#clojure – 2014-02-10

2014-02-10

00:01	benkay	if anyone in here has some java experience I'd appreciate a bit of help reading something that's totally opaque to me: https://code.google.com/p/bitcoinj/source/browse/core/src/main/java/com/google/bitcoin/core/Utils.java#130 i get that we're stepping through a 4-byte-wide byte array and setting each byte individually but this hairy (0xFF & (val >> 24)) stuff is breaking my poor brain
00:03	akurilin2	Hey folks. Are there any standard library functions that would help me simulate an ORDER BY + DISTINCT ON? E.g. get the latest records of types a,b,c?
00:05	akurilin2	Actually I guess there's distinct, which is cool, let me figure out whether it uses the earlier or later entries in the sequence.
00:05	arrdem	akurilin2: so.. our core sort takes an optional comparison argument...
00:05	arrdem	akurilin2: and yeah there's distinct...
00:05	arrdem	,(doc sort)
00:05	clojurebot	"([coll] [comp coll]); Returns a sorted sequence of the items in coll. If no comparator is supplied, uses compare. comparator must implement java.util.Comparator. If coll is a Java array, it will be modified. To avoid this, sort a copy of the array."
00:07	akurilin2	arrdem: aah there's no distinct-by, sad.
00:07	arrdem	akurilin2: no, but there's group-by...
00:08	arrdem	,(source group-by)
00:08	clojurebot	Source not found\n
00:08	arrdem	damnit clojurebot
00:08	arrdem	https://github.com/clojure/clojure/blob/master/src/clj/clojure/core.clj#L6514
00:09	akurilin2	No I got it, thanks, I've used it plenty before.
00:09	arrdem	yep. it's conj based.
00:09	arrdem	so you're order safe to use first or last to select the first or last value to exhibit a key.
00:10	akurilin2	so it sounds like I'd have to sort, then group by, then map to get the first of each group, then basically take the vals
00:11	arrdem	yarp.
00:11	arrdem	well.. map first over vals
00:11	akurilin2	oh yeah
00:20	akurilin2	arrdem: cool thank seems to work, had to dome some mental stretching there :P
00:21	arrdem	:D
00:27	jergason	are there any clojure git libraries?
00:27	arrdem	jergason: as in a Clojure git fs API?
00:27	jergason	so i want to manipulate a git repo - check out commits, list them, run git blame on files, ect
00:28	jergason	is there a library that can do this in pure clojure before I do a bunch of (sh) calls?
00:29	arrdem	$google github clojure git toolkit
00:29	lazybot	[heroku/heroku-buildpack-clojure · GitHub] https://github.com/heroku/heroku-buildpack-clojure
00:29	arrdem	I'm gonna guess not.
00:32	akurilin2	Working with libgit is not the most fun experience from what I hear.
00:32	akurilin2	Remember trying to use the Obj-C port, that was gross.
00:36	jergason	arrdem: yeah i'm pretty okay at googling stuff
00:36	arrdem	jergason: my appologies. you wouldn't belive the number of times I've answered someone's question like that.
00:37	jergason	np, i'm sure ill ask plenty of easily-googled questions :)
00:38	jergason	hrrrm, i might just use ruby or something then if this is just going to be a glorified shell script
00:38	turbofail	there's jgit
00:38	jergason	!
00:38	jergason	there is the noobness. i didn't even think of a java git lib
00:39	arrdem	's ok. I didn't either :P
00:39	jergason	:)
00:39	arrdem	(inc turbofail) ;; the fail is with us today
00:39	lazybot	⇒ 2
00:47	dacc	jergason: make us a nice clojure wrapper for it =)
00:48	jergason	i'm probably a few months away from being able to do that :)
00:49	turbofail	i feel like there was one already
00:50	turbofail	hm. looks like there are several possibly half-assed ones floating around
00:53	turbofail	on further inspection there is actually only one half-assed wrapper, the others are shelling out
00:53	arrdem	plz make not half assed one
01:09	muhoo	jgit sucks but is best i could find
01:17	jergason	does it suck for normal java reasons?
01:17	jergason	verbose etc
01:34	muhoo	slow, weird, verbose.
01:34	muhoo	but, it works
01:37	benkay	,(Integer/toBinaryString (long 0xd9b4bef9))
01:37	clojurebot	#<IllegalArgumentException java.lang.IllegalArgumentException: Value out of range for int: 3652501241>
01:37	benkay	what am i doing wrooooong
01:38	benkay	i just want four bytes representing that string to write into a socket and i can't figure this outtttt :(
01:39	muhoo	,(Long/toBinaryString (long 0xd9b4bef9)) ; ??
01:39	lazybot	muhoo: Definitely not.
01:39	clojurebot	"11011001101101001011111011111001"
01:39	muhoo	benkay: ^
01:39	benkay	oh well derp
01:39	benkay	thanks, muhoo.
01:39	muhoo	lazybot: piss off
01:39	muhoo	benkay: np
01:40	benkay	i have no java and clojure is my first lisp
01:40	benkay	just just wtf all the way down
01:40	muhoo	i feel ya
01:42	benkay	well, to represent that long.
01:45	muhoo	jergason: also, IIRC jgit is what eclipse uses internally for git, so it's a pretty safe bet for being maintained/bugfixed
02:02	benkay	so i want to stick ,0xd9 into a byte so that I can then send that out on a socket later, but this is the integer 217 which is too large for the java bytes (which go from -127 to 127). what am I conceptually missing in constructing this message?
