#clojure – 2012-11-01

2012-11-01

00:52	brainproxy	&(clojure.set/union [1] '(2))
00:52	lazybot	⇒ [1 2]
00:52	brainproxy	dynamic typing ftw!
00:52	brainproxy	:p
00:58	devn	zwat
01:08	Frozenlo`	Oh wow... I just discovered the function 'bean'
01:10	Frozenlo`	Is there a function to stringify every keyword in a map?
01:18	Frozenlo`	clojure.walk/stringify-keys
02:03	Gmind	#join #ruby
02:04	mrowe	lol
02:04	Gmind	Oops..
02:14	hughfdjackson	I love that lein creates a test folder by default
02:14	hughfdjackson	it's refreshing to have the defacto tool hint that that's just the right thing to do
02:29	Gmind	Oh, this channel is so crowd
03:15	wastrel	:[
03:28	jyu	hello I have a little question, how can i give the parameter a name when calling a function.
03:32	foodoo	jyu: Clojure doesn't support keyword arguments in general. But you can use destructuring with a hash-map for your own functions to support it
03:33	jyu	foodoo: thanks
03:42	foodoo	jyu: (defn foo [ & args ] (let [{:keys [a b c]} (apply hash-map args)] (list a b c)))
03:42	foodoo	jyu: usage example: (foo :b 2 :c 3 :a 1)
03:43	_ulises	foodoo: you can even (defn f [{:keys [a b c] }] (+ a b c))
03:43	_ulises	(if I recall correctly)
03:44	_ulises	,(defn f [{:keys [a b c] }] (+ a b c))
03:44	clojurebot	#<Exception java.lang.Exception: SANBOX DENIED>
03:47	_ulises	actually it'd be (defn f [& {:keys [a b c]}] (+ a b c))
03:48	_ulises	you can even then add an :or { ... } bit to it
03:49	jyu	_ulises: :) thanks got it
03:57	foodoo	_ulises: oh, cool. Thanks
04:12	dspp	i'm using enlive to do some scraping, im ending up with \ characters in the resulting strings
04:12	dspp	can't strip this out with trim, its just a slash \
04:12	dspp	why is this, how can i strip these out?
04:23	dspp	Apage43, thanks for the tip
04:24	dspp	I'd still like to know what this weird \ character is though
05:34	luxbock	using Emacs with paredit, is there a shortcut for changing parenthesis into square-brackets?
05:35	ejackson	luxbock: don't think so just splice and then slurp
05:35	luxbock	I typed my function call with [ ] by accident
05:35	luxbock	alright
05:52	ejackson	luxbock: i guess doing it the other way around would be faster and less error prone
06:13	wei_	get-in seems to have different behavior for clojure and clojurescript: (get-in [1 2 3] [nil] "not found") yields "not found" in clojure but "1" in clojurescript
06:13	wei_	is there any way to also get the "not found" result in clojurescript for nil reference into a vector?
06:19	wei_	ok, my workaround is (if (nil? index) "not found" (get-in [1 2 3] [index] "not found"))
07:56	mpan	Is there something which brings in another ns under a new name, similar to refer? The ns already exists and I wouldn't want to try loading it, because it was created in a different way.
07:57	mpan	Would require do it? That seems confusing because the usual usage of require is to cause loading, right?
08:57	xeqi	&(doc alias)
08:57	lazybot	⇒ ------------------------- clojure.core/alias ([alias namespace-sym]) Add an alias in the current namespace to another namespace. Arguments are two symbols: the alias to be used, and the symbolic name of the target namespace. Use :as in the ns macro in preference to calling this directly. nil
09:04	XOXO1	hey, I need some kind of basic web server in clojure (using noir?) which would accept POST requests with file uploads and it just needs to response that okay, file was uploaded
09:30	S11001001	lazybot: mail XOXO1 if all you're doing is accepting multipart/form-data, that doesn't really demand more than basic https://github.com/ring-clojure/ring capabilities
09:30	lazybot	Message saved.
09:30	pandeiro	does anyone else have problems when killing an nrepl server/client in emacs and then trying to restart? i get error in process filter: error during connect: connection refused
09:49	ambrosebs	Who would be pleased if my Conj talk was 40 minutes of demoing Typed Clojure at the REPL?
09:50	ambrosebs	I was thinking of putting some theory in there, but it seems secondary.
09:52	mpan	Is there a way to get filter and reduce of the same predicate in one pass?
09:52	S11001001	ambrosebs: theoretically, I'd like to hear about the limitations of the type system and ways to deal with them when ann-ing code
09:54	ambrosebs	S11001001: right, I think I can convey that without particularly going into the theory.
09:54	S11001001	not knowing the theory might lead to cargo-cult typing
09:56	ambrosebs	good point, I'll have to find a good balance.
09:56	ambrosebs	Thankfully I do have a dissertation I can point to for the more curious
09:59	ambrosebs	I'll definitely be using the live demos as vehicle to the theory, but it's probably not a good idea to go into much detail.
10:03	AntelopeSalad_	what would i have to type in the repl to start a server using this ring lib? https://github.com/aesterline/netty-ring-adapter
10:03	AntelopeSalad_	the instructions do not make it clear
10:03	AntelopeSalad_	i tried about 10 different things
10:05	ejackson	ambrosebs: yeah, too much theory is hard to absorb in a talk
10:05	ejackson	better to show how and why with pointers to the deep stuff to be considered offline with time
10:06	S11001001	all the why is theory, though, ejackson
10:06	ejackson	yeah, but its too hard to grok in a 40 min talk
10:06	ejackson	better get some concrete examples of what can be done
10:06	ejackson	so that you're motivated to figure out the theory, as well as have some concrete notions to tie the theory onto
10:08	ejackson	imho, anyway. i am quite slow to understand though
10:12	jimduey	ambrosebs: looking forward to your conj talk and unsession.
10:12	ambrosebs	jimduey: awesome :)
10:14	ambrosebs	jimduey: I'll walk you through the work so far on typing conduits
10:15	ejackson	i really liked conduits when I played with them
10:15	ejackson	typing them sounds insanely hard
10:16	ambrosebs	I don't know how jimduey implementing in Clojure TBH, amazing feat
10:17	ambrosebs	ejackson: It's fascinating stuff, and I have a pretty good idea what needs to be done in order to type the full conduit library
10:17	ambrosebs	well, aside from the really hard bits :D
10:17	ejackson	:)
10:19	ticking	Hm I'm curious what would be the most idiomatic way to share method implementations between records?
