#clojure – 2009-05-20

2009-05-20

00:08	duck1123_	How come I can't negate thrown? within an is macro?
00:09	duck1123_	I wanted to simply verify that a certain exception wasn't thrown
00:44	arohner	duck1123_: because "is" is a multimethod that dispatches on the first form passed to it
00:45	arohner	I assume you'd like to do (is (not (thrown? ...)))
00:45	arohner	you have a good use case. post a message to the group
01:33	seangrove	hey guys, wondering what the options are for automated testing before I dive into the clojure world?
01:41	cads	what converts :key into "key"?
01:41	hiredman	,(doc name)
01:41	clojurebot	"([x]); Returns the name String of a symbol or keyword."
01:42	cads	'doh I ought to know that
01:44	arohner	seangrove: there is test-is in clojure-contrib
01:46	seangrove	arohner: have you worked with it much?
01:46	arohner	yeah. It's running as we speak :-)
01:46	arohner	it's your standard test framework
01:46	arohner	oh, there's also fact
01:47	seangrove	ah, cool
01:47	arohner	http://github.com/weavejester/fact/tree/master
01:47	arohner	it's kind of like haskell's quickcheck, if you're familiar with that
01:47	seangrove	haven't dabbled with clojure at all yet, just wanted to make sure that was covered before I get too far into it
01:47	seangrove	arohner: nah, I haven't done any heavy functional programming
01:48	seangrove	mainly just some mall scheme and CLOS projects
01:48	cads	is there a function sym such that (apply #([% (class %)]) (sym "dog" 123 'cat :bat)) => [dog123catbat clojure.lang.Symbol]
01:49	hiredman	that would not work
01:49	hiredman	for a number of reasons
01:49	hiredman	,(symbol "foo")
01:49	clojurebot	foo
01:50	cads	,(symbol (str "rat" "cat" 123 "hellbat"))
01:50	clojurebot	ratcat123hellbat
01:50	replaca	,(symbol (str :bat))
01:50	clojurebot	:bat
01:50	cads	,(symbol (str "rat" "cat" 123 (name :bong) "apple"))
01:50	clojurebot	ratcat123bongapple
01:50	hiredman	a. (sym "dog" 123 'cat :bat) would try and call sym on the args "dog" 123 'cat :bat and there is no function named sym
01:51	hiredman	b. your function literal there takes one arg
01:51	replaca	hiredman: he as using sym as the free variable in his question
01:51	replaca	*as => was
01:51	cads	I think I could write sym in 3 lines
01:51	hiredman	wait
01:51	hiredman	so you got "sym" but couldn't find "symbol" on the api page?
01:52	replaca	java-utils/as-str might help here
01:52	replaca	hiredman: you're not understanding what he meant by "sym"
01:52	replaca	replace it with foo in his question
01:53	replaca	is there a function "foo" such that ... (foo "dog" 123...)
01:53	replaca	?
01:54	hiredman	replaca: yes, and the function symbol is more or less what he was looking for
01:54	cads	hired man, I'm supposing a function named sym which operates like str on a variadic list of operands to let you concatenate a symbol name from a few components of different types.
01:55	hiredman	,((comp symbol (partial apply reduce str)) "dog" 123 'cat :bat)
01:55	clojurebot	java.lang.IllegalArgumentException: Don't know how to create ISeq from: Keyword
01:55	cads	,(name "str")
01:55	clojurebot	java.lang.ClassCastException: java.lang.String cannot be cast to clojure.lang.Named
01:56	cads	,(symbol "te:st")
01:56	clojurebot	te:st
01:57	cads	,(defn sym [& glargs] (symbol (apply str glargs)) (sym "hyper" 'rat :cat)
01:57	clojurebot	EOF while reading
01:57	cads	,(defn sym [& glargs] (symbol (apply str glargs))) (sym "hyper" 'rat :cat)
01:57	clojurebot	DENIED
01:58	cads	hehe, it yiels hyperrat:cat
02:06	replaca	(defn catsym [& args] (symbol (apply str (map as-str args))))
02:06	replaca	where as-str is defined in clojure.contrib.java-utils
02:07	replaca	(catsym "dog" 123 'cat :bat) => dog123catbat
02:08	replaca	cads: I think that's what you're after, right?
02:09	replaca	~def as-str
02:09	replaca	~doc as-str
02:09	clojurebot	Titim gan �ir� ort.
02:09	replaca	oh well
02:34	jdz	as far as i can see there is no way to throw an exception with some data in it, right?
02:34	replaca	jdz: well, you can, but you need to create your own exception class and gen-class it
02:35	replaca	jdz: and then it will probably be wrapped
02:35	jdz	that is what i want to avoid doing
02:35	jdz	because gen-class requires compilation
02:36	jdz	something akin to CL's throw/catch would be cool, and could be implemented using Java exceptions
02:36	replaca	I used a method in cl-format where I built error information lower down with a special function
02:37	replaca	you might also want to look at Chouser's error-kit in contrib
02:37	jdz	that is actually what i need, non local transfer of control with some data passing
02:37	jdz	ye, will look at error-kit, maybe there's something cool.
02:37	replaca	in cl-format, I only wanted to be able to end up with a more sophisticated RuntimeException, so it may not be sufficient for you
02:38	replaca	I haven't tried error-kit yet, but I want to :-)
02:38	replaca	but I haven't written code that does anything but die on errors since Chouser wrote it
02:39	jdz	there's a pretty hackish thing in contrib/except
02:39	replaca	oh, I haven't looked at that. I thought it was just cleaning up error messages
02:41	replaca	looking at contrib/except, it doesn't look like what you're after
02:41	replaca	just convenience wrappers
02:43	jdz	hmm, looking at error-kit it seems i can use proxy after all...
02:43	jdz	somebody's really smart around here :)
02:44	replaca	yeah, Chouser's spent a lot of time understanding how Clojure really works
02:45	replaca	are you looking at his proxy + IDeref trick?
02:45	clojurebot	proxy is <Chouser> proxy teases with its ease of use, then suddenly betrays.
02:45	jdz	ye, the IDeref thing
02:46	replaca	yeah, that's clever. I want to play with it, too
02:48	replaca	time for bed!
03:14	cads	replaca, that's it :)
03:15	cads	~def str
04:16	opqdonut	can i somehow "store" a set of bindings that's in effect at some point, and eval something in that context later?
04:18	opqdonut	essentially, i have a macro that creates a datastructure with expressions inside it, and i want to be able to create a closure over that macro: "(defn [arg] (macro ... something using arg ...))"
04:18	opqdonut	the expressions aren't evalled until much later, and arg won't be in scope then
04:20	hoeck	opqdonut: with "bindings", do you mean dynamic vars?
04:21	hoeck	opqdonut: or just values bound in a let-form?
04:27	jdz	opqdonut: macros are run at compile time; when your program executes, all macros have been substituted by corresponding code.
05:02	opqdonut	hoeck: let-form
05:14	wnong	a terrible beginner here (to clojure as well as IRC)
05:16	wnong	i am a the getting started with the prag-prog book and am getting error with jline/consoleRunner (NoClassDefFoundError)
05:23	jdz	never used that, so all i can suggest is to use the exact version of the software that is used in the book.
05:23	wnong	i have the same version from the book.