02:02	benkay	the full 4-bytes i want to send is ,0xd9b4bef9
02:03	benkay	,0xd9b4bef9
02:03	clojurebot	3652501241
02:03	Cr8	,(unchecked-byte 0xd9)
02:03	clojurebot	-39
02:03	benkay	unchecked-byte?
02:04	Cr8	JVM numbers are all unsigned, but you can get what you probably want by forcing an unchecked cast
02:04	benkay	,(count (unchecked-byte 0xd9b4bef9))
02:04	clojurebot	#<UnsupportedOperationException java.lang.UnsupportedOperationException: count not supported on this type: Byte>
02:04	Cr8	that returns a byte, not a list of any time
02:04	Cr8	,(type (unchecked-byte 0xd9))
02:04	clojurebot	java.lang.Byte
02:04	benkay	yeah i see that now
02:05	Cr8	,(into-array Byte/TYPE (map unchecked-byte [0xd9 0xb4 0xbe 0xf9]))
02:05	clojurebot	#<byte[] [B@843e34>
02:05	benkay	oh man
02:05	benkay	,(vec (into-array Byte/TYPE (map unchecked-byte [0xd9 0xb4 0xbe 0xf9])))
02:05	clojurebot	[-39 -76 -66 -7]
02:05	benkay	groovy.
02:06	benkay	now i think i'm back at my "how do i see if this is going out over my socket as the correct hex" problem, but i think i'll leave that for tomorrow.
02:06	Cr8	https://github.com/apage43/hex
02:06	Cr8	is a thing I did, mostly to use at the REPL
02:06	Cr8	for stuff like that
02:07	benkay	ah neato.
02:07	benkay	that'll probably work for testing.
02:07	Cr8	,(apply str (map #(format "%02x" %) (into-array Byte/TYPE (map unchecked-byte [0xd9 0xb4 0xbe 0xf9]))))
02:07	clojurebot	"d9b4bef9"
02:09	benkay	many thanks sir! I do believe that I have enough to run with from here.
02:09	benkay	for at least another few hours ;)
02:09	Cr8	have fun
02:12	sm0ke	here is a idea, given a set of functions which take a type of input and give a type of output, is there something which facilitates writing a generic function which given input and desired output types, chains functions automaticaly and optimally ?
02:12	sm0ke	output type*
02:12	dsrx	sm0ke: there's something like that for haskell
02:12	sm0ke	dsrx: remeber the name?
02:12	dsrx	trying to
02:14	dsrx	sm0ke: djinn
02:14	sm0ke	it looks like a theorem prover, can that be employed?
02:16	sm0ke	hmm now that i think of it cant core.logic be used for this?
02:22	Cr8	ztellman's byte-streams does something like that
02:22	Cr8	https://github.com/ztellman/byte-streams
02:23	dsrx	sm0ke: the curry-howard correspondence says that a program and its type are directly related to a proof and its proposition
02:23	dsrx	... or something like that
02:23	ztellman	sm0ke: all you need is a memoized lookup of a exhaustive graph traversal
02:23	ztellman	nothing too fancy
02:23	dsrx	so it turns out theorem provers can generate a program given a type
02:24	dsrx	(and a bunch of other programs)
02:28	dsrx	and yeah, I imagine you could do something like that with core.logic
02:31	Cr8	now think about how you'd pull off these nifty cost hints https://github.com/ztellman/byte-streams/blob/master/src/byte_streams.clj#L613
02:35	sm0ke	ztellman: yea makes sense, wanted to void doing that
02:35	sm0ke	avoid*
02:36	sm0ke	as ideally i would have to write a shortest path algo
02:36	ztellman	sm0ke way simpler than anything else you just mentioned
02:36	ztellman	sm0ke or you can just use def-conversion in byte-streams
02:36	ztellman	it works on arbitrary types
02:37	sm0ke	ah thanks i will go though the code
02:39	sm0ke	would it make sense to seperate it out into a library?
02:40	sm0ke	or is it too trivial
02:40	ztellman	sm0ke maybe, haven't figured out any other real use cases besides the one I'm using it for
02:40	ztellman	I can't generalize without multiple concrete use cases
03:08	sm0ke	more i look at the code in byte_streams, i see exactly what i wanted, conversions with a cost. very nice indeed
04:41	marianoguerra	hi! what's the best way to do the following without race conditions and as efficient as possible?
04:41	marianoguerra	try to get a key from a map if it's there if not create it and return it
04:44	clgv	marianoguerra: this task occurs in memoization. you can read several variants over here: https://kotka.de/blog/2010/03/memoize_done_right.html
04:44	ambrosebs	marianoguerra: I assume you're talking about a mutable map?
04:45	clgv	marianoguerra: the above is assuming a clojure map
04:46	marianoguerra	I was thinking of an agent or an (atom {})
04:46	marianoguerra	but checking what's the correct way
04:47	marianoguerra	"Deutsche Bahn Version (or: „Delay is our core competence“)"
04:48	clgv	marianoguerra: check the atom version of memoize in the link above
04:48	clgv	yeah that's awesome ;)
04:48	marianoguerra	clgv: thanks!
04:48	opqdonut	to really get it right you will most probably need a lock
04:48	marianoguerra	will look into it
04:48	llasram	opqdonut: Drink more koolaid!
04:48	opqdonut	doing everything with one swap! is a very restricted model
04:48	llasram	~guards
04:48	clojurebot	SEIZE HIM!
04:49	opqdonut	especially since swap! is implemented wrong[tm]
04:49	llasram	opqdonut: How so?
