know if it's possible to get the filename from a javax maildir message?
15:40	raek	,(time (dorun (map (fn [_] (java.lang.Thread/sleep 20)) (range 100))))
15:40	clojurebot	"Elapsed time: 2127.817 msecs"
15:40	raek	,(time (dorun (pmap (fn [_] (java.lang.Thread/sleep 20)) (range 100))))
15:40	clojurebot	"Elapsed time: 90.17 msecs"
15:41	raek	pmap is useful when the computation to do for each element takes long time
15:41	raek	for very small computations, it is not recommended
15:42	rplevy	raek: thanks!
15:43	raek	np :)
15:44	SergeyD	rake: Nice clean example. But I can't figure out why the difference is 20x times on a 2-core processor :)
15:44	ihodes	anyone here know mucha botu Moustache?
15:45	raek	ihodes: you might need to use wrap-file-info too
15:46	SergeyD	raek: oh, I got it.
15:46	ihodes	raek: i'm not even there yet haha — i'm having trouble getting moustache to render my template
15:46	raek	oh, please share!
15:47	ihodes	http://pastebin.org/263049
15:47	raek	thought the thread pool was 2 x no-of-processors
15:47	raek	ihodes: you don't need to do the apply str step
15:48	LauJensen	[""] (response (templatename args))
15:48	LauJensen	ihodes: ^^
15:48	raek	Ring accepts sequences of strings (those that the templates return)
15:48	raek	ah, yes. that makes sense.
15:48	ihodes	raek: just getting rid of "render" didnt work.
15:48	ihodes	LauJensen: response is in which ns?
15:50	raek	ihodes: (defn ok [body] {:status 200, :headers {"Content-Type" "text/html; charset=UTF-8"}, :body body})
15:50	raek	^ this is what I use
15:50	raek	[""] (fn [_] (ok (index)))
15:50	raek	or, you could move that (fn ..) into ok
15:51	ihodes	raek: perfect!
15:51	ihodes	raek: that's awesome. thanks so much!
15:51	raek	the rigth hand side of the route rule should eval to a ring handler
15:51	raek	:)
15:51	ihodes	raek: didn't realize i needed to be passing a map back
15:51	ihodes	raek: i'll read more about ring. i just started using this about 20 mins ago :\ time for some real reading now. thanks so much
15:52	raek	ihodes: this is what you need to know: http://github.com/mmcgrana/ring/blob/master/SPEC
15:52	SergeyD	raek: Seems like the thread pool is about 30 on my computer
15:52	raek	basically, what keys are in the request and response
15:53	ihodes	raek: awesome! anything i should know about Moustache beyond what's in the little doc cgrand has?
15:53	raek	well, everything about moustache follows from syntax.html... if you know how to interpret it
15:54	SergeyD	raek: 32 actually, 16 for each core
15:54	raek	just remember that things (app [..] <here>) are threaded differently fomr things (app <here>)
15:54	Raynes	Moustache is so dead simple, the docs he provides is amazingly sufficient.
15:54	raek	*treated
15:54	Raynes	It's compojure that needs more documentation.
15:54	raek	even though thosw two cases overlap somewhat
15:55	raek	wrapping the first case in [ ] always yields the second case
15:55	ihodes	raek: i don't know how; what should i read in order to interpret that haha—i'm familiar with grammars etc, so maybe a guide eists somewhere to introduce this to me? would be useful for my continuing to work with enlive as well
15:55	raek	ihodes: by looking at the example he provides, I think most of should be pretty clear
15:56	raek	the grammar is nice if you want to be certain about the details
15:56	ihodes	raek: ah nevermind, this grammar makes perfect sense. and the docs look nice. i think i can handle this :)
15:56	raek	if you are familiar with grammars, then it should be pretty straight forward, I think
15:56	ihodes	raek: thanks again!
15:57	ihodes	raek: yeah it's not as bad as i was expecting; pretty much a cfg haha
15:57	Raynes	$seen cemerick
15:57	sexpbot	cemerick was last seen quitting 1019 minutes ago.
