#clojure – 2010-08-15

2010-08-15

04:00	LauJensen	Good morning all
08:00	rdsr	Can I type hint the protocol methods so that the methods in generated interface does not have Object as the argument type?
08:42	Bahman	Hi all!
08:50	no_mind	hi
08:54	emh	rdsr: I googled for that earlier this week, found http://www.listware.net/201006/clojure/44571-protocol-type-hints-java-interface.html
08:54	rdsr	emh, oh cool thks
08:55	rdsr	emh, I tried adding type hints but the generated interface contains only objects as types
08:57	emh	rdsr: yes, according to the thread seems like you have to use definterface to get non-Object types
08:58	rdsr	thks emh
09:39	Licenser	cookies?
09:40	Licenser	woooh it's working!
09:41	Licenser	now that I can talk again I wanted to say it's amazing how well clojure works with jpcap
09:51	defn	Licenser: Did you give me permission to port lein-search to cake?
09:52	Licenser	defn: it is open source :) you don't need my permission
09:52	defn	It's polite to ask :)
09:52	defn	Also, does anyone know if java-utils was renamed for 1.2.0?
09:52	defn	is it just .java now?
09:52	defn	specifically im looking for (as-file ...)
09:53	Licenser	defn: heh
09:53	defn	I keep forgetting which were changed and to what
09:57	Licenser	wow my knowledge about TCP IP is rusty :(
09:57	defn	Licenser: im sure you'll pick it up-
09:57	Licenser	defn: it can take a while, it's non trivial sadly
09:58	Licenser	and too well documented
09:59	defn	Licenser: know anything about protocol buffers?
09:59	Licenser	the google protobuf ting?
09:59	defn	yeah
09:59	defn	im curious why it's so efficient
09:59	defn	it just looks like kv pairs
10:00	Licenser	defn: because it's binary compiled
10:01	defn	Licenser: ah -- i found an interesting clojure library that allowed you to do them easily in clojure as your serialization format
10:01	Licenser	nods
10:01	Licenser	I know
10:01	Licenser	captured with clojure: http://gist.github.com/525514
10:02	defn	ha! interesting coincidence
10:04	Licenser	now I wonder how can I make out the TCP stream from that :(
10:21	gfrlog	can an agent ever have more than one error?
10:30	gfrlog	,(* 3 3)
10:30	clojurebot	9
10:51	Licenser	awsome I can follow TCP Streams now :D
10:57	technomancy	defn: as-file is just clojure.java.io/file now
10:59	LauJensen	Licenser: What are you working on ?
10:59	Licenser	LauJensen: jpcap & clojure
10:59	LauJensen	Yes - I mean from a highlevel POV - What kind of applicaton ?
10:59	LauJensen	s/applicaton/application/
10:59	sexpbot	Yes - I mean from a highlevel POV - What kind of application ?
11:01	Licenser	LauJensen: oh well I want some network debugging
11:01	LauJensen	Ok
11:01	Licenser	think of a wireshark specialized for my usual problems
11:01	Licenser	also I'm still in the testing stage :)
11:02	LauJensen	Wireshark has been good to me in the past, would be interesting if you came up with something like it
11:02	Licenser	LauJensen: it has been to me too, for now I want a somewhat compfy lib to handle basic cases :P
11:02	LauJensen	Can you give some examples of what you're targeting?
11:05	Licenser	(backgrond-capture-streams streams open-en1) => streams keeps an up to date list of TCP streams on en1
11:05	Licenser	where open-en1 is a capture interface I opened and strems is an (atom {})
11:06	Licenser	in general I hope for a tool to debug SIP, since it's part of my day job and it'd be way cooler then whireshark :P
11:07	LauJensen	Ok
11:08	Licenser	but that is a long run thing :P for the moment I am glad to play with package capturing ^^
11:08	Licenser	since I've my doubt that I'll ever manage to finish the entire project
11:14	LauJensen	Yea I understand- Its amazing how fast you can get work done, once you get paid for it
11:32	Licenser	LauJensen: yea my problem is I ain't :P
11:32	LauJensen	Yea I figured
11:33	Licenser	which is quite sad since I bet we could use a program like that
11:35	defn	technomancy: thanks
11:35	defn	technomancy: is this documented anywhere besides notes on a commit?
11:36	technomancy	defn: there was a mailing list thread about clojure.io
11:37	defn	technomancy: oh :X I must admit the mailing list is not the best way for me personally to follow new developments
11:39	technomancy	there may be more somewhere else
11:42	technomancy	maybe the 1.2 changelog?
11:56	bobbytek	what's the best ide for clojure?
11:56	bobbytek	counterclockwise?
11:56	clojurebot	Counterclockwise aka ccw at http://code.google.com/p/counterclockwise/
12:00	Licenser	LauJensen: out of curiosity, where does your interest come form?
12:10	kensho	what is the best way to mutate a java graphics context in a thread safe way in clojure? i.e. I want to call setColor and drawRect atomically from multiple threads.
12:13	kensho	Is this possible without locks?
12:14	Licenser	kensho: I guess using agents won't be wrong
12:15	kensho	Licenser: thanks, I'll look into it
13:41	Licenser	in the slime repl what was the specal variable for the last result?
13:41	qbg	*1
13:42	Licenser	thanks qbg
13:55	LauJensen	Licenser: my interest?
13:55	Licenser	LauJensen: in package capturing
13:56	LauJensen	I was just trying to imagine what kind of project you could be working on. Last time I used it was to reverse engineer a proprietary communications protocol
14:01	LauJensen	~seen etate
14:01	clojurebot	no, I have not seen etate
14:01	LauJensen	$seen etate
14:01	sexpbot	I have never seen etate.