04:49	clgv	opqdonut: yep. depending on how complicated it gets a `ref` might become mandatory
04:49	opqdonut	llasram: it returns the new value and not the old
04:49	clgv	why is that wrong?
04:49	pyrtsa	opqdonut: It's pretty easy to add the implementation that returns the old value.
04:49	opqdonut	or well, both variants are useful but I find myself often wanting the version that returns the previous value
04:49	llasram	opqdonut: Which 99% of the time is what you want
04:49	pyrtsa	clgv: Otherwise you can't see which value you swapped out.
04:50	opqdonut	e.g. implementing a queue with (atom []) is impossible using the current swap!
04:50	clgv	pyrtsa: your functions sees it when you swap it ;)
04:50	opqdonut	well you don't see the authoritative previous value
04:50	llasram	Sure, but implementing a queue with (atom []) is a terrible idea anyway, because you can't pop off the first element. A stack maybe :-)
04:50	opqdonut	your function might get run multiple times
04:50	llasram	Yes -- that's the whole point
04:51	pyrtsa	clgv: You often need to do something with the old value after the swap.
04:51	opqdonut	llasram: you could pop the first element if swap! returned the previous value
04:51	pyrtsa	opqdonut: https://gist.github.com/pyrtsa/7871080
04:51	llasram	opqdonut: I meant vectors don't allow you to efficiently pop the first element
04:51	pyrtsa	clgv: Example in the above link. ^
04:51	opqdonut	llasram: well, persistent-queue or whatever it's called then
04:51	llasram	Sure
04:51	opqdonut	llasram: or let's just say stack instead of queue
04:52	opqdonut	pyrtsa: sure, one can always compare-and-set!
04:52	llasram	Either way -- the point of atoms (and refs) is that you trade the risk of deadlocks for the risk of retries
04:52	pyrtsa	There is a persisten queue in Clojure, use (clojure.lang.PersistentQueue/EMPTY) to make an empty one and (conj q x), (into q xs) etc to use it.
04:53	llasram	Nothing is stopping you from calling `locking`
04:54	opqdonut	llasram: sure
04:54	opqdonut	and as I said, you will most probably need a lock
04:55	opqdonut	pyrtsa: yes, of course, but it's immutable. if you want to use one of those to synchronize between threads, you can't just put it in an atom
04:55	llasram	Most of the rest of the Clojure community will just need to agree to disagree with you
04:55	llasram	:-)
04:55	opqdonut	(or well, you can, but then you need to implement the logic using compare-and-set!)
04:56	pyrtsa	opqdonut: But that's exactly what I'm saying!
04:56	pyrtsa	...although it's cleaner with my above definition of exchange!.
04:56	opqdonut	I'm just mildly irked by not having swap-done-right! in the core
04:56	opqdonut	aka exchange!
04:56	opqdonut	yeah
04:56	pyrtsa	Yeah. It ought to be there.
04:59	clgv	pyrtsa: well. why dont you suggest it in jira?
05:01	pyrtsa	clgv: Haven't used it before. As a matter of fact, I will!
05:09	ddellacosta	opqdonut: can't you just get the old value with add-watch? seems like an easy way to do what you want
05:14	llasram	ddellacosta: Getting the old value with add-watch is pretty roundabout, especially given how the implementation works
05:14	llasram	Whenever you swap!, you have both the old and the new value. It's just the choice of which to return
05:29	Anderkent]away	llasram: eh, can you expand on that? swap! doesn't really let you return the old value.
05:30	llasram	Anderkent: The implementation in terms of compare-and-swap! has both the old and new values in order to compare-and-swap!. It's an interface decision which (or both) to return
05:30	Anderkent	ah. yeah, I agree with that
05:31	llasram	That's all I meant. You missed some early discussion where opqdonut claimed the current behavior to be incorrect
05:31	Anderkent	tbh I wish it returned the old one - with the old one and the function you can recreate the new one, can't really go the other way around
05:32	llasram	It does seem to be a common request. I've found it useful once or twice. Maybe it'll be in core some day :-)
05:32	llasram	In the mean time, really is easy enough to cart around in a utility namespace
05:32	pyrtsa	I'm adding it to Jira today as soon as I have a minute of time for it.
05:34	Anderkent	I guess; though when we hit that problem with a coworker (implementing a queue-like structure; i.e. a map where one element was a queue) and he proposed using compare-and-set! directly (i.e. reimplementing swap!) I was kinda meh'd
05:34	Anderkent	we went with transactions in the end, I think
05:34	llasram	Anderkent: That's what you do -- use compare-and-swap! directly to implement a swap!-like function which gives you the old value
05:35	llasram	compare-and-swap! is a feature, not an implementation detail :-)
05:35	pyrtsa	Indeed!
05:36	Anderkent	mhm, seemed a bit too low level for my taste; having to do the looping and retries and presumably failing after N retries all seems a little messy
05:36	llasram	But that's exactly what swap! does. Minus the failing -- it just retries forever
05:36	pyrtsa	I don't think swap! fails after N retries.
05:36	Anderkent	yeash, I know. I just want it abstracted away! :P
05:37	llasram	Well abstract it into a function yourself and you're good :-)
05:47	borkdude	Clojure rocks, just saying.
05:48	dsrx	disagree, a much better language is hask... oh, not the permitted time of day
05:52	llasram	heh
06:07	TheBrayn	what would you recommend for someone who has not done much functional programming yet to learn clojure?
06:08	broquaint	TheBrayn: 4clojure.com might be a good place to start.
06:08	broquaint	If you'd prefer to apply yourself to a book then I believe the Joy of Clojure may be apt.