15:59	Raynes	-> (/ 1019.0 60)
15:59	sexpbot	⟹ 16.983333333333334
16:01	amalloy	Raynes: sexpbot is yours, right? if the minutes aren't helpful, why not get him to tell you hours as well?
16:01	Raynes	If one more person asks me that, I'm going to make sexpbot timestamp intelligently.
16:02	Raynes	That's at least 6 people in the last two months. :p
16:02	amalloy	heh. well, i don't mind the minutes myself, but since the author seems to find minutes inconvenient...
16:02	raek	,((juxt #(/ % 60) #(rem % 60)) 1019)
16:02	clojurebot	[1019/60 59]
16:02	Raynes	It's been in my invisible TODO list forever. I just haven't gotten around to it. I haven't been getting around to much lately.
16:04	amalloy	,((juxt quot rem) 1019 60) ; raek?
16:04	clojurebot	[16 59]
16:05	amalloy	sexpbot: source?
16:05	KirinDave	Wait what?
16:05	KirinDave	Ah
16:05	Raynes	$whatis sexpbot
16:05	sexpbot	sexpbot = http://github.com/Raynes/sexpbot
16:06	raek	/facedesk
16:07	amalloy	raek: it's a good thing you're a founding member of the juxt fan club or we'd have to kick you out
16:07	raek	:-)
16:08	ihodes	Raynes: you should make sexpbot timestamp intelligently!
16:08	ihodes	;)
16:09	Raynes	He's right, you know.
16:09	Raynes	Just as soon as I get a good night's sleep rather than the 4 hour increments I've been getting over the last two days.
16:09	amalloy	Raynes: i'll fork and do it myself if you want. looks pretty easy in seen.clj
16:10	Raynes	I'd love you forever.
16:10	LauJensen	Such an affectionate young man
16:12	Derander	I'm trying to implement a file watcher daemon in clojure. Currently, I have a function like this: http://gist.github.com/63122 . In ruby, C, or other languages that I know, I'd slap that logic in a while loop and put a sleep of some reasonable amount of time in
16:13	Derander	I have no idea how to do this idiomatically in clojure
16:13	Derander	Most clojure programs that I read are more of the "compute this thing" or "listen for this event", not the "do this thing indefinitely" variety
16:13	Raynes	Derander: Sure that's the right gist?
16:13	amalloy	(loop [] (do-stuff) (Thread/sleep 1000)) ; would work, dunno if it's idiomatic
16:14	amalloy	er. and put a (recur) at the end, sorry
16:14	Derander	Raynes: that is definitely not, weird
16:14	Raynes	(while true ..)
16:14	Derander	http://gist.github.com/631226 is
16:14	LauJensen	amalloy: busywaiting... I think there is a FileWatcher class actually, or maybe Im thinking of C#
16:14	LauJensen	Raynes: (defn x [] (while true body (recur))) = (defn x [] (recur ...))
16:14	Derander	LauJensen: When I was googling, I found some sort of FileWatcher mechanism in java 7
16:15	Raynes	LauJensen: I didn't tell him to recur in (while true ..)
16:15	Derander	alright, I'll explore recur. Thanks for the info
16:15	Raynes	amalloy told him to recur at the end of the loop.
16:15	Raynes	:p
16:15	LauJensen	Raynes: Oh right, sorry for not being able to tell you two Americans apart
16:16	Derander	LauJensen: I imagined that in a snotty british accent.
16:16	Raynes	:>
16:16	Raynes	Not quite.
16:16	LauJensen	Haha!
16:16	KirinDave	Derander: Write (forever fn)
16:17	KirinDave	Derander: Use the trampoline to recur, and make sure that it returns a future so the repl doesn't block.
16:17	KirinDave	I'm not sure how you'd enable killing, but it's not like that's the intent, anyways. :)
16:17	KirinDave	You could probably put in a kill switch.
16:17	Derander	KirinDave: I've never heard of a trampoline, I've read about futures but have no idea how to use them. I'll do some reading.
16:17	KirinDave	Then you can write (forever #'alone)
16:17	Derander	haha.