10:20	hyPiRion	ticking: Make the implementations functions, and just use the function.
10:22	ticking	hyPiRion, hm k ^^ thanks ^^
10:26	ticking	hm btw is there a way to do it the other way around ^^? Having the same implementation for two protocol methods?
10:27	mpan	ticking: taking advantage of functions-as-values, defining the function, and then just sticking it as a value where you please... so really the same idea as the other case
10:29	ejackson	exactly: just call the same function for both extensions of the Protocol
10:29	S11001001	ticking: Clojure isn't OOA&D friendly
10:30	S11001001	ticking: and Clojure hates implementation inheritance. Hates.
10:30	ticking	mpan I was thinking of something among the line of (defprotocol A (foo [this])) (defprotocol B (foo [this])) (defrecord Fool [] [A B] (foo [this] "bar))
10:31	ticking	S11001001 yeah I noticed that sucks because I hate ;) code dublication ^^
10:31	mpan	but you're only duplicating the names
10:31	mpan	the definition happens only once
10:31	S11001001	ticking: we don't have duplication problem because we just write functions to operate on our data structures
10:32	S11001001	ticking: as a sort of transition to this style, I strongly suggest organizing your methods around the multimethod system instead of protocols/records. You'll get controlled implementation inheritance that way.
10:33	hyPiRion	ticking: I don't think you would (nor should) extend protocols so often that same names becomes an issue.
10:33	mpan	but yea, as others have said, you don't usually want a huge tree of relationships as Java would have you make
10:34	hyPiRion	So actually, having [A B] would make stuff more complicated in general.
10:35	ticking	S11001001 right, mutlimethods seem to be the way to go, sucks a bit because speed is a concern too ^^
10:35	S11001001	ticking: again, it's transitional; the clojure way is really just to write ordinary functions
10:36	S11001001	ticking: and you know what they say about premature optimization
10:36	ticking	S11001001 yeah totally agree on that ^^ but grinding through TBs of data is not fun :D
10:37	mpan	sounds interesting, though... where did you get TBs of data to begin with?
10:38	ticking	hyPiRion it's not by massive extension this occurs, I have two protocols which have their own records (scanned page data, only images, and only xml metadata) and one record which needs both (combined scanned images and metadata)
10:38	ticking	mpan ocr_ed books
10:38	ticking	with a ton of metadata ^^
10:38	mpan	sounds cool
10:39	hyPiRion	ticking: Ah, gotcha
10:39	hyPiRion	Not saying it's always bad, but it's sometimes a bit dangerous
10:39	dnolen	ticking: I would follow ejackson's advice if you want to share implementations - create a shared function that protocol implementations call.
10:40	jimduey	ambrosebs: would really like to see your ideas on typing conduit. At some point I'm going to get around to rewriting it using a protocol arrows library.
10:40	ticking	dnolen yeah but if an implementation is not shared between all records I probably end with function-sharedA function-sharedB e.t.c. seems kinda nasty^^
10:42	dnolen	ticking: without knowing more details about what design you have in mind - can't really tell. core.logic is pretty complicated tho and sharing implementations works fine for me.
10:42	ejackson	well if the function is different for different types...
10:42	ticking	ejackson yeah true, just feels like the langauges job to me ;D
10:42	dnolen	ticking: delegation is possibility as well if your protocol design isn't overly complex
10:43	dnolen	which it shouldn't be
10:43	ticking	dnolen yeah delegation is probably what I'm ending up with, the protocols introduction video just bashed them so hard that I had hoped there was another way :D
10:44	ticking	but thanks a lot everybody :D
10:46	mpan	ticking: which video was that? I think I'd be interested in watching it
10:47	ticking	mpan, http://vimeo.com/11236603 there you go ^^
10:47	mpan	thanks!
10:59	ambrosebs	jimduey: sounds great, can't wait!
11:12	Cubic	Maybe I have a thinking problem here, but is there any other way for a REPL to interact with a running program except by changing global state?
11:12	gfredericks	Cubic: that sounds plausible; changing local state doesn't sound promising.
11:15	Cubic	gfredericks: I mean I've been "trained" to understand globals as a Bad Thing, but I really want to develop/debug with a REPL. Would it make sense to expose some essentially local refs/atoms as globals in a sort of (by convention) repl only namespace?
11:19	jkkramer	Cubic: most repl development involves redefinition of functions and constants, and running functions, rather than mutating app data
11:20	jkkramer	mutating app data is ok during development, if your state is stored in refs of some kind
11:24	Cubic	jkkramer: I guess redefining functions and constants is easier, but I wanted to do things like add and edit entities to/in my world from the REPL to see how it works out
11:26	jkkramer	Cubic: then presumably your entities are stored in ref/atom/agent, so you are free to do that, sure. it can be fun, especially if your app keeps running
11:28	mpan	Is it safe to drop-while a lazy seq on a pred such that the evaluation may start causing exceptions after pred starts becoming false? Will chunking cause me problems here?
11:30	mpan	I am looking for "the first x in s such that not p(x)" but evaluating p on anything after x is likely to crash
11:30	Cubic	Yep. The way I was thinking was storing the relevant world state in a map ref, doing all the IO in one thread and the rest in one or more other thread(s). I'm still not sure if that's a good idea, but I'll see how it works out. Thanks either way
11:32	mpan	The source seems like it might be safe but I'm not sure I'm understanding it correctly.
11:32	Cubic	mpan: If there are any x for which p(x) is false you will evaluate p at least for one of them with drop-while. Or am I misunderstanding you?
11:33	mpan	p is monotonic over s
11:33	mpan	map p s would look like ... T T T T F F F F ...
11:33	mpan	I want the first F there
11:34	mpan	trying to evaluate anything after the first F is... I don't know, but ideally I don't want to try
11:34	dnolen	mpan: aren't you in control of the sequences?
11:34	dnolen	er sequence
11:34	mpan	I wrote it rather poorly for anything beyond the standard case
11:35	mpan	and the standard case is to stop trying to ask for elements after seeing an F
11:35	Cubic	mpan: I don't know how drop-while is implemented, but if it would keep evaluating after the first F I'd count that as a bug.