06:11	AimHere	if you're using the 4clojure route, remember to 'watch' a bunch of people so you can check out their solutions to things, so that you might find a better or more idiomatic way of doing your clojuring.
06:13	klokbaske	hi there! newest version of lighttable gives me "No reader function for tag js" when I evaluate eg (def another-array #js [1 2 3])
06:13	TheBrayn	broquaint: 4clojure looks like fun, thanks
06:15	broquaint	np, enjoy :)
06:25	klokbaske	oh, seems I didn't run the latest version. 0.5.20 just told me it was, but little does it know ...
06:28	borkdude	TheBrayn Just pick any book and start doing clojure development
06:28	borkdude	TheBrayn I like Clojure Programming and The Joy of Clojure
06:29	clgv	is there a way to produce exceptions with line number information of the input param of a macro?
06:31	borkdude	TheBrayn Here you can read something about the differences among clojure books: http://michielborkent.nl/comparingclojurebooks.html
06:41	llasram	clgv: I believe you can get that information from the metadata on &form
06:42	clgv	llasram: humm but how do I structure error handling within the macro?
06:42	llasram	Just throw an exception
06:43	llasram	That's what the core macros do. Well, although is also one reason for the amazing stack traces
06:44	bob2	such terrible stacktraces:(
06:44	clgv	llasram: my scenario is that I have a DSL where I specify a nested loop-recur so I have several parts of code that get inserted. loop initialization, recursion statement, "body"
06:44	llasram	ok
06:45	clgv	a concrete error is a name used in the "body" which does not occur among the loop variables or outer bindings
06:47	clgv	llasram: ^
06:48	llasram	I must be missing something -- what do you not get with just throwing an exception in the macro code reporting that error?
06:49	clgv	llasram: the line information in the inserted code snippet which I can access via &form or such
06:50	llasram	clgv: Well, add it? :-) Check out clojure.core/assert-args maybe
06:51	clgv	llasram: thats the question how to do that. I have a look
06:52	Anderkent	I'm still not clear on whether you want the line to be the line in the macro that hit the failed assertion, or the line in the code that callse the macro
06:53	clgv	e.g. (my-macro (let [x (* y y)] (conj res x))) if "res" is not defined in the generated surrounding context where that body is put, I want the line information of the expression using "res"
06:55	nano-	Is it possible to run Clojure under any of the JDK8 compact profiles?
06:56	nano-	(building and trying is ofc one solution, but asking is probably a bit faster)
07:00	Anderkent	compact2 might work? dunno, just looking at the packages
07:00	Anderkent	of course all java-using libs you depend on would also have to run on it
07:00	clgv	llasram: the assert-args approach wont work
07:01	llasram	Oh?
07:03	clgv	llasram: e.g. (my-macro (let [x (* y y)] (conj res x))) is put into some loop-recur environment. now I get an exception for an unknown "res" but there is no line information associated except the one of the macro
07:03	clgv	but without a macro I'd get the line information of (conj res x)
07:05	llasram	Right. In your macro you'll need yourself to detect the problem, grab the :line metadata (in this case of the (conj res x) form), and throw an exception mentioning that line
07:05	llasram	Or, wait
07:05	clgv	well that's a huge effort then
07:06	llasram	Why does it not work to just expand to code mentioning `res`, then let the compiler throw a normal exception when `res` is undefined?
07:07	clgv	llasram: oh wait. you are right. seems it was only the error during macro expansion before where no appropriate line information was present
07:08	llasram	Awesome
07:11	clgv	thx :D
07:18	clgv	llasram: only the metadata on the first line of the map that I pass to the macro has a line number that is off by one
07:18	clgv	that's weird
07:23	clgv	llasram: within a map which is a macro parameter only lists get metadata?
07:25	llasram	clgv: I haven't fiddled with it enough to say off the top of my head. Would need to dig into LispReader.java
07:26	clgv	llasram: when printing with print-meta true it seems like that
07:30	clgv	so with the current state I'd need to use lists for let-like bindings to get the correct line metadata. :/
07:31	clgv	https://github.com/clojure/clojure/blob/master/src/jvm/clojure/lang/LispReader.java#L7273
07:33	Anderkent	clgv: I think only lists get line numbers in general I'm afraid
07:33	Anderkent	clgv: it's been a huge pain for me in cloverage
07:34	clgv	Anderkent: and probably no plans on changing that?
07:34	Anderkent	clgv: you could possibly try to pull out the line number of the wrapping form, then read the source of the wrapping function, then manually corellate forms to lines
07:35	clgv	Anderkent: I'd add it to all persistent data structure literals and symbols
07:35	Anderkent	if you ever do, give me a ping, I could use it :D
07:35	Anderkent	clgv: that'd be nice, wouldn't it
07:35	Anderkent	maybe tools.reader does it? not sure
07:35	clgv	Anderkent: you didn't create a jira ticket I can vote on? ;)
07:35	llasram	I believe tools.reader gives you line information for all IObjs
07:35	Anderkent	nope:P
07:35	llasram	Not that that helps clgv, but might help Anderkent
07:36	llasram	Anderkent: Hah, I need to type faster :-)
07:36	clgv	well, to have that line meta only on lists is probably an optimization I guess
07:40	mskou72	hi
07:41	mskou72	There's a lot of outdated docs (for 1.2 and 1.3) online. Where should a newbie go for docs and examples?