14:01	Licenser	nods
14:02	Licenser	LauJensen: ah :) well clojure works great for this since it's nice for stuff
14:03	Licenser	wow that was a very non saying sentence :) I ment all the data mangling is really great
14:04	LauJensen	haha
14:05	LauJensen	Yea - Dude, I want a t-shirt with that printed
14:05	LauJensen	"since its nice for stuff"
14:05	LauJensen	Clojure is indeed, nice for stuff :D
14:05	Licenser	yes it is
14:44	LauJensen	ATTENTION PRODUCITIVTY FREAKS: Im about to wrap up my blogposts on productivity and Im looking to share some Emacs bindings that help speed things up. If any of you have some really good bindings would you mind hitting me with an msg or sending me an email with your 10 favorite bindings ?
14:48	gfrlog	Ctrl-X Ctrl-C vim
14:48	LauJensen	gfrlog: In my company that would get you fired, so dont joke like that :)
14:48	gfrlog	:-D
14:49	qbg	M-x butterfly
14:49	LauJensen	yea, butterfly is good
14:49	qbg	It flips the bits for me!
14:49	LauJensen	qbg: yea, like a bigger M-x zone-mode
14:50	qbg	M-x tetris makes you very productive ;)
14:50	LauJensen	Both of you guys scroll up, I said produtivity freaks - you sound like slackers :)
14:51	qbg	It makes you very productive at slacking off. Slacking off is hard work, after all...
14:55	Licenser	is there something like re-split?
14:56	Licenser	I totally fail to find that :(
14:56	gfrlog	clojure.contrib.str-utils2/split looks promising
14:56	gfrlog	,(doc clojure.contrib.str-utils2/split)
14:56	clojurebot	Pardon?
14:56	Licenser	gfrlog: thank you!
14:57	qbg	clojure.string/split
14:57	gfrlog	guess I'm not up to date on the latest organization
14:57	gfrlog	neither is the main site apparently
16:04	gfrlog	,(nil)
16:04	clojurebot	java.lang.IllegalArgumentException: Can't call nil
16:06	alpheus	There's a follow-redirects in clojure-http-client ("1.1.0-SNAPSHOT"). How do I access it in my code?
16:07	Licenser	I think (set! follow-redirects true) or even better (binding [ follow-redirects true] ...your code...)
16:07	alpheus	I've tried binding, but I get an exception with the message "No such var: clojure-http.client/follow-redirects"
16:07	Licenser	then your intel might be outdated?
16:08	alpheus	Intel? I am looking at source with the save version number. But maybe the jar that lein downloaded doesn't match the source?
16:08	Licenser	hmm hmmm
16:08	alpheus	s/save/same
16:09	LauJensen	alpheus: (def http-agent-defaults (assoc http-agents-default :follow-redirects false)) I think.
16:09	LauJensen	http://github.com/clojure/clojure-contrib/blob/master/src/main/clojure/clojure/contrib/http/agent.clj#L166
16:11	alpheus	Actually, after (ns clojure-http.client), evaluating follow-redirects still causes an exception. So the source I'm reading must really not match the jar, I think.
16:23	LauJensen	Do we have something in Clojure-land for producing guids ?
16:24	alpheus	dunno about clojure, but if you happen to be using Linux, there's /proc/sys/kernel/random/uuid -- just read a line from int
16:26	alpheus	AFAICT, clojure-http.client doesn't use http/agent, but just uses HttpURLConnection directly.
16:28	LauJensen	,(slurp "/proc/sys/kernel/random/uuid")
16:28	clojurebot	java.security.AccessControlException: access denied (java.io.FilePermission /proc/sys/kernel/random/uuid read)
16:29	LauJensen	tsk
16:29	LauJensen	Hard to get an easier UUID than that
16:30	alpheus	Yes. I replaced a few hundred lines of Ruby code with that one-liner.
16:30	LauJensen	hehe
16:30	LauJensen	ouch
16:47	alpheus	OK, the clojure-http.client jar and source were definitely out of sync. Proved it by loading client.clj by hand.
16:48	alpheus	So I built a jar called clojure-http-client-1.1.0-SNAPSHOT.jar. How do I replace the jar called clojure-http-client-1.1.0-20100502.112537-2.jar (that's apparently out-of-date) in my project?
16:53	gfrlog	alpheus: are you using leiningen?
16:53	alpheus	yes
16:53	gfrlog	when I went to clojars I saw two versions of that library
16:54	gfrlog	I have [org.clojars.raynes/clojure-http-client "1.1.0-SNAPSHOT"] in my project.clj
16:54	gfrlog	which was the newer of the two
16:54	gfrlog	and included the namespace switch mentioned on the project page
16:54	gfrlog	maybe that's good enough for you?
16:55	alpheus	Now it's starting to make more sense.
16:56	alpheus	As an experiment, I copied the jar I just built into the lib/ directory, and removed the other one. Everything's now working exactly as I'd hoped.
16:57	gfrlog	would that survive a "lein clean"?
16:57	alpheus	I obviously need to invest some time in learning how leiningen and clojars work
16:57	gfrlog	or cloning the project into a new directory?
16:57	alpheus	no, not at all. it's just a quick sanity check.
16:59	gfrlog	okeedoke
16:59	gfrlog	leiningen is pretty simple I think
17:00	alpheus	I'm a total newb. This is my firt attempt at a clojure program.
17:00	alpheus	first*
17:00	gfrlog	are you familiar with build tools in other languages?
17:01	alpheus	sure, from Common Lisp, and make. I've looked at ant
17:21	ender403	test
17:22	kencausey	ender403: pong
17:23	kotrin	ender403: test
17:23	ender403	hello all. i'm trying to use a java library that require me to define a class with some primitive fields and a java annotation per field
17:23	ender403	i tried doing this using deftype but I also need a default constructor that takes no arguments
17:26	ender403	should I use gen-class to create a class for thsi type?