11:35	TimMc	mpan: There is unchunk in amalloy's useful.
11:35	mpan	as in, I got the sequence from calling iterate
11:35	TimMc	Oh, it won't be chunked, then.
11:35	mpan	oh awesome!
11:35	mpan	Thanks guys!
11:38	TimMc	mpan: chunking is an opt-in behavior -- map, for instance, checks whether the incoming seq is chunked, and if so, produces a chunked output.
11:38	TimMc	I'd be really surprised if drop-while preserved chunking.
11:55	dnolen	TimMc: drop-while doesn't change chunking
11:56	dnolen	TimMc: I don't think any of the core seq ops defeat chunking - which is the correct behavior
11:58	TimMc	&(dorun (take 5 (map #(println %) (range 100))))
11:58	lazybot	⇒ 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 nil
11:58	TimMc	&(dorun (drop-while #(println %) (range 100)))
11:58	lazybot	⇒ 0 nil
11:58	TimMc	It looks like drop-while doesn't apply that predicate to anything beyond the first false result.
11:59	S11001001	dnolen: map on 2 or more seqs
12:07	uroborus_labs	idbknox: What are your thoughts on the pull request I sent you the other day for waltz?
12:13	dnolen	TimMc: oh right, because the function application for drop-while happens in step - still if you have a chunked seq underneath rest will actually load the next 32 items if it needs too. it doesn't make it go away.
12:13	dnolen	,(type (seq (drop-while (fn [n] (< n 5)) (range 100))))
12:13	clojurebot	clojure.lang.ChunkedCons
12:22	Cubic	I thought the question was about whether or not the predicate was applied, not about it's chunkiness.
12:26	duck1123	uroborus_labs: Have you tried superbobry's fork of waltz? (or my fork of his) Some of his changes are pretty nice
12:26	uroborus_labs	duck1123: No i have not but I will check it out.
12:27	uroborus_labs	I was just going to fork it and upload to clojars because git-deps is mucking things.
12:28	duck1123	I ended up fixing that same issue because those warnings were driving me nuts
12:28	uroborus_labs	duck1123: Yeah, agreed
12:29	uroborus_labs	duck1123: Looking at superobry's now
12:29	uroborus_labs	What is different?
12:30	duck1123	some of the function names are different, he rewrote a big part of it. Saner execution
12:30	duck1123	He also changed the logging. At one point I was on the fence about that, but I've come around to using his logging system
12:31	duck1123	My only gripe is it adds a newline at the end of the logs. My attempts to track it down and fix it have failed.
12:37	Cubic	Is it normal that clojure.math.numeric-tower throws out a ton of reflection warnings? I'd think that a numeric library would be better off avoiding reflection
12:39	uroborus_labs	Is there a better way to include other repos and libs than using lein git-deps?
12:40	technomancy	uroborus_labs: yes, real deps are always preferable to git-deps
12:41	uroborus_labs	technomancy: Agreed, but what is the best way then to use another repo, particularly a forked one? Upload it to clojars yourself?
12:42	technomancy	uroborus_labs: if the upstream maintainer has abandoned it, sure
12:42	uroborus_labs	technomancy: Ok cool
12:42	duck1123	What if they've just forked it but not provided a release
12:43	technomancy	if they've forked it then they're not the upstream maintainer
12:43	uroborus_labs	Hmm
12:43	technomancy	seeing a lot of unofficial forks on clojars is usually a sign that either the maintainer is doing a bad job, or the users of the lib aren't good at communicating
12:44	dnolen	Cubic: TimMc said drop-while doesn't preserve chunking - I was just pointing out that's not true. You could easily see multiple side effects fire off unexpectedly at some future point because the seq fns don't mess w/ chunking.
12:45	TimMc	"Preserve" is ambiguous there.
12:46	TimMc	As long as drop-while doesn't clojure.core/map the predicate across the input and then check the outputs, it's fine. :-)
12:46	clerksy	technomancy: Been trying to contact ibdknox through twitter/github/irc. No response.
12:46	dnolen	,(take 1 (map #(do (println %) %) (drop-while (fn [n] (< n 5)) (range 100)))))
12:46	clojurebot	(5
12:46	clojurebot	6
12:46	clojurebot	7
12:46	clojurebot	8
12:46	clojurebot	9
12:46	TimMc	oh no
12:46	clojurebot	10
12:46	clojurebot	11
12:46	clojurebot	12
12:46	clojurebot	13
12:46	clojurebot	14
12:46	clojurebot	15
12:46	clojurebot	16
12:46	clojurebot	17
12:46	clojurebot	18
12:47	clojurebot	19
12:47	clojurebot	20
12:47	clojurebot	21
12:47	clojurebot	22
12:47	clojurebot	23
12:47	clojurebot	24
12:47	clojurebot	25
12:47	clojurebot	26
12:47	technomancy	quick, someone figure out how to crash clojurebot
12:47	llasram	whelp
12:47	clojurebot	27
12:47	clojurebot	28
12:47	clojurebot	29
12:47	clojurebot	30
12:47	Fossi	-_-
12:47	dnolen	ha
12:47	clojurebot	31
12:47	clojurebot	5)
12:47	technomancy	clerksy: sounds like uploading a fork could be justified then, but it's a shame
12:47	TimMc	I specifically used lazybot for my demo earlier. :-)
12:48	clerksy	technomancy: Agreed, but he could just be really busy with lighttable :)
12:48	technomancy	if someone's volunteering to take over maintenance of a library you should always be able to find time to make that happen
12:50	mpenet	clerksy: what library are you thinking of?
12:51	clerksy	mpenet: waltz
12:51	clerksy	https://github.com/ibdknox/waltz/blob/master/src/waltz/state.cljs on line 30 get-in-sm is called before it is defined
12:52	clerksy	Throws compilation errors
12:52	mpenet	clerksy: same issue for me with jayq, I asked him earlier if he wanted help with this, as I have a fork with bugfixes & features
12:52	clerksy	mpenet: What was the response?
12:52	mpenet	clerksy: still waiting, I /msg him this time, we'll see. But as you said he must be really busy
12:54	TimMc	technomancy: DoS via bot amplification.