07:42	llasram	mskou72: 1.3 is OK. The changes since 1.3 have largely been incremental additions of new features
07:43	llasram	For functions which exist in 1.3 (the vast majority), clojuredocs.org is still an excellent resource
07:43	TheBrayn	https://gist.github.com/anonymous/8915135 this outputs ClassCastException java.lang.Long cannot be cast to clojure.lang.IFn my-stuff.core/eval1364/rec--1365 (form-init2646726723610872721.clj:5), why?
07:43	TheBrayn	it should emulate nth just with recursion
07:43	llasram	mskou72: For newer stuff and more detail, there are some good guides on http://clojure-doc.org/
07:44	llasram	mskou72: If you're into books, a new edition of /Joy of Clojure/ should be out any day now, and an advance edition is available now
07:44	mskou72	Ok, thanks :-)
07:45	llasram	TheBrayn: `((if ...))` -- you are calling the result of the `if` as a function
07:46	TheBrayn	oh I thought that was the function body
07:47	llasram	TheBrayn: Also note that Clojure doesn't automatically make functions tail-recursive, so this implementation will use a new stack frame for each recursion
07:47	TheBrayn	ok
07:48	llasram	You can fix that by calling `recur` in tail position instead of re-calling the recursive function by name
07:48	TheBrayn	cool, thanks
07:49	TheBrayn	I ofcourse made this much more complicated than needed but it's a good excercise nonetheless
07:59	sm0ke	whats the right way to transform maps in clojure
07:59	sm0ke	(into {} (map ... is ugly
08:01	llasram	sm0ke: A `map-vals` utility function is a fairly common solution
08:04	sm0ke	yea, probly should do that
08:05	jowag	,wat
08:05	clojurebot	eval service is offline
08:05	jowag	I am very sorry
08:05	llasram	,1
08:05	clojurebot	eval service is offline
08:05	llasram	Huh
08:05	sm0ke	does a closure over a transient also creates a copy?
08:06	llasram	sm0ke: It holds a reference to the transient, if that's what you mean.
08:06	sm0ke	llasram: yes?
08:06	jowag	##foo
08:06	daGrevis	is there a fn that would allow me to get nth elem from a seq and modified seq without that nth elem?
08:06	sm0ke	llasram: in case of immutable a copy is made right
08:07	llasram	No
08:07	llasram	Always references
08:07	sm0ke	hmm
08:07	sm0ke	really?
08:07	sm0ke	wouldnt that defeat the purpose of closures
08:07	llasram	The immutable types work by disallowing mutation, and returning structure-sharing copies on modification
08:07	TheBrayn	daGrevis: of course
08:07	TEttinger	daGrevis, like pop but with an index?
08:07	sm0ke	llasram: makes sense
08:08	clgv	"ClassCastException clojure.core.Vec cannot be cast to clojure.lang.IEditableCollection clojure.core/transient" - I thought I had a vector there. what is c.c.Vec?
08:08	sm0ke	if closures are internally class instances thats how it should work
08:08	sm0ke	hmm now that is think of it i am scared
08:08	llasram	?
08:09	sm0ke	so many places i used partial functions in my code without worrying about it
08:09	jowag	so, where do I report a bug in clojurebot?
08:09	llasram	daGrevis: Not that I'm aware of. Maybe with zippers? It's difficult to do without mutation
08:09	llasram	sm0ke: What are you worried about?
08:09	sm0ke	llasram: that the values i closed over were mutable
08:10	sm0ke	crap!
08:10	llasram	Well, are they? :-)
08:10	llasram	jowag: I believe hiredman is the person operating clojurebot
08:10	TheBrayn	why not just return a tuple of the nth element + concat of take and take-last?
08:11	jowag	llasram: thanks
08:11	llasram	jowag: https://github.com/hiredman/clojurebot
08:11	daGrevis	,(nth 3 [1 2 3 4 5])
08:11	clojurebot	eval service is offline
08:11	daGrevis	,(pop 3 [1 2 3 4 5])
08:11	clojurebot	eval service is offline
08:12	TEttinger	##(pop 3 [1 2 3 4 5])
08:12	lazybot	clojure.lang.ArityException: Wrong number of args (2) passed to: core$pop
08:12	daGrevis	TEttinger, kike pop with index, y
08:12	llasram	TheBrayn: You can, but the new sequence is actually a nested collection of references to the original sequence
08:12	TEttinger	##(pop [1 2 3 4 5])
08:12	lazybot	⇒ [1 2 3 4]
08:12	daGrevis	ok, lets start from the beginning. i have a vector
08:13	daGrevis	and i want to divide it in three parts
08:13	llasram	TheBrayn: If you keep doing that without fulling realizing the seqs, you'll just end up getting a deeper and deeper nesting
08:13	TEttinger	like partition or split?
08:13	daGrevis	[half of it] middle element [other half]
08:13	TEttinger	^ daGrevis
08:14	TEttinger	ah ok
08:14	Anderkent	llasram: tools reader would be nice, but I'm currently considering hooking into clojure.core/load-lib; I suppose I could hook in and replace RT/load with tools.reader/read followed by eval
08:14	Anderkent	but that feels scary
08:15	clgv	,(type (vec (range 10)))
08:15	clojurebot	eval service is offline
08:15	clgv	:/
08:15	clgv	&(type (vec (range 10)))
08:15	lazybot	⇒ clojure.lang.PersistentVector
08:15	TEttinger	##(let [midpoint (fn [idx coll] [(take idx coll) (nth coll idx) (drop (inc idx) coll)])](midpoint 3 [1 2 3 4 5]))
08:15	lazybot	⇒ [(1 2 3) 4 (5)]
08:15	daGrevis	i got middle index using quot. i can get it as elem using nth. no idea how can i get vector as first and second part
08:15	Anderkent	&(type (vec (int-array (range 10))))
08:15	lazybot	⇒ clojure.lang.PersistentVector
08:16	TEttinger	##(let [midpoint (fn [idx coll] [(vec (take idx coll)) (nth coll idx) (vec (drop (inc idx) coll))])](midpoint 3 [1 2 3 4 5]))
08:16	lazybot	⇒ [[1 2 3] 4 [5]]
08:16	daGrevis	TEttinger, correct answer should be [(1 2) 3 (4 5)]
08:16	Anderkent	daGrevis: 1 indexing!?