05:25	wnong	so far i have found ..googling that it is something to do with the absolute path (of jline.jar). I tried this too.
05:25	Chousuke	probably misconfigured classpath
05:25	Chousuke	it's in the wrong directory
05:25	Chousuke	the jar that is
05:25	wnong	i have all the jars in the /lib directory
05:25	wnong	including the clojure.jar
05:26	Chousuke	and how do you run clojure?
05:27	wnong	i have been doing simple test runing the clojure.jar itself
05:27	wnong	but now i am going by the book and trying to run the script (repl.sh)
05:27	wnong	and get this error "Exception in thread "main" java.lang.NoClassDefFoundError: jline/ConsoleRunner"
05:28	Chousuke	probably the repl.sh sets up the classpath wrong.
05:29	Chousuke	lisppaste8: url
05:29	lisppaste8	To use the lisppaste bot, visit http://paste.lisp.org/new/clojure and enter your paste.
05:29	Chousuke	paste its contents there
05:29	wnong	ok
05:31	lisppaste8	wnong pasted "prog-clojure-book-getting started" at http://paste.lisp.org/display/80539
05:31	wnong	done
05:32	Chousuke	is htat CLOJURE_DIR path correct?
05:32	Chousuke	does it contain your /lib?
05:32	jdz	and are the newlines really where they appear in the paste?
05:32	wnong	yes
05:32	wnong	yes
05:33	wnong	there are 3 lines
05:33	jdz	there is a space between : and classes
05:34	wnong	i fixed that now and same error
05:34	jdz	and $CLOJURE_JAR is not used, btw.
05:35	Chousuke	jdz: that doesn't matter; it's in the classpath anyway
05:35	jdz	yes, i'm just saying
05:35	jdz	if it's not used, it should not be theer
05:35	wnong	yes, i was trying to modify it according the wiki-book
05:35	jdz	there even
05:36	Chousuke	did you check if the lib dir actually contains jline-0.9.94.jar
05:36	wnong	I tried testing the jline itself going thru http://en.wikibooks.org/wiki/Clojure_Programming/Getting_Started#Linux
05:37	wnong	yes jline version is in the lib dir
05:38	wnong	from the link above, i was trying out : java -cp $CLOJURE_DIR/jline-VERSION.jar:$CLOJURE_JAR \ jline.ConsoleRunner clojure.lang.Repl
05:38	Chousuke	uh, VERSION?
05:38	wnong	ofcourse replacing the version with my jar-name
05:38	Chousuke	right.
05:39	Chousuke	and it's the 0.9.94 version?
05:39	yason	Which sounds better: a ref containing a hash-map that contains other refs as values, or a ref containing just a hash-map with direct values and the whole hashmap is updated with update-in?
05:39	wnong	yes it is
05:39	Chousuke	really weird
05:39	Chousuke	ensure that your CLOJURE_DIR has no typos
05:39	yason	In other words: is it ok to sow refs around like mad rather than use them sparingly?
05:39	Chousuke	that's the only thing I can think of anymore.
05:40	wnong	ok..will check again. thx for ur time
05:41	Chousuke	yason: if you're using many Refs, you're most likely trying to emulate mutability somehow.
05:41	Chousuke	yason: the fewer refs you have, the less code you have to coordinate with dosync.
05:41	yason	Chousuke: I have a single hashmap that contains most of the mutable state of my program.
05:42	hoeck	yason: depends on your granularity needs
05:42	Chousuke	a single ref for that is fine in many cases.
05:43	yason	Chousuke: it's a live database of stuff I need to update and read all the time. Currently I have one ref that is used to hold the whole hashmap
05:44	Chousuke	at the very least, you should avoid storing each database entry in a ref.
05:44	yason	hoeck: That brought me to this question: I need to change the data in the hashmap's values quite often but the values themselves are rather independent of each other
05:44	yason	hoeck: On the other hand, I need to read / filter values from the whole database at times, which would mean dereffing a lot of refs
05:45	hoeck	yason: maybe use atoms then for the hashmaps values?
05:45	yason	Chousuke: Ok, so refs can be considered expensive, complexity-wise?
05:45	yason	hoeck: If they're lighter, yes. I haven't got much "feel" for the differences of refs and atoms yet
05:45	Chousuke	well, they're trickier than just having immutable data, that's for sure :)
05:46	Chousuke	atoms are just holders for data that are atomically swappable, with no coordination of how the changes happen.
05:46	yason	Chousuke: How would you handle long-running mutable data in Clojure anyway? I can write most of the functionality of my program with immutable data but I need to store stuff somewhere
05:47	Chousuke	a ref is fine for that.
05:47	tWip	long running? how long? perhaps you could consider a database
05:47	Chousuke	clojurebot has its entire database of definitions in a single map with a single ref.
05:47	Chousuke	it's not very high traffic though.
05:48	Chousuke	but adding to the map is cheap even if it grows big.
05:48	yason	tWip: days, weeks. A database would be ok if I could easily shove Clojure objects there
05:49	tWip	using JDBC from Clojure is pretty easy
05:49	yason	tWip: isn't JDBC relational?
05:49	tWip	but I think Clojure would also be good for prevalence type solutions, if you want to have everything in memory
05:49	tWip	yeah jdbc is for sql databases mostly
05:50	Chousuke	if you want more granularity, you could also divide your database maps
05:50	cddr	In java, you can get at "this" during the constructor. Is that possible in a proxy?
05:51	yason	tWip: I'm more after object databases.
05:51	Chousuke	yason: make 4 maps (in 4 refs), then use some functions to transparently select which of the maps to write to when you add to the database; when you read, you can pretty easily combine the four maps
05:52	tWip	for some value of "object" I guess...
05:52	Chousuke	yason: though, overwriting might be a problem :/
05:52	Chousuke	hmm
05:52	Chousuke	nah, a single map would probably be faster :P
05:53	yason	Chousuke: I'll have think about scaling later anyway
05:53	yason	Chousuke: but using atoms as the values sounds good now that I realize refs do a lot more than atoms
05:55	yason	tWip: Objects as in live Clojure primities, to be more exact. (Or as live as possible.:))
05:55	Chousuke	hmm. if you wrap the values in atoms, it allows anyone to overwrite a value at anytime in the DB (atomically!), but they can't add to the database
05:55	Chousuke	without going through the main ref that is
05:56	yason	Chousuke: adding to the database probably won't be a problem
05:56	yason	Chousuke: the single ref will probably handle that gracefully
05:57	Chousuke	yeah
05:57	yason	Chousuke: seems that swap! maps the old value through a function that returns the new value
05:57	Chousuke	yeah.
05:57	yason	Chousuke: in my case that's ideal since the value is in the transitions, not the exact value of the atom at each given time
05:58	Chousuke	sounds good, then.
05:59	yason	Chousuke: What would've happened if I had put refs as values and wrapped a sweep over the values in a single dosync?
05:59	clojurebot	What is meta
06:00	yason	Chousuke: It would've compiled a list of refs in the transaction and somehow synced all of them for the calling thread?
06:00	Chousuke	yason: a dosync block restarts if there is a conflict
06:02	Chousuke	so you should actually do (doseq [[k v] @map] (dosync (alter v somefunc))) so that the ENTIRE map isn't resweeped
06:02	Chousuke	if you were using refs
06:03	yason	Chousuke: so in practice the transactions would be likely to restart many times during concurrent access?