12:55	mpenet	clerksy: fork+publish under your own group-id is fine for a temp solution though
12:56	TimMc	lazybot should probably be set to ignore clojurebot.
12:56	clerksy	mpenet: ok cool
13:32	ibdknox	clerksy: mpenet: I'm happy to push some of my repos into organizations so that they can be better cared by others :) My time is extremely limited
13:32	ibdknox	I've been trying to find someone to really help with korma, as I think it's the one that needs the most work and I really don't have time to work on that one
13:33	clerksy	ibdknox: That sounds like a great idea
13:34	ibdknox	clerksy: btw, fwiw, those are just warnings - CLJS doesn't suffer from the same limitations that CLJ does :)
13:35	ibdknox	clerksy: should be fixed, no question. But they're not functional issues
13:35	clerksy	ibdknox: I know, but I like not having the warnings ;)
13:35	mpenet	ibdknox: That would be great.
13:36	ibdknox	let me get through this release (early next week) and we can figure that out
13:36	ibdknox	mpenet: clerksy: sound good?
13:37	mpenet	ibdknox: yup, sounds good
13:37	clerksy	ibdknox: Sounds good to me
13:37	ibdknox	great :)
13:38	Cubic	TimMc: He doesn't already do that? So if anyone feels like it they can deadlock clojurebot and lazybot?
13:38	mpenet	ibdknox: btw, maybe an organisation is not necessary, you can just add collaborators, but that is your call.
13:39	TimMc	,(println "&(println \"Cubic: Well, they can do this.\")")
13:39	clojurebot	&(println "Cubic: Well, they can do this.")
13:39	lazybot	⇒ Cubic: Well, they can do this. nil
13:39	mpenet	Cubic: worst, flood the channel until one of them is killed
13:39	duck1123	lazybot prefixes it's output so it won't be picked up by clojurebot
13:40	TimMc	I haven't figured out how to get a reflective bot-quine going, but it can be used to aplify the attack.
13:40	TimMc	*amplify
13:40	mpenet	aww, no infinite ping pong session then
13:41	TimMc	I've privately tested a command that would eat up about 300 lines of of 80-column scrollback.
13:41	ibdknox	clerksy: mpenet: what are your clojars names - mpenet is right for now, I'll add you as collaborators to let you guys keep moving on this
13:42	mpenet	ibdknox: same nick on github
13:43	ibdknox	mpenet: added you to jayq, anything else?
13:43	mpenet	ibdknox: thanks, no that's good
13:44	ibdknox	you should be able to push to clojars as well
13:44	mpenet	great
13:54	msamiry	I'm interested in finding clojurians in Cairo - Egypt, is there any?
14:11	technomancy	interesting effort for webby folks building on existing js libs: http://www.webjars.org/
14:13	hiredman	technomancy: slick
14:15	technomancy	except it doesn't mention the :repositories entry to add?
14:15	technomancy	oh, they're all in central. that's interesting.
14:15	xeqi	neat, too bad they're not sorted
14:16	duck1123	hmm... If i add these, will cljsbuild pick them up? Anyone try?
14:17	hiredman	duck1123: "pick them up" in what way?
14:18	hiredman	the common case I imagine would be to just serve them as resources out of the jars
14:20	mpenet	ibdknox: all cleaneded, I released under 0.2.0 since there is a possible breaking change for people using it ($.ajax feature for guessing decoder from mimetypes)
14:20	duck1123	I was thinking about external libs like how cljs libs work, but I realize it's not that easy
14:21	hiredman	you would want an extern file for the libraries
14:21	hiredman	which, I still have no idea how to do
14:22	hiredman	there is some website where which supposedly has some code to auto generate them
14:22	hiredman	http://www.dotnetwise.com/Code/Externs/
14:22	hiredman	dunno if it works
14:25	zackzackzack	Is there a way to make functions available in a namespace that were defined outside of the namespace originally?
14:26	S11001001	zackzackzack: do you mean s,available in,be exported from,?
14:27	zackzackzack	Sort of. So the context is that I am writing a library.
14:27	S11001001	zackzackzack: and you want a "export commonly used stuff" facade over the more specific modules
14:27	zackzackzack	I think so
14:28	mpenet	zackzackzack: useful has a alias-ns fn for that
14:29	mpenet	zackzackzack: an example here: https://github.com/flatland/useful/blob/develop/src/useful/core.clj
14:29	dnolen	ibdknox: ping
14:29	S11001001	zackzackzack: there's no built-in facility for that. But namespaces are mutable and first-class, so you can iterate over ns-publics of a space, interning each. compojure had immigrate, and I guess what mpenet said is that in a box.
14:29	mpenet	it will alias the entire ns, but there is alias-var also I think
14:30	zackzackzack	Yeah! https://groups.google.com/forum/?fromgroups=#!topic/clojure/GAGF38uI1-o
14:30	zackzackzack	Great, thank you S11001001 and mpenet
14:52	clerksy	idbknox: I am in the middle of setting up a new github and clojars account. My current one is for a psuedonym while this is not.
14:57	TylerGillies	what does in means in "contains?" where its constant or logarithmic time?
14:57	nDuff	TylerGillies: What part of that isn't clear?
14:58	nDuff	TylerGillies: it's a promise that it'll be faster than O(n).
14:58	TylerGillies	nDuff: oh gotcha
14:58	TylerGillies	thanks
15:05	mklappstuhl	I have a funtction that works when retyped at the repl but doesnt when I define it in the a file ...
15:05	mklappstuhl	https://www.refheap.com/paste/6325
15:05	TimMc	mklappstuhl: Does it depend on something defined earlier in the REPL?
15:05	mklappstuhl	ArityException Wrong number of args (0) passed to: PersistentArrayMap clojure.lang.AFn.throwArity (AFn.java:437)
15:06	hiredman	mklappstuhl: testfund is not a function
15:06	TimMc	mklappstuhl: Oh, that's because you're trying to call {name trades} as a fn with no args.
15:08	mklappstuhl	TimMc, ahhh
15:08	TimMc	Also, what's with fund's arity? It takes two args yet you're calling it with 3 or 5.
15:09	tgoossens	Hi
15:09	mklappstuhl	TimMc, I had [name & trades] earlier and changed it trying to debug the isue
15:10	tgoossens	What would be an efficient way to parse this?