17:26	ender403	hi ken, kotrin
17:43	hoeck	ender403: right, with gen-class you can define custom ctors
17:45	ender403	hoeck, thanks. what is the syntax to add fields using gen-class?
17:45	hoeck	ender403: you can have only one single field :/
17:46	hoeck	its usually called state and contains a ref
17:47	hoeck	so using gen-class won't work in your case
17:48	gfrlog	is there a native closure function for getting the current time?
17:48	ender403	i'm trying to defining a class such that it will work to serialize in binary data using the preon library (http://preon.sourceforge.net/)
17:50	ender403	preon requires that I declare a simple java class with some annotations. like this class Rectangle {
17:50	ender403	@Bound private int x1;
17:50	ender403	@Bound private int y1;
17:50	ender403	@Bound private int x2;
17:50	ender403	@Bound private int y2;
17:50	ender403	}
17:50	ender403	i'd like to create a class exactly like this using clojure if possible
17:55	hoeck	sounds (currently) impossible to me doing this in clojure
17:56	hoeck	genclass is missing the ability to have mutable fields, and deftype defrecord do not have no-ar ctors
18:00	ender403	i read that rich doesn't want to make deftype into an all purpose java interop feature so deftype probably won't allow custom constructors in future
18:01	ender403	isn't gen-class supposed to be a java interop feature?
18:02	hoeck	imo a right decision, but gen-class should then support mutable or mutable-volatile fields
18:02	kencausey	It sounds to me like this is just a case where it makes more sense to implement that class in Java and move on
18:02	kencausey	If it's simple and requires a lot of Java-isms, why do otherwise?
18:03	hoeck	of course, and you have to compile it anyways, so little gain from writing it in clojure
18:06	ender403	it would be nice to have a macro that can generate the class in clojure. we can then have nice binary serialization like Erlang. But if there is no way current then I can write it in java
18:07	ender403	preon library looks very nice. has anyone here had experience with it?
18:08	ender403	brb. making pasta
18:15	ender403	hoeck, i'm still new to clojure. do you mean there is little to gain because the class is created at compile time instead of runtime?
18:19	hoeck	ender403: I just assume that because most java-serialization libs use named classes wich must be created by compiling a gen-class namespace in clojure
18:48	technomancy	Leiningen repl now listens on a socket as a bonus: http://github.com/technomancy/leiningen/commit/4f64c64afb455c3494873d469729b4a89f7f9f29
18:57	technomancy	I hate how long that function is, but building up a form for eval-in-project makes it harder to split out functionality. =\
18:59	technomancy	not on my watch =)
19:05	vIkSiT	lo all
19:06	vIkSiT	anyone using mongodb here?
19:06	vIkSiT	I was wondering if there was a way to use congomongo to do range queries on a collection.
19:07	vIkSiT	(For instance, if a collection has a 100 rows, is there a way to retrieve the first 10, the next 10, and so on)
19:09	phaer	What does "(~ var)" in clojure do? Or do you know whats the name of "~" in english, so i can look for it?
19:09	vIkSiT	the tilde operator?
19:09	vIkSiT	One use is when creating macros
19:10	gfrlog	is there another use?
19:10	technomancy	I think it's called unquote
19:10	vIkSiT	(def mymacro [n] `(apply str "this is" ~n))
19:10	vIkSiT	gfrlog, I'm not sure I know enough about Clojure to comment if there was ONLY one use :)
19:10	phaer	ah, makes sense. thank you
19:11	pdk	fkin hate how you cant even change your nick without leaving some channels on freenode
19:11	pdk	anyway ~ is called tilde
19:11	pdk	and in clojure it's used in macros to tell it to eval a part of a quoted form iirc
19:11	pdk	we're rebellious so we dont do cl style ,
19:13	phaer	pdk: Yes, but thats a plus point for clojure. The use of different kind of brackets for example makes it much more readable than common lisp in my opinion. :)
19:13	pdk	it's not so much a bracket difference here as just using a different symbol for the same function from cl
19:14	pdk	since people already use , as a reading aid in map literals
19:14	vIkSiT	hmm, so what would be the best way to iterate over a large collection of elements, taking n elements at a time?
19:15	pdk	(partition n n list)?
19:15	vIkSiT	pdk, would that be lazy?
19:15	pdk	(doc partition)
19:15	clojurebot	"([n coll] [n step coll] [n step pad coll]); Returns a lazy sequence of lists of n items each, at offsets step apart. If step is not supplied, defaults to n, i.e. the partitions do not overlap. If a pad collection is supplied, use its elements as necessary to complete last partition upto n items. In case there are not enough padding elements, return a partition with less than n items."
19:15	vIkSiT	aah
19:15	vIkSiT	thanks for the info
19:15	pdk	(partition 3 (take 15 (iterate inc 1)))
19:15	pdk	,(partition 3 (take 15 (iterate inc 1)))
19:15	clojurebot	((1 2 3) (4 5 6) (7 8 9) (10 11 12) (13 14 15))
19:17	vIkSiT	awesome
19:18	vIkSiT	pdk, so assuming I've got a list of elements whose total I do not know of.