06:03	Chousuke	depends on how small the transaction is.
06:04	Chousuke	in my dosync, the transaction only concerns one ref at a time; if you move the dosync outside doseq, it'll concern the whole map.
06:04	yason	Chousuke: I was imagining mapping/filtering (seq @myhashmap) or something inside a dosync
06:05	yason	Chousuke: Indeed it sounds more sensible to not dosync the whole operation
06:05	Chousuke	you should use doseq instead of a map, since map is lazy
06:06	Chousuke	that might not matter if the time when the transition happens doesn't matter to you.
06:06	Chousuke	ie. if everything is commutative
06:09	yason	Chousuke: In this case I had in mind doall would probably been it
06:10	yason	but anyway, thanks again for your thoughts
06:11	yason	I'll hopefully be able to contribute something back to the channel after hacking with Clojure long enough :)
08:25	mib_9i23ye	has anyone used clojure to develop for mobile devices?
08:32	cddr	how would one return the first element of a sequence for which (pred elem) returns true?
08:35	Chouser	(first (filter pred coll))
08:36	MikeSeth	thats an interesting question, is there a jme port anywhere?
08:36	cddr	Chouser: so does filter return a lazy sequence?
08:36	Chouser	cddr: yes
08:36	vy	Chouser: Isn't it quite inefficient that it requires to scan whole list regardless of the first non-NIL result?
08:37	cddr	cool
08:37	vy	Oops. Cool.
08:38	mib_9i23ye	j2me anyone?
08:39	rhickey	mib_9i23ye: I think j2me might be too minimal for Clojure
08:44	Cark	sooo there it is ... i have my first clojure program in production !
08:44	Cark	a license server for another program, i already talked about it
08:44	Chouser	Cark: nice!
08:44	Cark	i'm pretty happy =)
08:45	rhickey	Cark: congrats!
08:46	Cark	we tested it with 20 thousand simultaneous connections, looks like it will be good for a couple years
08:46	Cark	using nio
08:47	Cark	all of the ssl/nio stuff was done in clojure itself
08:47	Cark	actually everything was done in clojure
08:48	MikeSeth	i should start pushing this shop here off php and to clojure
08:49	Cark	what i didn't like (always easier to talk about the bad things) : spotty completion in emacs (not using slime)
08:49	Cark	also : multimethods didn't cut it for the all connection->sll connection->line connection->license connection thing
08:50	Chouser	Cark: because of flexibility or performance?
08:50	Cark	flexibility
08:51	Cark	m�hh how to explain this
08:52	Cark	i was missing something like call-next-method
08:52	Cark	i think that's the gist of it
08:52	Chouser	ah.
08:53	Cark	anyways, building anonymous functions and putting these in a map achieves the goal, so that was no big deal
08:54	rhickey	Cark: there is get-method
08:54	Chouser	there was some discussion of that, which resulted in a new feature, but I don't remember the specific ...
08:54	rhickey	,(doc get-method)
08:54	clojurebot	"([multifn dispatch-val]); Given a multimethod and a dispatch value, returns the dispatch fn that would apply to that value, or nil if none apply and no default"
08:55	Chouser	then again, fn's in a map are likely faster, so if they serve the purpose then great.
08:56	Cark	right, i didn't see how get-method would be usefull, it was adding another field in my map
08:56	rhickey	while not a blind next-method, you could (get-method foo-method my-parent-dispatch-val)
08:56	chessguy_work	can a macro raise an error?
08:56	MikeSeth	rhickey: is it your estimate that clojure wont be available on mobile platforms any soon?
08:57	Chouser	,(doseq x)
08:57	clojurebot	java.lang.IllegalArgumentException: doseq requires a vector for its binding
08:57	Chouser	chessguy_work: like that?
08:57	chessguy_work	yep :)
08:57	marklar	MikeSeth: You can write Android apps with it already
08:57	chessguy_work	so it's a compile-time error then
08:57	rhickey	MikeSeth: people are playing with it on Android now
08:57	Chouser	chessguy_work: sure. or the macro can emit code which throws the exception, and then you have a runtime error.
08:57	MikeSeth	mmmm delicious
08:58	marklar	MikeSeth: Check out http://github.com/remvee/clojurehelloandroid/tree/master
09:01	chessguy_work	how evil is it for a macro to evaulate some of the code passed t oit
09:01	Cark	chessguy_work : you mean using eval ?
09:02	chessguy_work	or just unquoting and applying a function
09:02	Cark	well that's one of the goals of macros
09:03	Cark	the doseq macro, it does evaluate the body
09:03	Cark	,(doseq [i '(1 2)] (println i))
09:03	clojurebot	1 2
09:03	Chouser	using 'eval' is hardly ever what you want, even in a macro
09:04	Cark	the println form gets evaluated
09:04	Chouser	Cark: yes, but not by using 'eval' in the macro.
09:04	Cark	right
09:04	chessguy_work	yeah, having done a lot of JS programming, i tend to read 'eval' as 'evil
09:04	chessguy_work	''
09:05	chessguy_work	,(expand-macro '(doseq [i '(1 2)] (println i))
09:05	clojurebot	EOF while reading
09:06	chessguy_work	,(expand-macro '(doseq [i '(1 2)] (println i)))
09:06	clojurebot	java.lang.Exception: Unable to resolve symbol: expand-macro in this context
09:06	chessguy_work	bah, what the heck is it calle
09:06	svdm	macroexpand
09:06	chessguy_work	,(macroexpand '(doseq [i '(1 2)] (println i)))
09:06	clojurebot	(loop* [G__2700 (clojure.core/seq (quote (1 2)))] (clojure.core/when G__2700 (clojure.core/let [i (clojure.core/first G__2700)] (do (println i)) (recur (clojure.core/next G__2700)))))
09:06	chessguy_work	eep
09:08	chessguy_work	err
09:08	chessguy_work	what did doseq evaluate there?
09:09	MikeSeth	rhickey: out of curiousity, has it ever occured to you that lispy languages need better code visualization during development?
09:09	MikeSeth	(the question goes to everyone really)
09:09	Chouser	it didn't evaluate anything -- it inserted the (println i) expression into its output, so that it would be evaluated at runtime.
09:10	chessguy_work	that's what i thought
09:10	chessguy_work	so the macro doesn't evaluate any of its parameters
09:10	chessguy_work	do other macros?
09:12	rhickey	MikeSeth: both enclojure and IntelliJ's La Clojure have structure browsers, enclojure's also has multimethod grouping
09:12	jdz	chessguy_work: macros don't evaluate their parameters, by definition.
09:12	chessguy_work	like, they literally can't?
09:12	MikeSeth	last time i tried enclojure it was too broken :(
09:13	MikeSeth	but that was a long time ago
09:13	jdz	chessguy_work: macro is just a function, which gets some clojure values in, and produces another clojure value.
09:13	jdz	chessguy_work: the values it gets in are not evaluated.
09:14	jdz	chessguy_work: as opposed to functions, which always receive their argumentns evaluated.
09:14	jdz	arguments even
09:14	Cark	but nothing prevents a macro evaluating its arguments
09:14	jdz	yes, one just has to remember when the macros are being executed, which is compile time, not runtime.