15:10	tgoossens	http://pastebin.com/1WwKD3Yb
15:10	TimMc	mklappstuhl: If that had been the issue it would have made helping very, very difficult.
15:10	tgoossens	it is a syntax for encoding a maze
15:10	tgoossens	first two numbers are dimensions
15:10	clerksy	ibdknox: My new github is "clerksy" and my new clojars is "clerksy"
15:10	clerksy	ibdknox: Thanks
15:10	tgoossens	For simplicity lets say that every token must be matched on a number
15:12	tgoossens	You don't have to make an actual algorithm (however that would always be nice :D) . just giving suggestions is fine :)
15:12	mklappstuhl	TimMc, sorry then. how would I have to change the testfund method?
15:13	TimMc	mklappstuhl: (defn testfund [] ...)
15:13	hiredman	function, not a method
15:21	TimMc	tgoossens: You haven't provided enough info about the format.
15:22	tgoossens	Ok
15:22	tgoossens	what info are you missing?
15:22	tgoossens	Is it clearer if i just say
15:22	tgoossens	that the all the commands could just as well be A B C D E ...
15:24	TimMc	Are newlines significant?
15:24	TimMc	Is that a comment?
15:25	TimMc	If so, can comments occur anywhere, or just at the beginning of a line? What if there is just some whitespace before the comment?
15:28	tgoossens	# is a comment line
15:32	tgoossens	but nvm. I think i got it :)
15:36	tgoossens	Anyone here a regex expert?
15:37	gfredericks	~anyone
15:37	clojurebot	Just a heads up, you're more likely to get some help if you ask the question you really want the answer to, instead of "does anyone ..."
15:38	tgoossens	Ok thanks for the tip
15:38	tgoossens	kinda new here ;)
15:39	uvtc	tgoossens: what are you trying to match?
15:39	tgoossens	How can I in regex match the tokens "Straight.S.55" and "DeadEnd.N" in http://pastebin.com/1WwKD3Yb . While ignoring the lines with # (comment lines)
15:39	bbloom	tgoossens: regexes are just one of many tools to use for this sort of stuff
15:40	bbloom	tgoossens: why not filter blank and comment lines first?
15:40	tgoossens	mmyes :D
15:41	tgoossens	I'm using a library that makes it very easy to do what i want to do. All i need is a way to regexmatch that token /D
15:41	tgoossens	:D
15:41	dakrone	anyone have an idea how to get around "java.lang.IllegalArgumentException: Mismatched return type: verify, expected: void, had: java.lang.Object" when reifying a method that has a return type of void?
15:41	tgoossens	(some project of last year, not a public lib or something)
15:42	Cubic	I suppose it could be done, but it's really much easier when you do it like bbloom said.
15:43	gfredericks	regexes aren't very good at negating things
15:43	gfredericks	though technically they can :)
15:43	hiredman	dakrone: put a nil at the end
15:43	tgoossens	mm ok
15:43	tgoossens	i'll filter out the comments
15:43	dakrone	hiredman: doesn't work, same error
15:43	tgoossens	that solves that
15:43	hiredman	dakrone: or maybe ^void before the name of the method
15:44	Cubic	#"^[^#]*(Straight\.S\.55\|DeadEnd\.N)"?
15:44	gfredericks	:)
15:44	tgoossens	ow
15:45	tgoossens	i wasn't clear enough
15:45	Cubic	I think you might need another 2 pairs of parantheseses, but other than that I think this works
15:45	amalloy	&(reify java.util.Collection (add [this x] nil))
15:45	amalloy	oops, that's boolean
15:45	lazybot	⇒ #<sandbox7657$eval421852$reify__421853 sandbox7657$eval421852$reify__421853@1dd105e>
15:45	mpenet	Should tools.namespace work on cljs files? It seems to be a trival modification. This would allow libs like codox to work on cljs without modifications (or reinventing the wheel). https://github.com/clojure/tools.namespace/blob/master/src/main/clojure/clojure/tools/namespace/file.clj#L30
15:45	tgoossens	I just gave those two as example
15:45	amalloy	&(reify java.util.Collection (clear [this] nil))
15:45	lazybot	⇒ #<sandbox7657$eval421864$reify__421865 sandbox7657$eval421864$reify__421865@f2d92c>
15:45	mpenet	or maybe I am missing something
15:45	amalloy	dakrone: ie, in general reify works out of the box for void methods. more detail is needed to solve your problem
15:45	tgoossens	i need to be able to match any command in the form xx.yy.zz or xx.yy
15:45	Cubic	Then insert there the regex for your token :O
15:45	dnolen	mpenet: it probably easily could
15:45	mpenet	I am asking here because my CA is still in the middle of the Atlantic I guess. No access to clj-dev
15:46	tgoossens	i'll google a bit more
15:46	tgoossens	haven't got much experience with regex
15:46	Cubic	xx and yy can be what? Digits, letters?
15:46	dakrone	amalloy: hiredman: this is what it looks like: http://p.draines.com/1351799152559b274549e.txt
15:46	mpenet	dnolen: Should I fill a ticket for that?
15:46	tgoossens	xx and yy strings. zz numbers
15:46	tgoossens	*digits
15:46	tgoossens	*chars
15:47	Cubic	what kind of strings?
15:47	dnolen	mpenet: it's up to you, but it sounds like a good idea. libs that "just work" w/ CLJS are always nice.
15:47	clojurebot	kind of interesting if unfold is the intermediate steps of a fold
15:47	mpenet	dnolen: will do then
15:47	tgoossens	nvm
15:47	tgoossens	i think i found it :)
15:48	tgoossens	oh
15:48	mpenet	dnolen: it is on the CLJ jira, or is there a separate one for these sub projects
15:48	tgoossens	all the examples are given in the link
15:48	dnolen	mpenet: all contrib projects have their own JIRA
15:48	mpenet	nevermind found it
15:48	tgoossens	i'm almost there
15:48	tgoossens	[^\s"']+\|"([^"])"\|'([^'])'
15:49	tgoossens	but i'll stop asking half-complete questions
15:49	tgoossens	so nevermind :)
15:52	TimMc	tgoossens: 1) Filter blank and comment lines, 2) remove that initial line, 3) split on whitespace, 4) split each element on \.