19:18	pdk	total as in number of items in the list or sum of all numerical items in the list
19:18	vIkSiT	hmm. the former
19:18	pdk	partition doesnt require the list it takes in to be a clean multiple of n in terms of length
19:19	pdk	,(partition 4 (take 15 (iterate inc 1)))
19:19	clojurebot	((1 2 3 4) (5 6 7 8) (9 10 11 12))
19:19	pdk	,(partition 3 (take 15 (iterate inc 1)))
19:19	clojurebot	((1 2 3) (4 5 6) (7 8 9) (10 11 12) (13 14 15))
19:19	vIkSiT	aah indeed
19:19	pdk	MAYBE
19:19	pdk	,(partition 4 4 (take 15 (iterate inc 1)))
19:19	clojurebot	((1 2 3 4) (5 6 7 8) (9 10 11 12))
19:19	pdk	clojurebot dont make this difficult
19:20	pdk	,(partition 4 1 (take 15 (iterate inc 1)))
19:20	clojurebot	((1 2 3 4) (2 3 4 5) (3 4 5 6) (4 5 6 7) (5 6 7 8) (6 7 8 9) (7 8 9 10) (8 9 10 11) (9 10 11 12) (10 11 12 13) (11 12 13 14) (12 13 14 15))
19:20	pdk	obvious way would probably be to do a step of 1 then grab every nth item and the last item i guess
19:20	vIkSiT	ah so the 1 there is an overlap step
19:20	pdk	this could use some more looking into
19:21	pdk	yeah the 3 argument version uses the second argument to control how far to advance between steps
19:21	pdk	,(partition 4 1 (range 1 9))
19:21	clojurebot	((1 2 3 4) (2 3 4 5) (3 4 5 6) (4 5 6 7) (5 6 7 8))
19:21	pdk	see how it steps with a step of 1
19:21	pdk	,(partition 4 2 (range 1 9))
19:21	clojurebot	((1 2 3 4) (3 4 5 6) (5 6 7 8))
19:21	vIkSiT	right
19:22	pdk	so partition 4 2 sort of makes half of each list overlap
19:22	vIkSiT	right
19:24	pdk	would need to know a few more details about what we want for this problem before deciding if partition's cutting-off behavior would be an issue
19:25	vIkSiT	well, in my case, I'm trying to compensate for the fact that my database won't let me do ranged queries on my data. So, I'm using a lazy "fetch" to get the entire dataset
19:25	pdk	cause apparently the 2 argument version if the length of the list isn't a multiple of n it'll just drop the excess
19:25	vIkSiT	and then partitioning it into manageable chunks, map a function on it, and finally dump it. so the exact partition doesn't make a difference ot me
19:26	vIkSiT	pdk, yes, the overlap won't be a problem
19:26	pdk	hmm not quite sure if partition would do the job on its own without the 3 arg version or some extra trickery to make it not drop stuff in the 2 arg
19:26	pdk	in that case 3 arg version with a step 1 seems to preserve all the data when the list length isn't a multiple of n
19:26	vIkSiT	pdk, not sure I understood your last statement
19:26	pdk	like when i tried to partition a 15 element list with n = 4 earlier
19:26	vIkSiT	right
19:26	pdk	in the 2 arg version of partition
19:27	pdk	note how it retrieved 12 items then just dropped the last 3 cause it couldn't make a 4 element list out of them
19:27	vIkSiT	aah.
19:27	vIkSiT	,(partition 4 4 (take 15 (iterate inc 1)))
19:27	clojurebot	((1 2 3 4) (5 6 7 8) (9 10 11 12))
19:27	raek	I think you can pass nil as one of the args to solve that
19:27	vIkSiT	right
19:27	vIkSiT	,(partition 4 nil (take 15 (iterate inc 1)))
19:27	clojurebot	java.lang.RuntimeException: java.lang.NullPointerException
19:27	pdk	oh yeah there's a 4 arg version
19:27	pdk	with the 3rd arg as padding
19:28	pdk	(doc partition)
19:28	clojurebot	"([n coll] [n step coll] [n step pad coll]); Returns a lazy sequence of lists of n items each, at offsets step apart. If step is not supplied, defaults to n, i.e. the partitions do not overlap. If a pad collection is supplied, use its elements as necessary to complete last partition upto n items. In case there are not enough padding elements, return a partition with less than n items."
19:28	pdk	the 3rd arg has to be a sequence
19:28	pdk	you could do either (repeat nil) or (repeatedly nil), i forget which
19:28	pdk	(doc repeat)
19:28	clojurebot	"([x] [n x]); Returns a lazy (infinite!, or length n if supplied) sequence of xs."
19:28	pdk	(doc repeatedly)
19:28	clojurebot	"([f] [n f]); Takes a function of no args, presumably with side effects, and returns an infinite (or length n if supplied) lazy sequence of calls to it"
19:28	pdk	yeah repeat
19:28	raek	yeah, pad = nil will make the last sequence be the rest of the elements
19:29	pdk	(partition 4 4 nil (range 1 16))
19:29	pdk	,(partition 4 4 nil (range 1 16))
19:29	clojurebot	((1 2 3 4) (5 6 7 8) (9 10 11 12) (13 14 15))
19:29	vIkSiT	hmm
19:29	pdk	,(partition 4 4 nil (range 1 15))
19:29	clojurebot	((1 2 3 4) (5 6 7 8) (9 10 11 12) (13 14))
19:29	vIkSiT	aah the third argument
19:29	pdk	!!!
19:29	pdk	that makes it preserve everything without overlap
19:29	vIkSiT	,(partition 4 4 nil (take 15 (iterate inc 1)))
19:29	clojurebot	((1 2 3 4) (5 6 7 8) (9 10 11 12) (13 14 15))
19:29	vIkSiT	ah that works
19:29	pdk	oughta take note of that little trick
19:30	vIkSiT	indeed
19:30	vIkSiT	thats pretty neat
19:30	pdk	though still dumb that it doesnt document what happens when you use nil as 3rd arg in 4 arg version
19:30	pdk	at least not explicity
19:30	pdk	cause first time you read it say "pad collection" you assume it has to be an actual collection
19:31	raek	well, I guess you could use () as padding too
19:31	raek	but yeah... it's a common usage that is not very obvious
19:32	pdk	(partition 4 4 () (range 1 15))
19:32	pdk	,(partition 4 4 () (range 1 15))
19:32	clojurebot	((1 2 3 4) (5 6 7 8) (9 10 11 12) (13 14))
19:42	chouser	there was a pretty extensive discussion about changing how partition with padding works
19:42	chouser	but I guess it didn't end up going anywhere
19:43	pdk	what was the gist of the proposed change
19:43	chouser	I think the final statements were feeling out whether the current partition-with-pad was perhaps mixing two things into a single function that ought to be separated.