09:15	Chouser	chessguy_work: you can use 'eval' in a macro, but usually it's not the right way to solve the problem at hand.
09:15	chessguy_work	Chouser, is taht the only way to evaulate in a macro?
09:17	asbjxrn	MikeSeth what do you mean by code visualization?
09:23	Chousuke	chessguy_work: you can use ~ inside syntax-quote
09:23	Chousuke	or hm
09:24	chessguy_work	to unquote it, right?
09:24	Chousuke	I guess that doesn't actually evaluate them
09:24	chessguy_work	but then you could (let [foo ~foo] (foo 1 2)) right?
09:25	Chousuke	,(let [foo 5] [`foo `(foo ~foo)])
09:25	clojurebot	[sandbox/foo (sandbox/foo 5)]
09:25	Cark	let's say you want to make a loop unrolling macro (unroll 3 (println "hello"), you could check at macro expansion time that the first argument is an integer and use it to make your 3 calls to println
09:26	Cark	or actually to output the 3 println forms
09:26	Chousuke	but checking explicitly for an integer would prevent it from working with variables
09:26	Cark	one could say that the first argument is evaluated
09:26	Chouser	right, but (let [x 3] (unroll x (println "hello"))) wouldn't work.
09:26	Cark	you wouldn't unroll a variable number of times
09:27	Cark	that's part of this macro's api
09:27	Chousuke	right.
09:27	Cark	or specification
09:28	Cark	though that's some braindead macro right there =)
09:29	Chousuke	(defmacro unroll [times body] `(do ~@(replicate times body)))
09:29	Chousuke	does that work?
09:30	Chousuke	hm, I guess it needs the &
09:30	Chousuke	no, I get a nullpointer exception ;(
09:30	chessguy_work	d'oh
09:30	chessguy_work	der suckage
09:31	chessguy_work	surely that's a bug
09:31	Chousuke	in my macro perhaps
09:31	chessguy_work	err
09:32	chessguy_work	sorry, a null pointer exception at compile time is a compiler bug, IMHO
09:32	Cark	(defmacro unroll [times & body]
09:32	Cark	`(do ~@(replicate times `(do ~@body))))
09:32	Cark	ugly =)
09:32	Chouser	,(intern 'sandbox '#^{:macro true} unroll (fn [times body] `(do ~@(repeat times (cons `do body)))))
09:32	clojurebot	#'sandbox/unroll
09:32	Chouser	,(unroll 3 (println "hi"))
09:32	clojurebot	"hi"
09:33	chessguy_work	although...i guess at this point, if you're evaluating things at compiletime, you've really got turing-complete compilation
09:33	Chouser	,(unroll 3 (print "hi "))
09:33	clojurebot	"hi "
09:33	Chouser	hm.
09:33	chessguy_work	,(macroexpand '(unroll 3 (println "hi")))
09:33	clojurebot	(do (do println "hi") (do println "hi") (do println "hi"))
09:33	Chouser	doh
09:34	chessguy_work	,(do (do println "hi") (do println "hi") (do println "hi"))
09:34	clojurebot	"hi"
09:34	Chousuke	chessguy_work: yes
09:34	chessguy_work	,(do (println "hi") (println "hi"))
09:34	clojurebot	hi hi
09:35	Chouser	,(intern 'sandbox '#^{:macro true} unroll (fn [times body] `(do ~@(repeat times `(do ~body)))))
09:35	clojurebot	#'sandbox/unroll
09:35	Chouser	,(unroll 3 (println "hi"))
09:35	clojurebot	hi hi hi
09:36	Chousuke	(defmacro unroll [times body] `(do ~@(repeat times body))) works too
09:36	cddr	what's the equivalent of funcall?
09:36	MikeSeth	asbjxrn: well, since lisp syntax is trivial, you do not have to write and edit code as plain text
09:36	Chouser	cddr: you don't need it. just put the variable in the first position.
09:37	cddr	ah
09:37	Chouser	cddr: lisp 1 ftw.
09:37	Cark	(defmacro unroll [times & body] `(do ~@(apply concat (replicate times body))))
09:37	Cark	that's better !
09:37	MikeSeth	asbjxrn: with some thought I suppose you can visualize code and relations between code parts as e.g. 2d graphs, 3d models and so on
09:40	cp2	Unspecified vulnerability in Java Runtime Environment (JRE) for Sun JDK and JRE 6 Update 10 and earlier; JDK and JRE 5.0 Update 16 and earlier; and SDK and JRE 1.4.2_18 and earlier allows untrusted applets and applications to gain privileges via unknown vectors related to "deserializing calendar objects."
09:40	cp2	hmm
09:43	Chouser	,(intern ns '#^{:macro true} _defmacro (fn [n & a] `(intern ns '~(with-meta n {:macro true}) (fn ~@a))))
09:43	clojurebot	#'sandbox/_defmacro
09:43	Chousuke	cp2: I tested a proof of concept of that vulnerability :P
09:43	Chousuke	it worked pretty well ;(
09:44	cp2	yeah, so it seems
09:44	asbjxrn	MikeSeth: I see. I've never been a big fan of software through pictures. Text is very "condensed", making pictures out of it doesn't really add much I find.
09:44	Chousuke	asbjxrn: perhaps you've never seen a good visualisation :)
09:45	MikeSeth	asbjxrn: well, I am talking here mostly about elimination of irrelevant visuals, quick navigation and so on, not about VB-style "yes you too can be a programmer" development
09:45	asbjxrn	That is entirely true, I never have :)
09:48	asbjxrn	MikeSeth: I understand, and I'm probably the least experienced programmer here, but I find proper indentation of lisp code gives a reasonably good and dense representation of the code.
09:49	MikeSeth	asbjxrn: without question. I'm just saying that original approaches could be superior
09:49	asbjxrn	When visualizing code in my head, it's more about abstractions and invariants than the tree as such.
09:52	Chousuke	with an image you could densely represent a much larger set of relations though.
09:52	Chousuke	so it might be good for getting an overview.
09:53	Chousuke	or quick code review: if it looks sensible, you probably don't need a redesign.
09:56	asbjxrn	Chousuke: It's possible, and it would be interesting to see something like that, but the challenge is that you need to see the details at the same time.
10:03	Chousuke	http://code.activestate.com/recipes/576766 D:
10:22	marklar	Is there a clojure way to write/read structs to/from a file? Or are we supposed to use the Java IO classes?
10:25	Chouser	you can print and read maps as text, if you want.
10:27	marklar	Chouser: In clojure? I've got it writing to a file using the Java IO, but was wondering if there is a cleaner way.
10:30	Chouser	,(prn {:a 1 :b 2})
10:30	clojurebot	{:a 1, :b 2}
10:30	marklar	Chouser: thanks!
10:30	Chouser	you could write that to a file, and read it in again later
10:31	Chouser	that's all I mean. It's very good for some scenarios, not so much for others.
10:31	marklar	Chouser: That was exactly what I was looking for I think, it prints to out right?
10:31	marklar	I just need something very simple
10:32	Chouser	Yep, use (binding [out file-stream-thing] (prn ...))
10:32	marklar	k, thanks again!