15:53	TimMc	No regex involved, except maybe for step 1.
15:59	dakrone	gah, I can never remember what the type hint for an array of strings is
16:01	dnolen	dakrone: who can?
16:02	hiredman	people who know the syntax for jvm type descriptors
16:02	gfredericks	that sounds almost tautological
16:02	dakrone	does anybody know what it is??
16:02	lazybot	dakrone: Definitely not.
16:03	dnolen	,(class (make-array String 0))
16:03	clojurebot	[Ljava.lang.String;
16:03	dakrone	there it is
16:04	gfredericks	but that's not readable; do you embed that in a string to do the type hint?
16:04	SegFaultAX	I just had the weirdest interview of my life. The candidate literally sprinted out of our office after I asked him the most basic technical question ever.
16:04	mpenet	woa, no project.clj file :( it's going to take more time to test than to patch
16:04	SegFaultAX	Like literally stood up, told me "if this is part of the qualification, I think we're done here" and ran out of our office.
16:05	gfredericks	SegFaultAX: I can't tell if that implied too hard or too easy
16:05	SegFaultAX	gfredericks: Sorry, it was too hard.
16:06	gfredericks	well at least he didn't waste your time any further
16:06	SegFaultAX	gfredericks: That's true. But the manner in which he did it was so bizarre.
16:07	SegFaultAX	gfredericks: Maybe I need to find even easier questions :(
16:08	dnolen	gfredericks: yes put it in a string
16:10	hyPiRion	SegFaultAX: On what scale would you rate the problem, based on difficulty?
16:10	hyPiRion	Is it like the medium questions on 4clojure, the hard ones, or is it like Project Euler hard?
16:10	SegFaultAX	hyPiRion: For someone who's been writing code for a couple weeks, maybe a 4 out of 10. For someone with years of experience, maybe a 1.
16:11	hyPiRion	Oh, okay.
16:11	gfredericks	reversing a string by calling its #reverse method?
16:11	SegFaultAX	hyPiRion: Let me tell you the exact question: given a dictionary/set/list/whatever you want of english words, write for me an algorithm "splittable?" that can determine whether the string can be split at any place such that the left and right sides are both in the dictionary.
16:11	SegFaultAX	hyPiRion: If the entire word is in the dictionary, you can return true immediately.
16:12	SegFaultAX	hyPiRion: The examples I offer are "helloworld" should return true, "lkajgh" should return false, and "hello" should return true.
16:12	hyPiRion	SegFaultAX: Well, yeah, that's pretty easy. I guess only one split.
16:12	SegFaultAX	hyPiRion: There are obvious ways to make it harder, but we didn't even get to that. Like how do you handle n-splits? What if you have a pathological dictionary and input string, etc.
16:13	SegFaultAX	This was just, given a string and a dictionary of some kind, determine if it's splittable.
16:13	Cheiron	Hi, I set :java-source-paths ["src/main/java"] in project.clj then lein javac why my Java class isn't compiled?
16:14	hyPiRion	SegFaultAX: Yeah, but I'd assume most people with at least two year of programming experience would figure our a solution
16:14	SegFaultAX	hyPiRion: This guy had 8 - 10 years. At least that's what he said.
16:14	hyPiRion	Maybe you dodged a bullet, or maybe my view on the world is strange.
16:16	duck1123	SegFaultAX: so should "hellolkajgh" return false?
16:17	gfredericks	(fn [dict word] (or (dict word) (some (fn [i] (and (dict (subs word 0 i)) (dict (subs word i)))) (range 1 (count word)))))
16:18	ytnnoj	hey, I'd like to take a value v, a list args = [a b c] and a function f(v, arg) and get f(f(f(v, a), b, c)
16:18	ytnnoj	is there an inbuilt macro that would do something like that
16:18	hammer	&(let f [(fn [dict word] (or (dict word) (some (fn [i] (and (dict (subs word 0 i)) (dict (subs word i)))) (range 1 (count word)))))] (f '("hello" "world") "helloworld")))
16:18	lazybot	java.lang.IllegalArgumentException: let requires a vector for its binding in clojure.core:1
16:18	dnolen	ytnnoj: you want reduce
16:19	duck1123	(doc reduce)
16:19	clojurebot	"([f coll] [f val coll]); f should be a function of 2 arguments. If val is not supplied, returns the result of applying f to the first 2 items in coll, then applying f to that result and the 3rd item, etc. If coll contains no items, f must accept no arguments as well, and reduce returns the result of calling f with no arguments. If coll has only 1 item, it is returned and f is not called. If val is supplied, returns the resu
16:19	ytnnoj	for arbritatry values of args, of course
16:19	hammer	&(let [f (fn [dict word] (or (dict word) (some (fn [i] (and (dict (subs word 0 i)) (dict (subs word i)))) (range 1 (count word)))))] (f '("hello" "world") "helloworld")))
16:19	lazybot	java.lang.ClassCastException: clojure.lang.PersistentList cannot be cast to clojure.lang.IFn
16:19	hammer	okay i give up
16:19	hammer	i just wanted to see if it worked
16:19	dnolen	,(reduce str "foo" ["a" "b" "c"])
16:19	clojurebot	"fooabc"
16:19	ytnnoj	oh of course i do dnolen, cheers.
16:20	jasonjckn	How does :state work in gen-class, my state method isn't being executed https://www.refheap.com/paste/6328
16:23	Apage43	,(let [word "hamstar" dict #{"ham" "star"}] (filter (partial every? dict) (for [i (range (count word))] (mapv (partial apply str) [(take i word) (drop i word)]))))
16:23	clojurebot	(["ham" "star"])
16:24	jasonjckn	nvm
16:25	duck1123	You probably want to start at the end of the input, that way if the word matches the dict it'll finish quickly
16:26	amalloy	duck1123: or start at 0 instead of at 1 :P
16:27	hyPiRion	Now let's do it for n splits.
16:28	Apage43	(take 1) in front of that one would stop at the first found split, and it tries 0 first, so, to get a short-circuiting boolean one just wrap it in (empty? (first
16:28	Apage43	er, rather, nil?
16:28	Apage43	er, rather, just (first
16:28	Apage43	bleh
16:28	duck1123	It seems like one of those questions where seeing how the person handles the challenge is more important than if they get the right answer.