16:19	Raynes	You can use futures in places where you'd do (.start (Thread. (fn [] ..))) a lot of the time.
16:19	LauJensen	haha
16:19	Derander	Raynes: I must confess that I've never done threaded programming in clojure either. I've written a total of 62 lines in the language. This is my first project
16:20	Derander	but I'll read read read and figure it out
16:20	Raynes	Gotta start somewhere.
16:20	Derander	mmhmm. it's more approachable than haskell at least
16:20	LauJensen	Derander: 62 lines is a very good start, it corresponds to 824 lines of Scala of 1400 lines of PHP
16:20	LauJensen	roughly...
16:20	Raynes	LauJensen: What about Ruby
16:20	raek	(def running? (atom true)) (future (loop [] (if @running? do-stuff clean-up)))
16:20	KirinDave	LauJensen: That is math I want to see
16:20	LauJensen	Raynes: 74 ? :)
16:20	Derander	Raynes: approximately equivalent
16:20	KirinDave	Raynes: 62 lines as well, but only if you use eigenclasses.
16:21	Derander	I've been writing ruby for 3 years and never really used eigenclasses
16:21	LauJensen	KirinDave: 62 lines, but behavior will be unpredictable
16:21	Raynes	Derander: With all due respect to Haskell (I adore the language), a brick wall is more approachable at high speeds.
16:21	Derander	Raynes: yes
16:21	LauJensen	Haskell is fantastic.
16:21	KirinDave	LauJensen: 62 lines, but it each line will take 1 year to execute, and won't run if you've required ActiveRecord.
16:21	Raynes	Yes it is.
16:21	LauJensen	hehe
16:21	Derander	better: 62 lines, and 120 lines of unit testing.
16:21	KirinDave	Haskell is beautiful. There is a difference. ;)
16:21	KirinDave	Haha, all written in advance
16:22	LauJensen	Derander: In Ruby, even unit tests wont save you
16:22	KirinDave	With a mock framework more complex and powerful than almost any other project in the language.
16:22	Derander	LauJensen: I usually write integration tests for my ruby projects anyway
16:22	Derander	but that's because I am filled with hubris
16:23	LauJensen	I guess if you build commercial software in Ruby, you do deserve a bit of punishment
16:23	KirinDave	LauJensen: If you need unit tests to successfully complete a project, ruby is a rough road.
16:23	KirinDave	Some languages require champions to ride them.
16:23	KirinDave	You can quote me on that.
16:23	Derander	the biggest project that I've worked on in ruby is 15k lines of code
16:23	LauJensen	Ouch!
16:23	Derander	it's actually pretty nice
16:24	Derander	if the templating was in haml life would be perfect
16:24	LauJensen	Derander: nice as in, you got rich from bugfixing it ?
16:24	KirinDave	LauJensen: It's not a condemnation.
16:24	LauJensen	KirinDave: What is it then ?
16:24	KirinDave	LauJensen: It's a compliment, actually. Ruby is a superbike.
16:24	Derander	LauJensen: nah, nice as in it works pretty well and is easy to grok and work with
16:24	KirinDave	LauJensen: The problem is you rev the throttle too fast and suddenly you're SuperDave.
16:24	LauJensen	Derander: impressive, I thought that was impossible with Ruby
16:24	Derander	when I finish pushing it to 1.9.2 it'll be damn fast too
16:24	KirinDave	Hanging from your bike's handlebars, body horizontal with the acceleration.
16:25	KirinDave	I have more metaphors, if you need them.
16:25	Derander	ruby was my first language after php. we've had a long and tender relationship
16:25	tensorpudding	if you have a function which takes [& args], how can you mangle (arg arg) to make it work as input to the function?
16:27	tensorpudding	wait, apply seems to be it
16:27	aav	tensorpudding: (applu myfunc args) - if i got your question right
16:27	aav	apply
16:28	shanmu	Hi, I am trying to represent db records as clojure records - I am trying to do defrecord dynamically
16:28	tensorpudding	except doh, it's a macro, not a function
16:28	KirinDave	Derander: Same.
16:28	KirinDave	Except the PHP part.