08:16	daGrevis	Anderkent, ?
08:17	Anderkent	daGrevis: (midpoint 3 [1 2 3 4 5]) should surely give you 4?
08:17	TEttinger	no... the index for 4 is 3. unless you want to split by value
08:17	Anderkent	oh
08:18	TEttinger	##(nth [1 2 3 4 5] 3)
08:18	daGrevis	i know! i can use split-at 2 coll and then rest on second elem
08:18	lazybot	⇒ 4
08:18	TEttinger	sure, you'd just need to restructure it
08:20	TEttinger	also, the way split-by is implemented is exactly how I did it
08:21	daGrevis	cool ^^
08:22	TEttinger	wait do you want the middle element to only be in the center? like balanced middle?
08:24	TEttinger	##(let [midpoint (fn [coll] (let [idx (quot (count coll) 2)] [(vec (take idx coll)) (nth coll idx) (vec (drop (inc idx) coll))]))](midpoint [1 2 3 4 5 6 7]))
08:24	lazybot	⇒ [[1 2 3] 4 [5 6 7]]
08:25	Anderkent	not sure about that, doesnt handle even collections well
08:25	Anderkent	the one with an explicit index does what you'd expect every time :P
08:26	daGrevis	TEttinger, yes, balanced middle
08:26	TEttinger	Anderkent, finding the exact center of an even-count collection isn't defined anyway right?
08:27	daGrevis	do clojure programs tend to call functions get-foo?
08:27	daGrevis	do clojure programs tend to call functions get-foo ?
08:27	Anderkent	TEttinger: yeah, that's why I'd rather have something that does something well defined for every input :P I'm just nitpicking at this point, I guess.
08:27	TEttinger	kinda
08:27	Anderkent	sorry :P
08:27	TEttinger	it's fine
08:28	TEttinger	daGrevis, people tend to use dashes in names, but single-word names are preferred when there's a good word for it
08:28	TEttinger	partition vs. cutAtIntervals
08:33	pyrtsa	Any idea why vectors, sets and maps aren't defining clojure.lang.ISeq? The interface itself (first/next/more/cons) would be easy to implement, of course not retuning the original type.
08:33	pyrtsa	s/defining/implementing/
08:37	pyrtsa	Or, in other words, what't the rationale of having `(seq? [])` return false?
08:37	Anderkent	pyrtsa: not sure if ISeq promises constant time next(), but if so that'd disqualify vectors
08:38	pyrtsa	Oh? Because of O(log32(N))?
08:38	Anderkent	hm, well, I guess you can make a wrapper vector that'd just translate every access to n+1
08:39	Anderkent	so I guess that's not strictly true.
08:39	pyrtsa	FWIW, (seq xs) is defined for vectors, as well as (first xs), (next xs), (cons xs).
08:39	minikomi	##((fn midpoint [idx v] (let [h (subvec v 0 idx) t (subvec v idx)] [h (first t) (subvec t 1)])) 5 (vec (range 15))) ; using subvec might be quicker?
08:39	lazybot	⇒ [[0 1 2 3 4] 5 [6 7 8 9 10 11 12 13 14]]
08:40	pyrtsa	I mean, (cons x xs).
08:40	Anderkent	pyrtsa: seq first next on vectors will convert it to a seq first, I think?
08:40	pyrtsa	Sure. But the conversion is shallow, of course.
08:43	clgv	pyrtsa: I think the relevant distinction here is between Seq and Seqable
08:43	TEttinger	##(seq? [])
08:43	lazybot	⇒ false
08:44	clgv	pyrtsa: vector, map, set are no Seq but can be turned into one by using `seq`
08:44	pyrtsa	But why such distinction? I know about clojure.core.incubator/seqable?, and to be honest, its definition creeps me somewhat. :)
08:48	clgv	well you can ask the question the other way round why should they implement ISeq? I am not sure if any of the books I read explained the reasons therefore in details
08:49	Anderkent	it seems a useful thing to know, whether your collection is a seq or a different data type - it changes what things like conj do
08:49	clgv	but it is pretty consistent way of marking sequences that do not need any conversion to be used by sequence functions and those data structures which need to be turned into a seq first
08:49	Anderkent	if you don't care, you can just (seq thing) and know it's a seq now.
08:50	pyrtsa	Nod.
08:50	clgv	probably JVM is a reason as well, since eah specific class knows best how its most efficient sequence implementation has to look like
08:51	circ-user-HInXe	nick ajs
08:51	clgv	^^ as a reason for Sequable
08:54	pyrtsa	Looking back at the code I've written, there seem to be two uses for `seqable?`: 1) to verify early (as a :pre condition) if an argument is valid, and 2) to make a function behave differently for Seqables (many values) and non-Seqables (one value).
08:55	clgv	pyrtsa: the one-value case is the one I encountered often when visualizing data
08:56	clgv	it's the case where you'd kill for (atomar? x) to return true for non-collections/sequences ...