10:32	hoeck	marklar: you may want to use clojure.contrib.duck-streams/spit too
10:32	marklar	hoeck: I'll take a look, I need something very simple though so I think prn will work
10:32	marklar	thanks
10:33	hoeck	marklar: that's the simplest way of writing sth. to a file I can imagine :)
10:33	marklar	hoeck: then I'll definitely take a look :)
10:36	alrex021	How do I pronounce clojure? Is it like closure? :)
10:36	rhickey	alrex021: yes
10:36	alrex021	thx
11:18	Chouser	is the only way to compile from the command-line like: java -cp clojure.jar clojure.main -e "(compile 'my.thing)" ?
11:19	Chouser	well, I guess it's sufficient.
11:20	rhickey	there is clojure.lang.Compile
11:21	Chouser	hmph. I was looking for that in main.clj
11:22	Chouser	ok, thanks.
11:33	triddell	I see two main json libraries, danlarkin's and stuart sierra's in contrib... anyone know why someone might use one over the other?
11:33	danlarkin_	triddell: at this point I'd say probably go for stuart's
11:33	danlarkin_	unless you see some compelling reason to use mine
11:34	liwp	has anyone seen the following on windows with slime:
11:34	liwp	16:19 <Chouser> ok, thanks.
11:34	liwp	16:25 *** svdm QUIT Read error: 104 (Connection reset by peer)
11:34	liwp	16:25 *** svdm JOIN
11:34	liwp	16:29 <triddell> I see two main json libraries, danlarkin's and
11:34	liwp	stuart sierra's in contrib... anyone know why
11:34	liwp	someone might use one over the other?
11:34	liwp	16:30 <danlarkin_> triddell: at this point I'd say probably go for
11:34	liwp	stuart's
11:34	liwp	> has anyone seen the following on windows with slime: java.io.FileNotFoundException: Could not locate swank/swank__init.class or swank/swank.clj on classpath: (NO_SOURCE_FILE:0)
11:34	liwp	urgh, sorry about that
11:34	AWizzArd	Is there a function (maybe in contrib) that allows me to ask yes/no questions as input from the keyboard?
11:35	liwp	has anyone seen the following on windows with slime: java.io.FileNotFoundException: Could not locate swank/swank__init.class or swank/swank.clj on classpath: (NO_SOURCE_FILE:0)
11:35	liwp	basically slime fails to run clojure and keeps polling until I abort it
11:36	triddell	danlarkin_: I saw he stated "If you want indented output, try the clojure-json library at http://github.com/danlarkin/clojure-json"... which would be nice, but I didn't know of any other major differences.
11:36	cp2	AWizzArd: dont think so
11:36	dnolen	liwp: are you using Clojure 1.0?
11:36	cp2	maybe you should make one!
11:36	cp2	and submit
11:36	cp2	to contrib
11:36	liwp	dnolen: svn HEAD
11:36	dnolen	liwp: using 1.0
11:37	dnolen	er use i mean.
11:37	Chouser	AWizzArd: read-line is the closest thing I know of.
11:37	liwp	dnolen: OK, I'll give it a try. The same setup works fine on OS X at home, so I'm a bit confused. Thanks.
11:39	AWizzArd	Chouser: probably that's enough for me
11:39	danlarkin_	triddell: yeah they're very similar at this point. stuart's has the advantage of being in contrib, so one less dependency
11:41	triddell	danlarkin_: ok, thanks for the feedback
11:43	alrex021	I've seen the ClojureBox for Win project. Looks really cool. Are there any instructions (or recommendations) for getting a similar setup for a Mac OS X?
11:45	liwp	dnolen: FYI: it's broken with 1.0 as well. I must have messed up my slime environment somehow. I'll have to look into it more carefully some other time. Cheers.
11:47	marklar	liwp: i've been dealing with the same issue on a linux box, haven't been able to figure it out either :(
11:50	chessguy_work	,(let [f (fn [x] (. Character isDigit x))] (f 5))
11:50	clojurebot	false
11:50	chessguy_work	,(let [f (fn [x] (. Character isDigit x))] (f '5'))
11:50	clojurebot	Unmatched delimiter: )
11:50	AWizzArd	I have a file src/de/company/foo.clj which looks like (ns de.company.foo) (defn main [& args] (apply println "Args:" args)). Now I would like to call that main function from the command line. How can I do this? java -cp src/;clojure.jar de.company.foo.main 10 20 does not work.
11:50	chessguy_work	,(let [f (fn [x] (. Character isDigit x))] (f "5"))
11:50	clojurebot	java.lang.IllegalArgumentException: No matching method found: isDigit
11:52	chessguy_work	,(let [f (fn [x] (. Character isDigit x))] (f \5))
11:52	clojurebot	true
11:53	cp2	marklar: marklar marklar, marklar marklar marklar!
11:55	marklar	cp2: I've always thought the same thing
11:55	cp2	heh
11:55	cp2	sorry, i just saw your name and it reminded me =)
11:55	marklar	cp2: np :)
11:59	chessguy_work	,(\5)
11:59	clojurebot	java.lang.ClassCastException: java.lang.Character cannot be cast to clojure.lang.IFn
12:00	chessguy_work	,\5
12:00	clojurebot	\5
12:01	AWizzArd	java -cp src/;clojure.jar clojure.main -e "(do (use 'de.company.foo) (main 1 2 3))" also does not work. Any ideas how I can call that main function from the command line?
12:01	rhickey	AWizzArd: did you try that with latest trunk?
12:02	AWizzArd	no, I will do this right now
12:03	Cark	,(doc defsymbolmacro)
12:03	clojurebot	java.lang.Exception: Unable to resolve var: defsymbolmacro in this context
12:04	Cark	there is a thread talking about symbol macros on the mailing list, is that from a library ?
12:06	Cark	ok found it in clojure contrib =)
12:06	Cark	now why would this be bad ?
12:12	Chousuke	what, defsymbolmacro?
12:13	cp2	11:02 < Cark> ok found it in clojure contrib =)
12:13	Cark	symbol macros, rich is saying he's "a bit concerned by the need of symbol macros"
12:15	cp2	what is a symbol macro?
12:15	Cark	some kind of a macro that willl be expanded with just the symbol
12:16	Cark	like i could expand the symbol 'hello to (println "hello")
12:16	cp2	oh
12:16	cp2	hm
12:16	cp2	not sure i like that idea =\
12:17	Cark	that's a silly example, but there are cases where it would be usefull
12:17	cp2	yeah i can imagine
12:17	Chouser	CL has them?
12:17	Chousuke	yes
12:52	rhickey	any good tokyo cabinet libs for Clojure?
12:53	rsynnott	well, there's a java api...
12:54	rhickey	rsynnott: got that
12:55	rsynnott	it's a simple enough library in terms of interface provided that a wrapper might nearly be overkill
12:55	rhickey	a wrapper would let it work with all Clojure map and seq functions
12:55	clojurebot	map is LAZY
12:56	rsynnott	oh, I suppose so, yes
12:57	cddr	has anyone written a "clojure for cl programmers" tutorial?
12:58	danlarkin_	I have a couchdb library that I'm working on... but that's not tokyo cabinet...
13:01	Chouser	we're using tokyo cabinet, but not from clojure (yet)
13:01	technomancy	anyone using S3 from clojure?
13:01	Chouser	fascinating array of "no"s there.