16:28	KirinDave	Ruby was my first dynamic language that I loved.
16:28	KirinDave	:)
16:29	shanmu	I am using metadata to get column names as a map of string, and when I try to defrecord using them, hitting a classcastexception unable to cast String to IObj
16:30	KirinDave	So... I've been wondering.
16:30	KirinDave	What's the right way to make something that responds to @deref correctly?
16:30	tensorpudding	i'm really getting annoyed by how restrictive macros seem to be
16:31	KirinDave	tensorpudding: What in particular is bothering you?
16:31	tensorpudding	well, you can't apply them
16:32	tensorpudding	or use them as functions
16:32	Chousuke	if you need to apply a macro you probably have a weird macro :/
16:32	KirinDave	tensorpudding: You have to understand, when you say (defmacro ...) what you're saying is (open up my compiler and add....)
16:32	KirinDave	tensorpudding: The very act of applying things is generally a runtime process.
16:32	KirinDave	And as such the compiler has no place
16:32	tensorpudding	i didn't decide to make or and and macros
16:32	KirinDave	Ah.
16:32	KirinDave	The Classic Complaint™
16:33	KirinDave	and and or aren't actually macros.
16:34	KirinDave	They're special forms.
16:34	Chousuke	well, except that they are macros :P
16:34	tensorpudding	hmm
16:34	KirinDave	Chousuke: Implementation detail ;)
16:34	Chousuke	and no they're not really special forms
16:34	tensorpudding	,(reduce or true [true false true])
16:34	clojurebot	java.lang.Exception: Can't take value of a macro: #'clojure.core/or
16:34	Raynes	KirinDave: No, they seriously aren't special forms.
16:35	Chousuke	if is a special form, or and and are just macros
16:35	Raynes	-> (doc and)
16:35	sexpbot	⟹ "Macro ([] [x] [x & next]); Evaluates exprs one at a time, from left to right. If a form returns logical false (nil or false), and returns that value and doesn't evaluate any of the other expressions, otherwise it returns the value of the last expr. (and) returns true."
16:35	KirinDave	Raynes: Ah, i'm thinking CL.
16:35	Raynes	Not a special form.
16:35	tensorpudding	,(reduce (fn [x y] (or x y)) true [true false true])
16:35	KirinDave	Sorry.
16:35	clojurebot	true
16:35	KirinDave	tensorpudding: Yes
16:35	Chousuke	but the thing is that macros work with code
16:35	tensorpudding	if you wrap it in a lambda, it works
16:35	KirinDave	tensorpudding: Right, because a function is generated.
16:35	Chousuke	apply works with runtime data.
16:35	tensorpudding	i'm trying to translate something from haskell into clojure
16:35	tensorpudding	to learn the former better
16:35	Raynes	-> (some false? [true false true])
16:35	sexpbot	⟹ true
16:36	KirinDave	Raynes beat me to it.
16:36	Raynes	That essentially accomplishes the same thing without any macros.
16:36	amalloy	Raynes: sent a pull request for sexpbot
16:36	Raynes	amalloy: Thank you, kind sir. Your work is appreciated.
16:37	Raynes	KirinDave: It must be Haskellers Try Clojure week. I've seen a ton of guys from #haskell in here lately. :D
16:38	Chousuke	heh.
16:38	KirinDave	Raynes: Reaching for something. :)
16:38	amalloy	hmmm. does github have a "merge with upstream" function or do i have to figure out how to do it from git?
16:38	Chousuke	amalloy: pull from upstream, push into your fork? :/
16:39	KirinDave	Hum
16:40	KirinDave	I'm working on asciidoc-ing my clojure & the web chapter for my book
16:40	KirinDave	This is the final code for a URL shortener. I know it works, but any failures to emulate idiomatic code would be appreciated.
16:40	KirinDave	(err, pointing them out, that is)
16:40	KirinDave	https://gist.github.com/f3509ac3fe7859fe527d
16:41	KirinDave	And before you ask, the silly storage stubs are there so that couchdb can be hooked in later.
16:55	shanmu	Hi, Can I dynamically declare a record?