08:56	Anderkent	I never needed sequable?, but I needed list? that works; that turns out to be seq?
08:56	pyrtsa	clgv: `atomar?` meaning `(complement seqable?)`, you mean?
08:59	blrm	join #robottelo
09:00	blrm	sorry, mistype :)
09:00	clgv	pyrtsa: returning true for single values so to say as oposed to data structures with multiple values ;)
09:00	clgv	list? and vector? are often handy for macros
09:01	pyrtsa	Yeah, that's what the above is.
09:01	pyrtsa	I use vector? a lot in macros.
09:01	clgv	does seqable? return true for seqs?
09:01	Anderkent	clgv: depends on how you implement it
09:01	pyrtsa	It should.
09:01	Anderkent	clgv: my preferred implementation would be ,(try (seq x) true (catch IllegalArg... false)) :P
09:01	pyrtsa	clgv: It does.
09:01	clgv	Anderkent: yeah I meant the mentioned implementation ;)
09:02	pyrtsa	,(clojure.core.incubator/seqable? (seq '(1 2 3)))
09:02	clojurebot	eval service is offline
09:02	pyrtsa	Oh.
09:02	Anderkent	&(clojure.core.incubator/sequable? (seq '(1 2 3)))
09:02	lazybot	java.lang.RuntimeException: No such var: clojure.core.incubator/sequable?
09:02	Anderkent	oh well
09:02	clgv	&(require 'clojure.core.incubator)
09:02	lazybot	⇒ nil
09:03	clgv	&(clojure.core.incubator/sequable? (seq '(1 2 3)))
09:03	lazybot	java.lang.RuntimeException: No such var: clojure.core.incubator/sequable?
09:03	pyrtsa	https://github.com/clojure/core.incubator/blob/master/src/main/clojure/clojure/core/incubator.clj#L83-92
09:03	clgv	old version ;)
09:03	cleatoma	&(clojure.core.incubator/sequable? (seq '(1 2 3)))
09:03	lazybot	java.lang.RuntimeException: No such var: clojure.core.incubator/sequable?
09:03	cleatoma	Oops
09:03	cleatoma	&(clojure.core.incubator/seqable? (seq '(1 2 3)))
09:03	lazybot	⇒ true
09:04	Anderkent	is there a way of making google include the ? in 'list?'. Verbatim doesnt do it
09:04	Anderkent	I remember there was some issue with list?, where the reader could construct a list that wasnt a list?
09:04	Anderkent	but can't remember the details
09:05	scape_	,(byte 129)
09:05	clojurebot	eval service is offline
09:05	llasram	Anderkent: Hmm. I don't remember the reader constructing non-list? sequences, but you can certainly e.g. return one from a macro just fine
09:09	Anderkent	hm, maybe I'm misremembering. Can't find it now ;/
09:11	daGrevis	naive implementation of quicksort https://gist.github.com/daGrevis/8862352
09:26	scape_	,(byte 129)
09:26	clojurebot	eval service is offline
09:27	Anderkent	scape_: it'll throw an exception afaik
09:27	Anderkent	&(byte 129)
09:27	lazybot	java.lang.IllegalArgumentException: Value out of range for byte: 129
09:27	scape_	why do I see (byte) 129 in java?
09:28	scape_	http://stackoverflow.com/a/12471677/1298523
09:28	Anderkent	&(unchecked-byte 129)
09:28	lazybot	⇒ -127
09:28	Anderkent	is what you want, I suppose
09:28	scape_	oh
09:28	scape_	they're signed? I know so little about bytes :-\
09:28	Anderkent	everything's signed in java AFAIK
09:29	scape_	thx let me work with that, it's supposed to be -127 too, so that must be it
09:49	stuartsierra	Many Java APIs return ints to represent bytes.
09:51	gfredericks	stuartsierra: do you think that's because of signedness issues or because ints are more first-class in the bytecode?
09:52	stuartsierra	dunno
09:52	Anderkent	I'd think signedness; let the compiler worry about the byte code :P
09:52	scape_	I was confused because this is a typecast I found in a java example: (byte) 129;
09:55	Anderkent	when you're working with a network protocol that's based on unsigned bytes, you probably want your constants to be positive; casting with (byte) just chops off the high bytes, so the bit pattern shuold be the same
09:56	Anderkent	&(bit-and (int (unchecked-byte 127)) 0xff)
09:56	lazybot	⇒ 127
09:56	evantravers	esh
09:57	scape_	it's curious why websockets were designed the way they were-- the entire bit masking concept seems silly, but I sort of understand why they set the frames up the way they did-- bc javascript is not threaded I assume
09:57	Anderkent	obviously 127 was a bad example
09:57	gfredericks	,[(byte 127) (byte 278)]
09:57	clojurebot	eval service is offline
09:57	gfredericks	&[(byte 127) (byte 278)]
09:57	lazybot	java.lang.IllegalArgumentException: Value out of range for byte: 278
09:57	gfredericks	&[(unchecked-byte 127) (unchecked-byte 278)]
09:57	lazybot	⇒ [127 22]
09:58	gfredericks	I'm bad at this
09:58	gfredericks	&[(unchecked-byte 129) (unchecked-byte 278)]
09:58	lazybot	⇒ [-127 22]
09:58	Anderkent	not sure where you're going
09:59	gfredericks	I'm done going there and it wasn't worth the effort
09:59	Anderkent	haha :D
09:59	scape_	hah
09:59	gfredericks	I was confirming to myself that the sign of the output was just related to the 8th bit, not any higher bits
09:59	Anderkent	obviously if you get n > 255 then no magic will make it work with a sign byte
10:00	clgv	two-complement ;)
10:37	clgv	&(transient (vector-of :int))
10:37	lazybot	java.lang.ClassCastException: clojure.core.Vec cannot be cast to clojure.lang.IEditableCollection
10:37	clgv	will that be fixed?