16:29	duck1123	walking out is not the right way to handle the challenge
16:30	hyPiRion	duck1123: Walking out is not the right way to handle any challenge, I believe
16:30	hyPiRion	Except maybe if the challenge is to walk out.
16:36	TimMc	SegFaultAX: Maybe a dictionary killed his mother.
16:37	gfredericks	q
16:37	duck1123	Is that what they mean by dictionary attack?
16:39	mabes	,(name :foo/bar)
16:39	clojurebot	"bar"
16:40	mabes	I take it that the above is relating to namespace considerations? I was expecting "foo/bar"
16:41	TimMc	Yes, a symbol is a local part with an optional namespace part.
16:41	TimMc	err, that's a keyword, sorry
16:42	mabes	yeah, I get it.. it helps in cases where you have:
16:42	mabes	,(name ::foo)
16:42	clojurebot	"foo"
16:42	mabes	,::foo
16:42	clojurebot	:sandbox/foo
16:42	TimMc	Fun fact: A keyword has a symbol under the surface.
16:43	TimMc	&(.sym :foo/bar)
16:43	lazybot	⇒ foo/bar
16:43	mabes	ahh, that will do the trick
16:43	mpenet	dnolen: Patch sent, I allow anything matching #"\.clj.?$", this would allow clj, cljs, cljc, cljx etc to be supported also, maybe it's overkill, we'll see what Stuart says.
16:43	mabes	,(str (.sym :foo/bar))
16:43	clojurebot	"foo/bar"
16:44	mabes	TimMc: thanks for that fun fact :)
16:46	SegFaultAX	duck1123: Yes, that's correct.
16:47	SegFaultAX	duck1123: There isn't anything tricky going on with the problem. It's really as simple as it sounds.
16:47	luxbock	I'm trying to follow along the book Joy of Clojure, chapter 3.4.2 where he talks about using java.awt.Frame to draw a xor-pattern
16:48	luxbock	I added up the code he talks about into a file (I didn't create a new lein project, but instead put it into the /src/ folder of an existing project that I Was trying out some other stuff with
16:48	TimMc	SegFaultAX: That's a fun question, I'm stealing it.
16:49	SegFaultAX	TimMc: It's really fun when you ramp up the difficulty.
16:49	luxbock	but I when I run the code I get CompilerException java.lang.ClassNotFoundException: jawa.awt.Frame
16:49	luxbock	here's what I have in my .clj file: http://pastebin.com/5Jkpgezu
16:49	luxbock	is it an issue with namespace?
16:50	raek	luxbock: you got cut off at "...I Was trying out some other stuff with"
16:50	TimMc	Yes, you need an :import.
16:50	luxbock	oh
16:50	TimMc	and a ns statement at the top, for that matter
16:50	raek	luxbock: no, you have a typo in "java" :)
16:50	raek	line 5
16:50	raek	it says "jawa"
16:51	luxbock	oh god damnit
16:51	TimMc	Oh hah! That too.
16:51	luxbock	I went through that file more than a few times
16:51	TimMc	That's right, if it's package-qualified, it doesn't need an import.
16:51	technomancy	"jawa" is javanese
16:52	luxbock	I figured that since the line got a proper syntax highlight that it must be correct, so didn't realize to look more carefully
16:53	SegFaultAX	luxbock: That'll teach ya that (not= syntactic-correctness logical-correctness)
16:54	nbeloglazov	Hi. Is there way to force printing from separate thread go to visible buffer, so I can see output?
16:54	luxbock	yeah, I'm trying to learn Emacs and Clojure at the same time, not having that much experience with programming to begin with so it's rather slow beginnings
16:54	amalloy	i dunno, i've never written code that made it past the reader and didn't correctly solve every present and future problem i wanted to solve
16:54	SegFaultAX	amalloy: Magical fingers, I guess?
16:55	technomancy	amalloy: you forgot to say "...in haskell"
16:55	amalloy	haha, i was thinking that technomancy
16:55	amalloy	SegFaultAX: more like selective memory
16:55	raek	nbeloglazov: you could try (alter-var-root #'out (constantly out)) in your repl
16:55	nbeloglazov	raek: does it have same effect if I create bound function via bound-fn?
16:56	nbeloglazov	from repl
16:56	raek	nbeloglazov: that should work too
16:57	nbeloglazov	raek: no, it doesn't help :(
16:57	luxbock	also, now that I managed to create that java-frame, how do I kill it without messing up the process that runs nrepl inside Emacs?
16:57	nbeloglazov	I think it's more complicated than just repl
16:57	raek	nbeloglazov: what repl are you using? swank? nrepl?
16:57	nbeloglazov	lightable :)
16:58	raek	oh, is this printing via clojure functions or direct cals to System.out?
16:58	nbeloglazov	via println
16:59	nbeloglazov	clojure functions
16:59	TimMc	amalloy: Really? You've never wanted to solve any problem?
17:00	amalloy	TimMc: yeah, i just like adding features
17:00	raek	do out and @(future out) return the same thing?
17:00	TimMc	Or you've never written code? :-P
17:00	TimMc	raek: I guess you're asking whether dynamic bindings are conveyed?
17:01	raek	hrm. right. forgot that 'future' does that now...
17:08	nbeloglazov	Cool. If I do (.write out "Hello") it is printed. If I do (def out out) (.write out "Hello") it just stored in out and not printed.
17:08	nbeloglazov	(.write out "Helll") in last command
17:11	technomancy	sounds like the repl is assuming disposable streams on a per-entry basis
17:11	technomancy	early versions of nrepl.el had that bug
17:15	nbeloglazov	techmonancy: When I do (def out out) and (= out out) immediately after that. It returns false. How it is possible?
17:15	nbeloglazov	technomancy*
17:16	TimMc	nbeloglazov: What he's saying is that a new output stream is being created for each execution.
17:16	TimMc	*evaluation
17:16	technomancy	yeah, it's a bug
17:16	reque	hey
17:16	technomancy	head to the issue tracker I guess?
17:16	technomancy	since apparently you aren't allowed to fix it yourself
17:18	reque	i'm trying to get a browser repl going using cljs-template, but I can't seem to figure out how to connect to the browser repl. Anyone got any experience with it or similar?
17:29	nbeloglazov	I've created a bug. It seems I'll need to collect all output from different threads to atom and then print it from time to time for now.