16:57	shanmu	I get an error when trying (def name-list ["id" "name"]) (defrecord org name-list)
16:58	shanmu	,(do (def name-list ["id" "name"]) (defrecord org name-list))
16:58	clojurebot	DENIED
16:58	chouser	you'll need to use eval or something
17:01	amalloy	or a macro. i think i gisted something like this a while back, let me find it
17:04	amalloy	well, i guess it's not that close. but the general idea is (defmacro record-from-strings [fields] `(defrecord ~@(map symbol fields))) (record-from-strings ["org" "id" "name"])
17:04	kjeldahl	http://stackoverflow.com/questions/678867/how-to-defn-a-function-from-string-in-clojure
17:10	shanmu	amalloy, kjeldahl: thanks!
17:23	Raynes	$seen cemerick
17:23	sexpbot	cemerick was last seen quitting 18 hours and 24 minutes ago.
17:23	Raynes	amalloy_: My love. <3
17:40	amalloy_	Raynes: woo!
17:40	amalloy	~seen cemerick
17:40	clojurebot	no, I have not seen cemerick
17:40	amalloy	heh. win for sexpbot
17:41	Raynes	Maybe he stores the info in memory. If so, a restart would wipe it (which is a possibility).
17:42	KirinDave	Oh I hope he doesn't have a data store for seen
17:42	Raynes	A seen database can get really huge.
17:43	Raynes	sexpbot using mongodb for seen (among a lot of other things).
17:44	amalloy	clojurebot uses (ref {}) to store the seen list
17:44	Raynes	Yeah, in memory. It probably just restarted since cemerick left.
17:44	amalloy	yeah
17:59	shanmu	I had to use (defmacro record-from-strings [fields] `(defrecord ~@(map symbol fields))) (record-from-strings ["org" "id" "name"]), does that look alright?
18:00	shanmu	oops! wrong paste
18:00	shanmu	I had to use (defmacro record-from-strings [rec-name fields] `(defrecord ~(symbol rec-name ) ~(map symbol fields))) (record-from-strings "org" ["id" "name"])
18:01	shanmu	does that look alright?
18:01	amalloy	yeah, looks reasonable. i forgot it doesn't want the @ there
18:02	shanmu	amalloy: thanks a lot! this solves my problem nicely
18:03	amalloy	shanmu: just keep in mind it will only work with literals - you can't do like (def x ["foo"]) (record-from-strings "org" x)
18:04	shanmu	oh.. ok... I was planning to do defrecords for table columns in a db (obtained via c.c.sql)
18:05	amalloy	macros can't have access to runtime data. they happen before runtime
18:05	amalloy	but there is a way to do it if your db info will be known at compile time
18:05	shanmu	amalloy: thanks! is there any way to do it? the db info would be available during compile time...
18:06	amalloy	shanmu: see http://clojure-log.n01se.net/date/2010-10-17.html, especially the link to http://gist.github.com/630662
18:06	shanmu	amalloy: thanks! looking thorugh
18:17	clojurebot	The Tao, considered as the Rabbit, began.
18:40	amalloy	shanmu: that help solve your problem?
18:42	shanmu	amalloy: no, not exactly... the problem is I donot know my macros... reading up on them now
18:42	shanmu	I tried (defmacro record-from-strings [fields] `(defrecord ~@(map symbol fields))) (record-from-strings ["org" "id" "name"])
18:42	shanmu	oops!
18:42	shanmu	I tried (defn create-record [rec-name fields] `(defrecord ~(symbol rec-name) ~(map symbol fields)))
18:42	shanmu	then (defmacro make-record [rec-name fields] `(~(create-record rec-name fields)))
18:43	shanmu	but (let [n "recName" f ["f1" "f2"]] (make-record n f))
18:43	shanmu	errors with Don't know how to create ISeq from: clojure.lang.Symbol
18:44	amalloy	shanmu: (let [blah] (stuff)): if stuff is a macro, it can't access the contents of blah because that's happening at runtime
18:45	amalloy	the way i would solve a problem like yours is to define (define-sql-records) as a macro, which calls (lookup-sql-tables), a function
18:46	amalloy	define-sql-records takes no arguments, or maybe one that's the name of the db to use or something; lookup-sql-tables queries the database and returns some code, which define-sql-records splices into its return value so that it's available at compile time
18:49	shanmu	amalloy\|gone: thanks for the advice! will try something
18:51	raek	making some Active Record-like system?