13:02	technomancy	it looks like it's a simple matter of a bunch of rest calls, not really much room for a library to abstract things
13:03	rhickey	I wonder how well tokyo cabinet's threading/lock stuff plays with Java
13:03	rhickey	technomancy: not tyrant, the cabinet directly embedded
13:04	rhickey	how about jdbm?
13:11	danlarkin_	crickets
13:16	cddr	has anyone written a "clojure for cl programmers" tutorial?
13:16	cddr	sorry, didn't mean to report my earlier comment
13:17	technomancy	cddr: there's a video of a presentation rich gave to a Lisp group that would probably be helpful
13:17	technomancy	it's not too specific to CL, but it addresses folks coming from a lispy background
13:22	AWizzArd	btw rhickey, with the new checkout I can run java -cp src/;clojure.jar clojure.main -e "(do (use 'de.company.foo) (main 1 2 3))" and indeed call my main method with that successfully.
13:22	AWizzArd	Thanks for the hint :)
13:29	rhickey	cddr: not that I know of, I presume you've seen http://clojure.org/lisps
13:33	technomancy	oh snap; the oreilly scala book is free online
13:33	technomancy	... and the second word of the first chapter is "enterprise". didn't see _that_ coming. =)
15:01	triddell	rhickey: Clojure-DBM is an interface to key/value databases like Tokyo Cabinet... http://github.com/weavejester/clojure-dbm/tree/master ... from the man behind compojure. I think it was just an initial but something to fork and enhance maybe
15:02	sopel	hey
15:03	sopel	any idea how to cast a Clojure vector to a byte array so that i can write it to a java outputstream?
15:05	Chouser	vector of...? Integers?
15:05	sopel	yeah
15:05	Chouser	,(into-array Byte/TYPE (map byte [1 2 3 4]))
15:05	clojurebot	#<byte[] [B@8ccbe7>
15:06	sopel	oh, thanks.
15:06	sopel	i'm working with sockets, it's pretty cumbersome.
15:08	durka42	hmm, SwingUtilities/invokeLater swallows exceptions doesn't it
15:13	Chouser	rhickey: What would you require of a Contributor's employer to be comfortable with patches originally owned by that employer?
15:13	Chouser	a specific signed document with their name and yours?
15:13	Chouser	a general written corporate policy that employees my submit patches to open source projects?
15:13	Chouser	s/my/may/
15:14	rhickey	Chouser: should be a CA from employer
15:14	sopel	how to find the names of classes of base clojure datastructures? i want to derive them
15:15	sopel	for multimethod dispatching
15:15	Chouser	so the employer would then retain the copyright
15:15	Chouser	,(class {})
15:15	clojurebot	clojure.lang.PersistentArrayMap
15:15	Chouser	sopel: like that?
15:15	Chouser	,(class (hash-map))
15:15	sopel	:)
15:15	clojurebot	clojure.lang.PersistentArrayMap
15:15	Chouser	,(class (hash-map :a 1))
15:15	clojurebot	clojure.lang.PersistentHashMap
15:15	rhickey	Chouser: yup
15:15	sopel	yea pretty simple. thanks. still learning.
15:15	Chouser	rhickey: ok, thanks.
15:15	rhickey	from the CA: "If this contribution is on behalf of a company, the term 'you' will also mean the
15:15	rhickey	company you identify below"
15:16	Chouser	sopel: note that there are many specific derived types, and some important interface types.
15:16	rhickey	ooh, La Clojure adds a REPL!
15:16	sopel	ok.
15:17	mattrepl	after "use"ing a Clojure source file containing derives that don't explicitly reference a hierarchy, there seems to be no affect on the global hierarchy. is that correct behavior?
15:17	noidi	is it possible to use ctags with clojure?
15:18	Chouser	rhickey: ah. hm -- but that looks like my signature would be sufficient if my employer has a policy that authorizes me to sign the contract.
15:18	sopel	Chouser; oh, LazilyPersistentVector alread ypopped up.
15:19	mattrepl	for example, using c.c.json.write and trying (json-str (doto (java.util.ArrayList.) (.add 1) (.add 2))) will print using ArrayList#toString instead of the proper formatting "[1,2]"
15:19	rhickey	Chouser: this has to be true:
15:19	rhickey	With respect to your contribution, you represent that:
15:19	rhickey	o it is an original work and that you can legally grant the rights set out in these terms;
15:20	danlarkin	mattrepl: c.c.json probably doesn't have a default dispatch for j.u.ArrayLists, you can add one with defmethod though
15:20	something	I'm trying to consume PersistentStructMap from jruby, not sure how to access the values in the map.
15:20	mattrepl	but after a load-file of c.c.json.write.clj, the hierarchy is in place and json-str works correctly.. or rather, the multimethod print-json works
15:20	mattrepl	danlarkin: that's what I thought, but there is one for java.util.Collection which should work for ancestors such as java.util.ArrayList, and it does after the load-file
15:21	rhickey	Chouser: so, depending on your policy/authorization, yup
15:22	sopel	hum, interesting. i wrote some bytes to a socket but didn't flush it. and the other end read null bytes from the socket. when i flush it it gets the right data on the other end.
15:22	rhickey	Chouser: I've gotta run soon, does that make sense?
15:22	Chouser	rhickey: well, I'm trying to help form the policy
15:22	Chouser	I've never had so much fun. :-P
15:22	rhickey	fair enough :)
15:22	Chouser	yep, thanks.
15:22	danlarkin	mattrepl: curious...
15:23	rhickey	the Sun CA is certainly commercial friendly, they got Google and Red Hat to sign it. Contributors really give up nothing, just share
15:25	danlarkin	mattrepl: use parents/ancestors/descendants and isa? to query the hierarchy
15:25	danlarkin	after the use
15:26	mattrepl	danlarkin: did that, and didn't see modification as I would expect if print-json isn't dispatching correctly
15:29	danlarkin	mattrepl: after using I run: (ancestors java.util.Collection) and it returns #{:clojure.contrib.json.write/array java.lang.Iterable}
15:29	mattrepl	good, at least it's just me... one moment
15:31	mattrepl	yeah, I just restarted the session and I'm only seeing Iterable after the use. must be configuration error, thanks for sanity check
15:32	something	I tried to use the indexer in jruby or get method to access PersistentStructMap. It always return nil. the toString of the map shows there is data.
15:33	danlarkin	mattrepl: what version of clojure are you running?
15:33	mattrepl	it's post 1.0, I believe
15:34	marklar_	How does one checkout the 1.0 version of clojure from the googlecode page? I've never used svn before...
15:34	danlarkin	mattrepl: ok, just checking... I'm on latest SVN
15:38	hiredman	marklar_: I would just download the zip
15:39	marklar_	hiredman: I need to make some changes
15:39	hiredman	I am not 100% sure but I think the zip includes sources
15:40	danlarkin	marklar_: you can just use the path when you checkout... so it's whatever... svn co code.google.com/clojure/tags/1.0
15:40	marklar_	hiredman: you know, I think you're right
15:40	marklar_	danlarkin: thanks, tags was what I was looking for
15:43	cemerick	argh, I've been down this path before: if I require some namespace with an alias (:require [com.foo.bar :as bar]), how do I name a keyword using that alias? :bar/kw is kw in the bar namespace, not the com.foo.bar namespace...