19:44	chouser	I'm sure it's on the google group or IRC log. Not sure which.
19:47	chouser	hm, or maybe assembla
19:49	chouser	pdk: https://www.assembla.com/spaces/clojure/tickets/120
20:05	ender403	hi chouser
20:07	ender403	is it possible to declare the following java class using gen-class or deftype? class Rectangle {
20:07	ender403	@Bound private int x1;
20:07	ender403	@Bound private int y1;
20:07	ender403	@Bound private int x2;
20:07	ender403	@Bound private int y2;
20:07	ender403	}
20:09	chouser	Is @Bound some sort of annotation?
20:11	chouser	hm, I don't think either will do private fields.
20:12	ender403	yes Bound is an annotation provided by the preon library
20:12	ender403	it doesn't have to be private field. When I tried defining this type using deftype I couldn't get a default constructor which takes no arguments. preon needs this
20:13	ender403	preon overview is here http://www.scribd.com/doc/8128172/Preon-Introduction
20:13	chouser	ugh. I don't think so.
20:14	chouser	deftype only provides a ctor with one arg per field, and no way to provide more ctors
20:14	chouser	gen-class only allows zero or one fields, no more.
20:15	ender403	hmm, is it possible to embed java code inside clojure code?
20:15	raek	maybe it's just simpler to write that class int java, and interact with it using clojure's usual java interop
20:16	chouser	no, but you can define your little class in a .java file, then use that as needed from clojure
20:16	raek	ender403: there is, for example, a lein-javac plugin for leiningen that allows you to have both clojure and java sources in the same project
20:17	ender403	ok. i was hoping to do it in clojure so I can get nice binary serialization like Erlang has
20:17	raek	iirc, the build tool Cake has support for java sources out of the box
20:19	ender403	thanks raek. lein-javac looks interesting
20:27	gfrlog	how do you handle circular definitions in clojure? like if functions a and b want to refer to each other
20:28	gfrlog	do we vacuously pre-def one of them C++ style?
20:28	gfrlog	then re-def it later?
20:29	raek	yes.
20:29	gfrlog	okeedoke
20:29	raek	you can simply do a (def a)
20:30	gfrlog	thanks
20:30	spewn	Isn't the preferred way to do it with declare?
20:30	raek	or it you want to create multiple variables: (declare a b)
20:30	phaer	raek: thanks from me too. lein-javac seems to be exactly what i was looking for at the moment :)
20:30	gfrlog	I learned a new word
20:31	raek	spewn, I think declare signals the intentions more clearly, so I guess so
20:35	gfrlog	I can swap an atom inside an agent function, right?
20:35	raek	sure
20:37	dreish	I had slime working just yesterday, and now, restarting after my Mac crashed, slime-connect reports success but does not create a buffer for the REPL.
20:37	raek	but I have no idea of what makes most sense in your situation
20:37	raek	dreish: is there a repl in the inferior-lisp buffer?
20:38	dreish	raek: The only buffers currently open are scratch, Messages, and slime-events
20:38	tomoj	dreish: do you have slime-repl installed?
20:39	dreish	tomoj: That might be it. I think I moved some things around in my .emacs file and now my slime/contrib dir is no longer in the load path.
20:39	dreish	I had to symlink a couple of slime things just to get emacs to start at all, and I missed slime-repl.el.
20:41	raek	I finally surrendered to ELPA
20:42	raek	and installing slime though it wasn't that bad
20:44	tomoj	you get a crippled slime I guess
20:45	raek	old version?
20:46	tomoj	it doesn't contain the whole thing
20:46	tomoj	right?
20:46	tomoj	I mean, you get more stuff if you install slime yourself?
20:47	raek	I dunno... I haven't missed anything so far :)
20:47	dreish	This doesn't look good. My slime-repl clojure buffer has a CL-USER> prompt.
20:48	dreish	And I still had to load slime-repl.el manually.
20:48	raek	you used slime-connect to start it?
20:48	dreish	Yes.
20:48	tomoj	haha
20:48	dreish	It still didn't create a repl buffer.
20:48	raek	do you have clojure-mode?
20:48	dreish	Yes.
20:49	tomoj	wait, I'm confused... "it stil didn't create a repl buffer" "my slime-repl clojure buffer..."
20:49	dreish	Which I created after manually loading slime-repl.el and running M-x slime-repl
20:49	dreish	I think I need to just give up and go with ELPA.
20:49	dreish	It can't be much worse than this.
20:50	raek	I think slime-repl starts the repl itself, rather than connect to an existing
20:50	raek	what happens if you don't run slime-repl and only slime-connect?
20:50	tomoj	I don't think you should be running M-x slime-repl
20:50	dreish	It says "Hack and be merry" or whatever, but I just have the buffers I listed at 8:38.
20:50	raek	dreish: http://www.assembla.com/wiki/show/clojure/Getting_Started_with_Emacs
20:51	raek	there is some relevant info there
20:51	raek	perticularily about how slime finds out which program to start
20:52	raek	I only have to run slime-connect to get started
20:52	gfrlog	does "java.lang.Exception: Can't await in agent action" ring a bell to anybody?
20:52	dreish	That worked for me, too, last time.
20:53	raek	gfrlog: what are you trying to do? (on a more larger scale)
20:54	dreish	Oh, I already have ELPA. :-?