19:07	FUser	what's the difference between :while and :when in macro "for"?
19:07	FUser	,(for [x (range 15) :when (> x 5)] x)
19:07	clojurebot	(6 7 8 9 10 11 12 13 14)
19:08	FUser	,(for [x (range 15) :while (> x 5)] x)
19:08	clojurebot	()
19:09	FUser	,(for [x (range 10) :when (> x 5) :while (> x 5)] x )
19:09	clojurebot	(6 7 8 9)
19:10	FUser	,(for [x (range 15) :when (even? x) :while (> x 5)] x)
19:10	clojurebot	()
19:10	shanmu	raek: not a full fledged activerecord like system
19:10	FUser	im confused
19:13	Lajla	FUser, obviously looking at this, while stops after the first false, and when continues.
19:13	Lajla	But just doesn't put it into the list if it's false.
19:14	FUser	but check this out
19:15	FUser	,(for [x (range 15) y (range 10) :when (even? x) :while (> x 5)] x)
19:15	clojurebot	(6 6 6 6 6 6 6 6 6 6 8 8 8 8 8 8 8 8 8 8 10 10 10 10 10 10 10 10 10 10 12 12 12 12 12 12 12 12 12 12 14 14 14 14 14 14 14 14 14 14)
19:15	FUser	identical case as when it returns empty list
19:15	FUser	except it has an additional variable that isn't used
19:15	FUser	and it suddenly behaves totally differently
19:16	Lajla	I have no idea how that works no.
19:16	Lajla	Could even be a bug.
19:16	Lajla	Ask raek, he is omnipotent.
19:17	FUser	:D
19:17	FUser	you mean omniscient surely
19:19	Lajla	Well, if you can do all things, that includes knowing all things, or at least answering all quaestions.
19:27	FUser	:)
19:37	chouser	:when and :while control the previously specified loop
19:59	shanmu	amalloy: its working now!!
19:59	shanmu	(defmacro record-from-table [table-name]
19:59	shanmu	`(defrecord ~(symbol table-name) ~(map symbol (get-columns-map table-name))))
19:59	shanmu	(defn get-columns-map [table-name]
19:59	shanmu	(sql/with-connection db
19:59	shanmu	(map #(:column_name %1)
19:59	shanmu	(resultset-seq (-> (sql/connection)
19:59	shanmu	(.getMetaData)
19:59	shanmu	(.getColumns nil nil table-name nil))))))
19:59	shanmu	sorry, pasting it
20:00	amalloy	shanmu: great! yeah, i was about to say paste
20:01	shanmu	http://pastebin.com/4ihxfMED
20:01	shanmu	amalloy: thanks for the guidance! just like you said, a macro which calls a function to get column list
20:05	KirinDave	shanmu: #(:sym %1)
20:05	KirinDave	shanmu: Or
20:05	KirinDave	,(map :col [{:col 1}])
20:05	clojurebot	(1)
20:05	zakwilson	If I want clojure-contrib 1.3 from Leiningen, where do I find appropriate version strings? I see that it has been split in to core and [other stuff] but [org.clojure/clojure/contrib/core "1.3.0-alpha1"] seems like it should work (and doesn't)
20:05	KirinDave	shanmu: Understand?
20:06	duck1123	I have a bunch of files I'm reading for my tests (lazytest) and
20:06	duck1123	previously I had them in the root of my application, but I want to
20:06	duck1123	move them to src/test/resources (using maven) is slurp the best way to
20:06	duck1123	contiue reading them, and if so, should I specify the path in the src
20:06	duck1123	directory, or is there a better way?
20:06	duck1123	ahh, sory about the multi-line
20:08	dysinger	duck you could find the files in your classpath if you put them in maven's src/test/resources dir
20:09	shanmu	KirinDave: sorry, didnot understand that...