15:44	hiredman	this does sound familar
15:45	cemerick	Rich set me straight on it about 6 weeks ago, but it's come up again, I can't find my code from then, and I can't find the conversation in the channel log. ;-/
15:45	danlarkin	how about using the two-argument form of keyword?
15:46	cemerick	danlarkin: ech. There really is an elegant reader form that does it.
15:46	hiredman	2009:Apr:30:05:14:01 rhickey : ::myns/foo
15:47	danlarkin	human google
15:47	hiredman	well, I have my personal logs here on disk so I can just grep
15:47	cemerick	hiredman and clojurebot are actually just skynet v0.9
15:49	danlarkin	speaking of, what the heck is this company thinking, naming themselves cyberdyne and building robots?!? http://www.cyberdyne.jp/english/index.html
15:49	cemerick	nice
15:50	p_l	danlarkin: afaik only exoskeletons for now
15:50	danlarkin	kara thrace harbinger of death
15:50	cemerick	that's like the old joke about Wang coming to America, and not understanding what all the snickering was about.
15:55	hiredman	clojurebot: functional programing?
15:55	clojurebot	Huh?
15:55	hiredman	clojurebot: functional programing is <reply>http://jackcoughonsoftware.blogspot.com/2009/05/teaching-functional-programming-to-kids.html
15:55	clojurebot	Ik begrijp
15:56	p_l	danlarkin: also, notice the name of the system...
15:56	p_l	danlarkin: "Hybrid Assistive Limb, and HAL are trademarks (TM) or registered trademarks (�) of CYBERDYNE, Inc."
15:56	danlarkin	HAL? yeah, they're so hip
15:57	danlarkin	downfall of humanity right here
15:57	p_l	Nah, they are only making simple stuff. Fear the AGI research :D
15:58	something	is there an opposite of the bean function?
15:59	hiredman	nope
15:59	hiredman	because beans are class and to make new classes you need to AOT compile
15:59	hiredman	well
15:59	hiredman	beans are instances of classes
15:59	something	AOT?
16:00	hiredman	ahead of time
16:00	something	i see.
16:00	something	it seems like it's possible
16:01	hiredman	http://gist.github.com/34229 is a old (possibly not working) and lacking features (like types) macro for easily generating beans
16:02	hiredman	of course you would still need a function to map a map to a bean
16:02	sopel	how to make map non-lazy?
16:02	hiredman	(doc doall)
16:02	clojurebot	"([coll] [n coll]); When lazy sequences are produced via functions that have side effects, any effects other than those needed to produce the first element in the seq do not occur until the seq is consumed. doall can be used to force any effects. Walks through the successive nexts of the seq, retains the head and returns it, thus causing the entire seq to reside in memory at one time."
16:02	hiredman	(doc dorun)
16:02	clojurebot	"([coll] [n coll]); When lazy sequences are produced via functions that have side effects, any effects other than those needed to produce the first element in the seq do not occur until the seq is consumed. dorun can be used to force any effects. Walks through the successive nexts of the seq, does not retain the head and returns nil."
16:02	something	hiredman, cool thanks
16:03	sopel	i'm using str for debugging, any neater way to print datastructures?
16:03	sopel	'cause (str somevector) isn't giving me the contents ;)
16:03	hiredman	prn-str
16:04	hiredman	,(prn-str [1 2 3 4])
16:04	clojurebot	"[1 2 3 4]\n"
16:04	hiredman	,(str [1 2 3 4])
16:04	clojurebot	"[1 2 3 4]"
16:05	sopel	thanks
16:11	HAppyKAmikaze	hello world
16:12	something	was that from clojure?
16:43	cemerick	mmm, the "new" enclojure release is yummy so far
16:44	mrsolo	why isn't cons lazy by default?
16:46	technomancy	mrsolo: there's some overhead to lazy sequences
16:49	mrsolo	i take it the performance hit is very noticeable?
16:49	technomancy	mrsolo: it _would_ be noticeable if you had to invoke it for every single cons call.
16:59	cemerick	I wonder...
16:59	cemerick	,(ref 5)
17:00	clojurebot	#<Ref@47dc79: 5>
17:00	cemerick	,(ref 5 :meta {:foo 0})
17:00	clojurebot	#<Ref@11e7af3: 5>
17:00	cemerick	hrm, that throws a classcast for me...
17:01	durka42	works for me
17:03	cemerick	durka42: yeah, it's a sporadic error on my end. Not sure what to make of it.
17:44	twism	how do you check if a map contains a key?
17:44	danlarkin	(doc contains?)
17:44	clojurebot	"([coll key]); Returns true if key is present in the given collection, otherwise returns false. Note that for numerically indexed collections like vectors and Java arrays, this tests if the numeric key is within the range of indexes. 'contains?' operates constant or logarithmic time; it will not perform a linear search for a value. See also 'some'."
17:45	twism	if anyone knows that off head... im too lazy to browse the api
17:45	twism	thanks danlarkin
17:45	danlarkin	find-doc is your friend
17:46	hiredman	also, maps are Collections so they have a .contains method
17:46	hiredman	,(.contains {:a 1} 1)
17:46	clojurebot	java.lang.IllegalArgumentException: No matching method found: contains for class clojure.lang.PersistentArrayMap
17:46	twism	oh cool
17:46	hiredman	oh
17:46	hiredman	wait
17:46	hiredman	Maps are not Collections
17:46	hiredman	duh
17:46	twism	lol
17:46	hiredman	everything else is
17:46	hiredman	,(.contains [1 2] 1)
17:46	clojurebot	true
17:47	hiredman	,(.contains [0 2] 1)
17:47	clojurebot	false
17:47	twism	,(contains? {:a 1} :a)
17:47	clojurebot	true
17:47	hiredman	,({:a 1} :a)
17:47	clojurebot	1
17:48	twism	thanks guys
17:48	twism	#clojure is the best
17:48	hiredman	,(.keySet {:a 1})
17:48	clojurebot	#< [:a]>
17:48	hiredman	,((.keySet {:a 1}) :a)
17:48	clojurebot	java.lang.ClassCastException: clojure.lang.APersistentMap$2 cannot be cast to clojure.lang.IFn
17:48	hiredman	,(get (.keySet {:a 1}) :a)
17:48	clojurebot	nil
17:49	hiredman	:(
17:53	twism	hiredman: isnt that right?
17:54	hiredman	,(:a (.keySet {:a 1}))
17:54	clojurebot	nil
17:54	hiredman	,(class (.keySet {:a 1}))
17:54	clojurebot	clojure.lang.APersistentMap$2
17:54	twism	,(doc get)
17:54	clojurebot	"([map key] [map key not-found]); Returns the value mapped to key, not-found or nil if key not present."
17:54	hiredman	,(get #{:a} :a)
17:54	clojurebot	:a
17:54	twism	hmm
17:54	hiredman	,(map #(.getName %) (.getMethods (class (.keySet {:a 1}))))
17:54	clojurebot	("contains" "size" "iterator" "hashCode" "equals" "removeAll" "add" "clear" "toString" "isEmpty" "addAll" "toArray" "toArray" "remove" "containsAll" "retainAll" "wait" "wait" "wait" "getClass" "notify" "notifyAll")
17:55	hiredman	,(.contains (.keySet {:a 1}) :a)
17:55	clojurebot	true
17:55	hiredman	:(
17:58	hiredman	http://farm3.static.flickr.com/2289/2126533900_30e4e1cd4f.jpg
18:25	danlarkin	if I'm conjing into an array-map, it should remain in the order the elements get conjed on, right?