18:05	luxbock	when I use emacs and I execute a faulty command, is there a way to disable the screen from flooding with the error messages? I could do just fine with a more succinct version
18:05	luxbock	with nrepl that is
19:20	brainproxy	how can I test the relationship between (class '()) and (class '(1))
19:22	hiredman	,(empty? (clojure.set/intersection (supers (class '())) (supers (class '(1)))))
19:22	clojurebot	false
19:23	hiredman	,(contains? (clojure.set/intersection (supers (class '())) (supers (class '(1)))) java.util.List)
19:23	clojurebot	true
20:27	mklappstuhl	Is there a way to start a repl in function so that I can inspect stuff inside the execution scope?
20:27	frenchyp	hello
20:28	frenchyp	can I use leiningen checkout dependencies to include 1 lein project into a 2nd one?
20:29	mklappstuhl	frenchyp, why not?
20:29	ivan	mklappstuhl: there's a macro for that
20:29	mklappstuhl	you have to have it on clojars or somewhere of course
20:29	mklappstuhl	ivan, does this macro have a name? :)
20:29	frenchyp	mklappstuhl: oh, I am working on both at the same time
20:29	ivan	; local-bindings and eval-with-locals are from http://gist.github.com/252421
20:29	ivan	; Inspired by George Jahad's version: http://georgejahad.com/clojure/debug-repl.html
20:30	frenchyp	1 is the lib i m working on, the other is wrappnig code to interact with it
20:30	frenchyp	so pushing to clojars each time will get tiresome :)
20:34	tmciver	frenchyp: I don't use lein2, but I know in lein1 you could do it 2 different ways: you could 'lein install' one project jar in your local .m2 dir, or you can use a checkout deps.
20:44	frenchyp	tmciver: can't get the checkout deps to work, i'll try the other one
20:46	tmciver	frenchyp: are you using 'checkout-deps-share' as shown in this sample project.clj: https://github.com/technomancy/leiningen/blob/preview/sample.project.clj?
20:46	tmciver	frenchyp: I think checkout deps would make your life easier.
20:51	frenchyp	i'll test as soon as I can (upgrading to latest lein 2 now , and tethering thru my phone so it takes for ever)
20:51	frenchyp	but , if i'm reading the default values properly , it should be setup alreeady
21:04	frenchyp	doesn't work
21:13	frenchyp	ah, got it
21:13	frenchyp	the project I was linking to in checkouts had a '-' in its name. somewhow it didn't like it
21:14	frenchyp	hmm, spoke too fast, wrong dir
21:53	devn	How is it true that compojure still depends on 1.2.1?
21:55	devn	shrug
21:58	gfredericks	switch it to 1.5 and run the tests?
22:08	xeqi	it doesn't use anything thats changed or new past 1.2.1
22:46	myguidingstar	Hi. Anyone can tell me how to write a handler (with custom evaluator) for nrepl server?
23:09	xeqi	myguidingstar: are you wanting to make a middleware that handles a custon op?
23:11	myguidingstar	well, I'm not sure what op is
23:12	myguidingstar	All I want is to change the evaluator with my custom one
23:12	myguidingstar	Anything else should be kept
23:18	xeqi	sounds like you want to write a custom middleware for the "eval" op. perhaps looking at the builtin one can help; https://github.com/clojure/tools.nrepl/blob/master/src/main/clojure/clojure/tools/nrepl/middleware/interruptible_eval.clj#L153
23:19	myguidingstar	Thanks for your help, I'm reading it, too
23:21	xeqi	if you're able to rewrite the :code form then just delegating to the eval op might be easier, like how load-file works
23:27	myguidingstar	xeqi: I don't understand, or maybe I didn't read enough. Where is :code form or load-file?
23:27	xeqi	https://github.com/clojure/tools.nrepl/blob/master/src/main/clojure/clojure/tools/nrepl/middleware/load_file.clj#L61
23:28	xeqi	that one is a nrepl middleware that implements a new :op
23:28	xeqi	"load-file", and rewrites the :code to use the eval :op
23:29	xeqi	well, really just creates the :code form from the stuff it wants from its custom op arguments
23:30	xeqi	but it might be possible to create a middleware that checks for :op "eval" and rewrites the :code form, depending on how custom your evaluator is
23:32	myguidingstar	I'd try it now
23:34	xeqi	heh, I think thats what piggieback does
23:34	xeqi	though it gets pretty complex since it sets up a cljs environment
23:34	xeqi	https://github.com/cemerick/piggieback/blob/master/src/cemerick/piggieback.clj#L182
23:35	frenchyp	tmciver so the pb is when I add the checkout dep to :dependencies. If I don't do that it works...
23:38	myguidingstar	omg I didn't know piggieback existed
23:38	myguidingstar	thanks a lot, xeqi
23:38	xeqi	myguidingstar: was that what you wanted to do?
23:39	xeqi	run a cljs repl through nrepl
23:39	myguidingstar	No, but that will help a lot. In fact I'm writing a repl for clojurejs
23:42	myguidingstar	I guest I will just have to make a clone of `cljs-eval`
23:42	xeqi	frenchyp: what kind of error are you having with checkouts?
23:44	frenchyp	xeqi: i am trying to setup a 2 projects, one a sub project of the other. I am trying with brand new empty projects. If I declare one in the :dependencies of the other, it doesn't work. If I don't add the dependency and just have the checkouts symlink, it works
23:45	frenchyp	it doesn't work = could not resolve dependencies when running lein repl
23:46	xeqi	frenchyp: yeah, the checkout feature is meant to share the source from the sub-project to the parent; it doesn't change the dependency resolution in any way
23:47	xeqi	so it expects the sub-project to be in a repo somewhere for finding out sub-sub-deps and such
23:50	frenchyp	xeqi: hmm, ok. Am I reading the lein tuto wrong?
23:50	frenchyp	in the checkout section, it says "Libraries located under the checkouts directory take precedence over libraries pulled from repositories, but this is not a replacement for listing the project in your main project's :dependencies; it simply supplements that for convenience.
23:50	frenchyp	"
23:55	xeqi	frenchyp: that means it expects the sub project to be in :dependencies, and availiable in a repo. Then it lets the checkout source take precedence over the jar file it pulls down