20:10	KirinDave	shanmu: You don't need the #().
20:11	KirinDave	shanmu: :keywords are invokable.
20:11	shanmu	KirinDave: got it! sorry, bad habit of adding an anon func as soon as I type map!
20:11	KirinDave	shanmu: Same. :)
20:12	ninjudd	zakwilson: [org.clojure.contrib/complete "1.3.0-SNAPSHOT"] should work
20:12	shanmu	KirinDave: thanks! will remember that keywords are invokable... :)
20:13	zakwilson	ninjudd: looks good so far. Thanks.
20:13	duck1123	dysinger: right, my question was what is the best way to now read those files from the classpath. (slurp "entry.xml") is no longer working.
20:13	clojurebot	Alice felt a violent blow underneath her chin - it had struck her foot! She was walking with nameless simplicity.
20:14	duck1123	I know there's an easy way to do it, but I can't remember it, and google is failing me
20:14	ninjudd	zakwilson: no problem. or you can just add the specific libs from contrib you want. e.g. [org.clojure.contrib/find-namespaces "1.3.0-SNAPSHOT"]
20:16	zakwilson	ninjudd: I'll start with everything, and get specific if I decide by uberjar is too uber.
20:19	ninjudd	zakwilson: that works. though if you are writing a library that you're going to put on clojars, it is probably best to have specific deps. looks like complete pulls in 61 jars!
20:20	zakwilson	I am not writing a library, and I am not putting what I'm writing on clojars.
20:20	zakwilson	Sorry. Super proprietary. Going to make me a zillionare. Etc.
20:21	ninjudd	zakwilson: i see. well then you'll be able to afford all 61 jars ;)
20:21	duck1123	I need to go through my projects and go down to the specific contribs I need
20:23	zakwilson	I'll even be able to afford sixty-TWO jars!
20:26	shanmu	thanks a lot people! #clojure rocks, as always....
20:26	dysinger	duck1123: (-> (Thread/currentThread) (.getContextClassLoader) (.getResourceAsStream "x/y/z.txt") (slurp))
20:27	dysinger	(slurps file from classpath)
20:29	duck1123	dysinger: ok, so maybe it wasn't as easy as I thought. Thanks a lot
20:30	amalloy	dysinger: the parens aren't necessary for one-arg functions in ->, eg
20:30	amalloy	,(-> 1 inc inc inc)
20:30	clojurebot	4
20:30	duck1123	you wouldn't by any chance to know how to get lazytest failures to fail the build?
20:31	amalloy	(whether that makes things more readable is down to taste, of course)
20:34	dysinger	duck1123: it'd be pretty easy to make a fn to slurp-from-cp
20:34	dysinger	and just use that instead of plain slurp
20:35	duck1123	just did. looking for instances of slurp atm. :)
20:35	dysinger	easy peasy
20:39	duck1123	dysinger: what are you using for code coverage with maven atm?
20:58	dysinger	duck1123: I'm using lein mostly
21:25	cemerick	duck1123: code coverage is orthogonal to build tooling
21:25	cemerick	I presume something like Clover would get the job done, though it'd be a bit of a mess insofar as it's not clojure-aware.
21:26	cemerick	Presumably there's some code coverage tools that have plugin architectures amenable to informing them about non-Java languages.
21:43	jstirrell`	hi
21:44	jstirrell`	what's the easiest way to get a new collection of, say, every third item from an old collection?
21:45	cemerick	,(take-nth 3 (range 20))
21:45	clojurebot	(0 3 6 9 12 15 18)
21:45	cemerick	jstirrell`: ^^
21:45	jstirrell`	sweet thanks!
22:04	Raynes	cemerick: Evening.
22:05	cemerick	Raynes: howdy
22:05	Raynes	How was your weekend?
22:05	cemerick	Very busy. Working on the book at the moment.
22:05	Raynes	Ooh, goody.
22:06	cemerick	Rewriting stuff this early on is surely a bad sign. :-/
22:06	Raynes	Only means you're a perfectionist. When kept in check, that can be a very valuable asset.