20:54	raek	it sounds like you are using agents in the wrong way
20:55	raek	gfrlog: can you paste the code?
20:55	gfrlog	um
20:55	gfrlog	I'm trying to have most of my code running within an agent
20:55	gfrlog	and it includes a lot of http
20:55	gfrlog	throught clojure.contrib.http.agent
20:55	gfrlog	so if there's an issue with one agent function calling another agent and waiting for the result...
20:55	raek	why in an agent? to have it in its own thread?
20:56	gfrlog	hmm
20:56	raek	you can use (future ...code...) to run something in another thread
20:56	gfrlog	well I want other threads to be able to add code to the main thread
20:57	gfrlog	I suppose imagine a main thread and some worker threads
20:57	gfrlog	where each worker thread ultimately results in an http call
20:57	gfrlog	but I'd rather the http calls be one-at-a-time
20:57	gfrlog	so I do them in the main thread
20:57	gfrlog	via the agent
20:57	raek	where does the waiting come in?
20:58	gfrlog	waiting for the result of the http
20:58	raek	that sounds more like futures
20:58	gfrlog	but I don't want the next http call to start before the last one finishes
20:58	raek	if you do a @ on a future, it will block until the values is completed
20:58	gfrlog	hmm.
20:59	raek	why not simply do the calls one after another?
20:59	raek	in one thread
20:59	gfrlog	so
20:59	gfrlog	the main thread I want to make http calls whenever one of the worker threads finishes, or 30 seconds elapses, whichever is first
21:00	raek	anway, to answe the original question: yes, agents are not allowed to wait for other agents in their action functoin
21:00	gfrlog	so I could tell him to sleep for 30 seconds, but then how does he get interrupted by the workers?
21:01	gfrlog	I'll rework it I guess
21:02	gfrlog	thanks
21:02	raek	I'm sorry... I can't come up with a simple solution right now...
21:02	gfrlog	I know, this setup is rather complex
21:03	gfrlog	I have a hard enough time coming up with good ideas even with full knowledge
21:03	raek	futures have a future-cancel function too, might be good to know
21:03	gfrlog	yep
21:03	gfrlog	I'm using futures for the worker threads actually
21:04	raek	I guess the worker agents could send a "I'm ready for work" message to the main agent or something
21:05	gfrlog	but then the main thread is still an agent
21:06	gfrlog	I wonder if I can sidestep it by just wrapping the await in a future
21:06	raek	maybe one could send an "add job" message to the main agent, which adds the job to a queue (the main agent's sate)
21:06	gfrlog	ah hah! the sidestep worked
21:07	gfrlog	so an agent can wait on a future, but not another agent
21:07	gfrlog	that seems arbitrary from my naive point of view
21:07	raek	an agent kind of is a thread
21:08	raek	and a future is a thread too
21:08	raek	so the agent is not waiting
21:08	gfrlog	I bet the clojure.contrib.http.agent library would be better rewritten with futures anyhow
21:08	raek	because another thread does the waiting
21:08	gfrlog	the agent waits for the future though
21:09	raek	...and then when a worker agent is ready, it sends a function to the main agent that will pop the next job and send it to itself
21:10	raek	anyway, agents are asynchronous in their nature
21:10	gfrlog	right
21:11	raek	but they might be the closest fit anyway, if one needs the queing facility
21:11	gfrlog	yeah, I think it works fine without that one issue
21:11	gfrlog	I'm actually only using one agent
21:11	gfrlog	the "worker threads" are all futures
21:15	spewn	How does one typically deal with having a defstruct foo and passing instances of that struct to a function as an arg named foo?
21:15	spewn	I can only think of renaming one or the other to struct-foo or a-foo, but that seems awkward.
21:16	gfrlog	,(keys {})
21:16	clojurebot	nil
21:16	gfrlog	ick
21:16	gfrlog	,(or (keys {}) [])
21:16	clojurebot	[]
21:19	gfrlog	spewn: if you were using deftype, you could have Foo and foo
21:19	spewn	gfrlog: I'll have to wait until 1.2 comes out for that, I think
21:20	gfrlog	yeah :(
21:43	pdk	how is 1.2 release looking now anyway
21:43	pdk	i see on the dl page it's at rc3
21:48	gfrlog	,(.substring (str :keyword) 1)
21:48	clojurebot	"keyword"
21:54	brweber2	gfrlog you could use (name :keyword), of course I have no idea what you are trying to do...
21:54	gfrlog	(name :youreright)
21:54	gfrlog	,(name :onemoretime)
21:54	clojurebot	"onemoretime"
21:54	gfrlog	ah ha
21:55	gfrlog	,(doc name)
21:55	clojurebot	"([x]); Returns the name String of a string, symbol or keyword."
21:55	gfrlog	good man!
21:55	gfrlog	that'll clean my code up
22:27	eckroth	how do I convert a map to a string? (str [1 2 3]) works but not (str {:a 1 :b 2})
22:28	defn	eckroth: how do you want it to look
22:28	defn	like "a 1 b 2"?
22:28	eckroth	defn: anything that can parse into a map with (read-string)
22:28	eckroth	oddly try-clojure.org halts on (read-string "{:a 1 :b 2}")
22:29	defn	Raynes: 21:28 < eckroth> oddly try-clojure.org halts on (read-string "{:a 1 :b 2}")
22:29	eckroth	defn: thanks for forwarding to Raynes :)
22:29	defn	:D
22:29	defn	eckroth: so you want to...serialize this to a file or?
22:29	eckroth	defn: exactly
22:30	defn	iirc you need to use print-dup
22:30	defn	one sec
22:30	a_strange_guy	,(print-str {:a 1, :b 2})
22:30	clojurebot	"{:a 1, :b 2}"
22:30	defn	yeah but that wont work with read-string i dont think, will it?