18:29	hiredman	danlarkin: until it becomes a HashMap, I believe so
18:30	danlarkin	oh, I see... this must be becoming a hash-map somewhere then
18:30	hiredman	around 8 to 10 entries
18:31	danlarkin	oh, even if I instantiate with (array-map)? I thought that would force it to remain one
18:31	hiredman	,(class (conj (array-map :a 1 :c 2 :d :e :f :g :h :i :j :k :l :m :n :o :p :q :r :s :t :u :v :w :x :y :z :z) [:b 2]))
18:31	clojurebot	clojure.lang.PersistentHashMap
18:32	danlarkin	darn
18:32	danlarkin	I want an ordered map
18:32	rsynnott	funny, really, you rarely see ordered maps
18:32	clojurebot	The most exciting phrase to hear in science, the one that heralds new discoveries, is not 'Eureka!' (I found it!) but 'That's funny ...' -- Asimov
18:33	hiredman	possibly a sorted set of keys?
18:33	rsynnott	I suppose they'd be a little expensive
18:33	hiredman	hey
18:33	hiredman	yeah, you could attach a sorted set of keys as metadata
18:34	hiredman	of course you would need to wrap conj and assoc, etc
18:34	danlarkin	yeah, no good
18:35	hiredman	I dunno that it is no good
18:35	hiredman	you just shove it in to d.lakin.sorted-map
18:35	hiredman	exclude the core versions
18:36	hiredman	(:require [d.larkin.sorted-map :as sm])
18:36	hiredman	sm/assoc etc etc
18:36	danlarkin	:-/
18:37	danlarkin	that's too much
18:38	danlarkin	it's a library, I can't really make people use dan.larkin.ordered-map/get for one map returned by one function
18:38	danlarkin	I wouldn't want to, even
18:46	hlship	working on a macro to define a function in another namespace
18:46	hlship	so as to have access to that namespace's private defs
18:46	hlship	context: using Clojure as a property expression language inside a template file (for cascade)
18:46	hlship	http://paste.lisp.org/display/80573
18:47	hlship	doesn't seem to work, however
18:47	hlship	user=> (load-file "/Users/Howard/work/clojure/cascade/src/test/clojure/com/howardlewisship/cascade/test_utils.clj") java.lang.Exception: Unable to resolve symbol: double-talk in this context (test_utils.clj:23)
18:47	hlship	It's as if the (in-ns) doesn't work.
18:48	technomancy	wasn't there a "with-ns" added that temporarily changes the namespace?
18:48	hiredman	hlship: the namespace that def uses is determined durring compile time
18:48	hiredman	so you need to use intern
18:48	hiredman	(doc intern)
18:48	clojurebot	"([ns name] [ns name val]); Finds or creates a var named by the symbol name in the namespace ns (which can be a symbol or a namespace), setting its root binding to val if supplied. The namespace must exist. The var will adopt any metadata from the name symbol. Returns the var."
18:52	hlship	thanks for the tip, but I'm not seeing what exactly should be "intern"-ed
18:53	hiredman	hlship: the function
18:53	hlship	So, use (intern) rather than (defn)?
18:54	hiredman	some like (intern (create-ns "some.namespace") (fn [] "foo"))
18:54	hiredman	,(doc create-ns)
18:54	clojurebot	"([sym]); Create a new namespace named by the symbol if one doesn't already exist, returns it or the already-existing namespace of the same name."
18:55	hlship	can you point me at an example of someone doing this? It looks close to what I want anyway, as the final impl will use anonymous or inline functions
18:55	hiredman	lemma take a look at your paste
18:57	hiredman	hlship: so you are passing in something like (defn foo [] "foo") ?
19:01	hiredman	because intern explicitly takes a namespace argument you wouldn't have to deal with in-ns back and forth
19:11	mrsolo	hm clojurecheck, does it work?
19:20	Chousuke	hiredman: you can use find-ns :p
19:21	hiredman	Chousuke: sure sure
19:22	hiredman	but create-ns acts just like find-ns, but it will create the ns if it does not exist
19:22	mrsolo	oh. well quickcheck aspet isn't implemented yet..
19:54	_hrrld	I'm trying to use print and read to serialize a vector of strings and I'm getting owned:
19:54	_hrrld	(with-in-str (with-out-str (print ["a b c" "d e f"])) (read))
19:55	durka42	,(pr-str ["a b c" "d e f"])
19:55	clojurebot	"[\"a b c\" \"d e f\"]"
19:55	durka42	,(read-string (pr-str ["a b c" "d e f"]))
19:55	clojurebot	["a b c" "d e f"]
19:56	_hrrld	n1, thanks.
20:01	mrsolo	I am seeing something like this in core.clj.. what does it mean? [#^clojure.lang.IObj x m]
20:01	mrsolo	type casting?
20:01	ataggart	type hint i think
20:02	ataggart	to avoid reflection
20:02	mrsolo	ah thanks
20:03	mrsolo	another question. what is (fn* ....)?
20:03	mrsolo	(fn* let [& decl] (cons 'let* decl))) for example
20:04	ataggart	dunno
20:04	ataggart	I dont see it being declared in any of the clj
20:04	mrsolo	core.clj
20:05	mrsolo	being used that is
20:05	ataggart	ya, I means fn* itself being declared. Given how high it is in clore.clj, I'd imagine it's in the java side
20:05	ataggart	in the processing
20:06	ataggart	erm.. reading
20:06	ataggart	just a guess tho
20:07	ataggart	ya, check out Compiler.java
20:07	ataggart	it's a symbol created amongst all th eother special forms
20:09	mrsolo	oic.. thx
20:59	beelee	,(-> 2 (fn [x] (+ 1 x)))
20:59	clojurebot	java.lang.RuntimeException: java.lang.RuntimeException: java.lang.UnsupportedOperationException: nth not supported on this type: Integer
21:00	beelee	but if i bind that fn to something it works?
21:03	beelee	,(let [f (fn [x] (+ 1 x))] (-> 2 f))
21:03	clojurebot	3
21:04	dysinger	,(System/exit 1)
21:04	clojurebot	java.security.AccessControlException: access denied (java.lang.RuntimePermission exitVM.1)
21:04	dysinger	:)
21:05	durka42	beelee: -> is translating the first form into (fn 2 [x] (+ 1 x))
21:06	durka42	,(-> 2 ((fn [x] (+ 1 x))))
21:06	clojurebot	3
21:06	beelee	hmm, i see what you mean
22:47	danlarkin	rhickey: any thoughts on adding an ordered-map to clojure? I have a use case that requires one
22:56	rhickey	danlarkin: there is a sorted map
22:57	danlarkin	I want to sort by insertion order
22:59	rhickey	then it's going to be a slow map
22:59	rhickey	might as well just search for things in a vector
23:30	sshack	quick question. what do the []'s do?
23:31	hiredman	vector literal
23:31	hiredman	,(class [])
23:31	clojurebot	clojure.lang.PersistentVector
23:31	sshack	ahh. Okay, I see now.