22:06	cemerick	heh, indeed
22:06	cemerick	Otherwise, it was a perfect autumn day.
22:07	Raynes	Indeed. It's been a beautiful week here. Nice and cool.
22:07	moogatronic	+1 for perfect autumn day.
22:07	cemerick	Warmish but breezy, mostly sunny, tons of fantastic treescapes.
22:07	cemerick	Breakfast out, then a small hike, then a visit to Old Deerfield.
22:08	Raynes	Well, I'm off to bed. I'll talk to you in the morning. I should be up, so just ping me whenever you get around to calling expedia (DOT COOOOOOOOM).
22:08	moogatronic	we're a little moisture starved in mid-southern indiana , burn ban in effect for most counties, going to impact my up-coming backpacking trip
22:08	cemerick	Raynes: Figure after 8:30 anyway
22:08	Raynes	Right, I should be up or nearly up.
22:09	Raynes	Goodnight, fair #clojure.
22:09	moogatronic	goodnight. =)
22:13	clojurebot	"You're nothing but a pack of cards, after all? Can you play croquet with the bitterest grief; and soldiers under sentence of execution?"
22:14	cemerick	wow, that was random
22:14	maravillas	clojurebot must be reading lewis carroll lately
22:16	maravillas	curious...both that quote and the one at 8:12 are from http://twitter.com/#!/WonderTao
22:17	maravillas	"Text built by mashing up Alice's Adventures in Wonderland with the Tao Te Ching. Created with Python & Markov chains."
22:17	cemerick	very random
22:36	Lajla	So how are things going in noTCOland now?
22:38	cemerick	same as they always have; see recur
22:38	cemerick	or trampoline
22:38	Lajla	I know them.
22:39	Lajla	I'm just feeling like that remark.
22:39	Lajla	I am actually the second best clojure programmer in the world, second only to the microsoft Chief Software Architect.
22:40	chouser	cemerick: hi
22:41	cemerick	chouser: bonsoir
22:41	Lajla	Terve
22:41	chouser	I feel like I've been gone
22:42	cemerick	There was certainly a disturbance in the Force.
22:43	chouser	heh
22:48	chouser	wow, vars are suddenly in flux?
22:48	cemerick	indeed
22:51	chouser	someone asked me at strangeloop, "what's next"
22:51	cemerick	"how much time do you have?"
22:52	chouser	which of course is a common question. and I'm always tempted to guess
22:53	chouser	I said "I don't know" and "maybe pods" ...not realizing dynamic binding and future/send had already changed
22:54	cemerick	mmm, the future/send change is pretty friggin' huge
22:54	cemerick	That "fixes" vars and binding quite a lot.
22:59	coldhead	thanks clojurebot
23:02	chouser	yep
23:45	cemerick	This seems like an unfortunate asymmetry:
23:45	cemerick	,(-> ::foo str keyword)
23:45	clojurebot	::sandbox/foo
23:45	cemerick	vs.
23:45	cemerick	,(-> 'foo str symbol)
23:45	clojurebot	foo
23:47	cemerick	Meaning, namespaced keywords aren't trivially string round-trippable.
23:56	chouser	,(-> :foo str keyword)
23:56	clojurebot	::foo
23:56	chouser	,(-> `foo str symbol)
23:56	clojurebot	sandbox/foo
23:57	chouser	,(-> :foo name keyword)
23:57	clojurebot	:foo
23:57	chouser	cemerick: I think I'm missing your point
23:57	cemerick	I'm tired; perhaps my point is foolish.
23:58	cemerick	I think one should be able to do (-> ::foo str keyword) and get the proper keyword out.
23:58	cemerick	As things stand, you have to munge the string before keywordizing it.
23:58	hiredman	,(pr-str ::foo)
23:58	clojurebot	":sandbox/foo"
23:59	cemerick	sure
23:59	cemerick	my point was about asymmetry, not about not being able to do the thing
23:59	chouser	ah, I see what you're saying
23:59	chouser	pr-str is the inverse of eval, though