22:30	defn	,(read-string "{:a 1, :b 2}")
22:30	clojurebot	{:a 1, :b 2}
22:30	a_strange_guy	,(read-string (print-str {:a 1, :b 2}))
22:30	clojurebot	{:a 1, :b 2}
22:31	defn	i must be wearing orthopaedic shoes...
22:31	eckroth	a_strange_guy: a great
22:31	eckroth	a_strange_guy: ah great I mean
22:31	defn	eckroth: http://clojuredocs.org/v/2020
22:32	defn	eckroth: you want to use print-dup -- im failing to remember why that is the case right now
22:34	defn	eckroth: ah -- it's if you're serializing something in your map which is a java object:
22:34	a_strange_guy	print-dup conserves the type of maps and numbers etc
22:34	eckroth	defn: ok right
22:34	eckroth	,(binding [print-dup true] (str {:a 1 :b 2}))
22:34	clojurebot	"#=(clojure.lang.PersistentArrayMap/create {:a 1, :b 2})"
22:34	defn	,(read-string (print-str {:a (java.util.Date.), :b 1}))
22:34	clojurebot	java.lang.RuntimeException: java.lang.Exception: Unreadable form
22:35	a_strange_guy	(binding [print-dup true] +)
22:36	a_strange_guy	,(binding [print-dup true] (print +))
22:36	clojurebot	#=(clojure.core$_PLUS_. )
22:36	defn	,(read-string (binding [print-dup true] (str {:a (java.util.Date.), :b 1})))
22:36	clojurebot	java.lang.RuntimeException: java.lang.IllegalArgumentException: No method in multimethod 'print-dup' for dispatch value: class java.util.Date
22:36	defn	a_strange_guy: you're a strange guy.
22:37	a_strange_guy	xD
22:37	a_strange_guy	if you want to serialize POJOs then you have to extend the print-dup multimethod to Serializable
22:38	eckroth	there should be no expectation that "any" POJO can be serialized
22:39	a_strange_guy	neither is every datastucture printable
22:40	eckroth	a_strange_guy: such as records at the moment, right? they cannot be printed
22:40	eckroth	a_strange_guy: and multimethods I assume
22:41	a_strange_guy	,(binding [print-dup true] (print #'clojure.core/print-method))
22:41	a_strange_guy	,(binding [print-dup true] (print #'clojure.core/print-method))
22:42	clojurebot	#=(var clojure.core/print-method)
22:42	a_strange_guy	,(binding [print-dup true] (print @#'clojure.core/print-method))
22:42	clojurebot	java.lang.IllegalArgumentException: No method in multimethod 'print-dup' for dispatch value: class clojure.lang.MultiFn
22:42	eckroth	,(binding [print-dup true] (print (fn [x] (+ 1 x)))
22:42	clojurebot	EOF while reading
22:42	eckroth	oh bother
22:42	eckroth	,(binding [print-dup true] (print (fn [x] (+ 1 x)))
22:42	clojurebot	EOF while reading
22:42	eckroth	whoops again
22:43	a_strange_guy	,(binding [print-dup true] (print (fn [a] (+ a a)))
22:43	clojurebot	EOF while reading
22:43	a_strange_guy	,(binding [print-dup true] (print (fn [a] (+ a a))))
22:43	clojurebot	#=(sandbox$eval502761$fn__502762. )
22:43	eckroth	so, usless
22:44	a_strange_guy	actually it works
22:44	a_strange_guy	you can read functions
22:45	defn	it's also dangerous
22:45	a_strange_guy	unless they are closures
22:45	defn	and the sandbox would be smart to not let you just read in any old string
22:45	eckroth	a_strange_guy: not if you lose state, right? that printed function makes no sense once you restart java
22:46	a_strange_guy	toplevel fns do work actually
22:46	a_strange_guy	if you AOT
22:46	eckroth	a_strange_guy: I mean anon functions
22:47	eckroth	a_strange_guy: I don't know why print can't just represent as the same datastructure that defines the anon func (that is, '(fn [x] ...))
22:48	a_strange_guy	because clojure compiles everything
22:48	a_strange_guy	the source is gone then
22:48	defn	there's a way to pretty print an anon function fwiw
22:48	eckroth	a_strange_guy: but the datastructure remains
22:48	defn	from a string i mean
22:48	a_strange_guy	nope, no datastructure
22:49	eckroth	a_strange_guy: what is no datastructure? an anon func has no associated datastructure?
22:50	a_strange_guy	an anon func is just a singleton instance of a java class
22:50	a_strange_guy	(more or less)
22:50	eckroth	a_strange_guy: why say singleton instance? why not just instance?
22:51	defn	singleton implies anon i believe
22:51	defn	a_strange_guy: is that right?
22:51	defn	or at least in the ruby object model it does
22:52	a_strange_guy	singleton means that there is only one instance
22:53	a_strange_guy	you don't need more than one if the fn is not a closure
22:53	a_strange_guy	but I am not sure if clojure compiles that way
22:54	a_strange_guy	(letfn [(make-fn [] (fn [] nil))] (indentical? (make-fn) (make-fn)))
22:55	a_strange_guy	,(letfn [(make-fn [] (fn [] nil))] (indentical? (make-fn) (make-fn)))
22:55	clojurebot	java.lang.Exception: Unable to resolve symbol: indentical? in this context
22:55	a_strange_guy	,(letfn [(make-fn [] (fn [] nil))] (identical? (make-fn) (make-fn)))
22:55	clojurebot	false
22:55	a_strange_guy	forget singleton
22:55	defn	a_strange_guy: you're a strange guy.
22:56	defn	im sorry but i cannot resist
22:56	defn	great nick
22:56	defn	reminds me of a strange loop
22:56	a_strange_guy	from G.E.B?
22:58	defn	the author of GEB wrote a book called "I am a Strange Loop"
22:58	defn	http://en.wikipedia.org/wiki/I_Am_a_Strange_Loop
22:59	Raynes	eckroth: Aye, read-string shouldn't even been allowed. Thanks for letting me know.
22:59	a_strange_guy	damn, I have to read that
22:59	defn	a_strange_guy: :) It's fantastic.
22:59	Raynes	Oh wait.
23:00	Raynes	I was thinking of read-line.
23:00	Raynes	That is odd. It doesn't seem to be halted though.
23:00	defn	a_strange_guy: I also very much recommend "The Mind's I"
23:00	defn	by Hofstadter and Dennett
23:01	Raynes	-> (read-string "{:a 1 :b 2}")
23:01	sexpbot	=> {:a 1, :b 2}
23:01	Raynes	Odd.
23:01	defn	that could be naughty.
23:01	a_strange_guy	defn: thanks, GEB already blew my mind
23:01	Raynes	That's beside the point.
23:02	Raynes	I have no clue why it would make it freeze (if only temporarily).
23:02	a_strange_guy	this seem to be its successor ^^
23:02	Raynes	(read-string "3") works
23:02	defn	-> (read-string "#=(var foo)")
23:02	sexpbot	java.lang.Exception: EvalReader not allowed when read-eval is false.
23:03	Raynes	It only fails with maps it seems.
23:03	defn	hm, weird
23:03	Raynes	defn: Licenser isn't that stupid. ;)
23:03	defn	Raynes: haha :)
23:03	Raynes	It even works with vectors.
23:03	eckroth	Raynes: yeah, odd :)
23:04	Raynes	Licenser: When you get back, check out the backlog. This might be a job for the almighty Heinz. I don't even know where to start.
23:04	defn	-> (map class [{} [] '()])
23:04	sexpbot	=> (clojure.lang.PersistentArrayMap clojure.lang.PersistentVector clojure.lang.PersistentList$EmptyList)
23:04	Raynes	Seems like Licenser does more work on tryclojure than I do these days. ._.
23:04	defn	what's the REPL say
23:06	defn	Raynes: might wanna look at clojure.lang.RT as well
23:07	Raynes	defn: ?
23:07	defn	just thinking out loud
23:07	Raynes	I don't know what you mean.
23:08	Raynes	Or what you're implying.
23:08	Raynes	:p
23:09	defn	Raynes: no implication, simply places I would begin investigating-- read-string uses clojure.lang.RT
23:09	Raynes	Oh.
23:10	Raynes	I know what I need to do. I can find out if this is a sandbox problem, and if it is, I can just whine to Heinz.
23:10	Raynes	Luckily, try-clojure has a bit of a public API I can use without the Javascript console.
23:11	Raynes	java.lang.NullPointerException
23:12	Raynes	Well, I see the problem.
23:13	Raynes	defn: http://try-clojure.org/magics?code=(read-string%20%22{:a%201%20:b%202}%22)
23:13	Raynes	Tell me what you see.
23:13	Raynes	;)
23:14	defn	Raynes: hah
23:14	defn	Raynes: you know, i had a similar problem with walton
23:15	defn	functions which were in the sandbox which just refused to show up
23:15	Raynes	The problem is that the console does nothing but sit there if nothing is returned by it's callback.
23:15	Raynes	eckroth: It isn't locking up, the cursor is just disappearing. After you type the read-string stuff, you can move the cursor with the arrow buttons and then delete what you typed.
23:16	Raynes	But the actual reason that it isn't returning anything is the real problem.
23:16	Raynes	I'll work on that later tonight or tomorrow.
23:24	defn	google analytics is great
23:25	defn	(doc root-cause)
23:25	clojurebot	Gabh mo leithscéal?
23:25	defn	->(doc root-cause)
23:25	sexpbot	java.lang.Exception: Unable to resolve var: root-cause in this context
23:27	Raynes	->(doc clojure.stacktrace)
23:27	sexpbot	java.lang.SecurityException: Code did not pass sandbox guidelines: (#'clojure.core/print-namespace-doc)
23:27	Raynes	->(doc clojure.stacktrace/root-cause)
23:27	sexpbot	=> ------------------------- clojure.stacktrace/root-cause ([tr]) Returns the last 'cause' Throwable in a chain of Throwables. nil
23:27	defn	danke
23:27	defn	(read-string (read-string
23:27	defn	grr
23:27	defn	-> (read-string (read-string "{:a 1 :b 2}"))
23:27	sexpbot	java.lang.ClassCastException: clojure.lang.PersistentArrayMap cannot be cast to java.lang.String
23:41	no_mind	are there any plans to get rid of JVM from clojure in near or distant future ?
23:43	Raynes	No. Why would there be?
23:43	Raynes	:o
23:43	Raynes	Clojure in Clojure is one current focus, and it'll make it easier for their to be other implementations of Clojure, but as of now, the JVM is the primary platform for Clojure.
23:54	mmarczyk	hey guys, I've just done a write-up on Compojure... I'm not (at all) a web programmer, but I am becoming interested in the field, so it was an interesting exercise... anyway: any comments?
23:54	mmarczyk	http://stackoverflow.com/questions/3488353/whats-the-big-idea-behind-compojure-routes/3490479#3490479
23:54	mmarczyk	hi Raynes :-)
23:56	Raynes	mmarczyk: Hai.
23:58	Raynes	mmarczyk: I want to read that, but I don't know if I have the energy.
23:58	Raynes	;P
23:58	mmarczyk	oh come on :-P
23:59	mmarczyk	on a more serious note, I'd love to find out whether the general picture is correct and what I'm missing...
23:59	mmarczyk	no hurry, though ;-)
23:59	Raynes	Actually, this is something that might help me understand Compojure better.