10:38	Anderkent	eventually probably?
10:38	Anderkent	http://dev.clojure.org/jira/browse/CLJ-787
10:39	Anderkent	hm, not sure if that's exactly the same thing
10:39	Anderkent	but it's similar enough
10:39	Marc_	auth
10:40	Guest43154	?
10:40	clgv	Anderkent: at least it is vetted
10:41	TheBrayn	what was the function called that took multiple functions as an argument and applied them to a value?
10:41	AimHere	juxt?
10:41	clojurebot	juxt is completely uncontroversial in its splendour
10:42	Anderkent	&((juxt inc dec) 2)
10:42	lazybot	⇒ [3 1]
10:43	TheBrayn	like: (<func> f g k x) which does (f (g (k x)))
10:43	AimHere	That sounds more like comp
10:43	TheBrayn	yeah that's what I meant, thanks
10:44	clgv	&((comp dec inc) 0)
10:44	lazybot	⇒ 0
10:44	Anderkent	#wastingcycles
10:44	Anderkent	:P
10:44	TheBrayn	#yolo
10:44	AimHere	Maybe the compiler optimized away the inc/dec
10:44	gfredericks	&((apply comp (repeat 1000000 inc)) 42)
10:45	clgv	AimHere: very unlikely
10:45	lazybot	Execution Timed Out!
10:45	gfredericks	&((apply comp (repeat 100000 inc)) 42)
10:45	lazybot	⇒ 100042
10:46	AimHere	,((apply juxt (repeat 100000 inc)) 42)
10:46	clojurebot	[43 43 43 43 43 ...]
10:49	luxbock	anyone able to point out what I'm doing wrong with the following Instaparse: https://gist.github.com/luxbock/8918050
10:50	luxbock	I keep running into this same error with the expected parentheses "(" at the end of the file
10:52	MarcFromNYC	can someone tell me what's wrong with this macro:
10:52	MarcFromNYC	(defmacro foo [x] `(let [y# (map inc x)] (prn ~@y#)))
10:52	MarcFromNYC	Results in CompilerException java.lang.RuntimeException: Unable to resolve symbol: y# in this context
10:53	michaniskin	MarcFromNYC: why do you unquote y#?
10:53	tim____	MarcFromNYC: you don't need to unquote a auto generated sym
10:53	AimHere	Isn't that because you have an unquoted y# there?
10:53	MarcFromNYC	i want to splice it in
10:53	Anderkent	MarcFromNYC: you splice things that are calculated in the macro, not things that are in the code that the macro generates
10:53	MarcFromNYC	this is a simple case, i could apply in this case, but what if prn were a macro instead.
10:54	Anderkent	MarcFromNYC: you can't splice things that come form runtime, you don't know what could be inside x
10:54	Anderkent	unless it's a constant list/vector, in which case you can do
10:55	Anderkent	(defmacro foo [x] (let [y (map inc x)] `(prn ~@y)))
10:55	MarcFromNYC	Anderkent: thx. i'll try that.
10:58	jonathanj	,(help juxt)
10:58	clojurebot	#<CompilerException java.lang.RuntimeException: Unable to resolve symbol: help in this context, compiling:(NO_SOURCE_PATH:0:0)>
10:58	Anderkent	luxbock: try running with :total true and make sure it's what oyu'd expect?
10:58	Anderkent	Also I get different errors with the gist you posted (Expected "\"")
11:00	luxbock	running it with :total true gives me: [:project-alist [:projects]] so I guess I'm doing multiple things wrong
11:00	luxbock	I think I need to spend some more time trying to understand how context free grammars work to begin with
11:02	luxbock	I'm trying to capture 'name' 'kw' and 'val'
11:02	Anderkent	luxbock: for me it fails on the name, since you eat the closing double quote in your regex
11:02	luxbock	ah yeah
11:02	Anderkent	luxbock: you should be really careful about using regexes in instaparse; they don't backtrack
11:04	luxbock	yeah I think I understand what I was doing wrong now
11:07	Anderkent	luxbock: yeah getting them right can be a little tricky - took me at least a day to make https://github.com/JacekLach/doctest/blob/master/resources/doctest/doctest.grammar work :P
11:09	luxbock	yeah for me this is a simplified case as the actual config can contain comments and other expressions before and after
11:10	luxbock	doesn't seem like there's any easier way though
11:12	scape_	if anyone is savvy with websockets, I'm trying to replicate this but am having a hard time: http://stackoverflow.com/questions/8125507/how-can-i-send-and-receive-websocket-messages-on-the-server-side/12471677#12471677 with this: https://gist.github.com/viperscape/8918565
11:15	scape_	am getting array oob :-\
11:19	Anderkent	scape_: inputStream.read() gives you an int, not a byte. Why do you cast mlen?
11:20	arrdem	,(println "arrdem: ping") ;; testing sauron-mode
11:20	clojurebot	arrdem: ping\n
11:20	Anderkent	wait, are you trying the writing or the reading
11:20	scape_	decode/read right now. and your right, it does
11:20	scape_	the java example on that SO post is so-so
11:21	Anderkent	also no need to do the (unchecked-byte 127), you can give an int (0xff) to bit-and
11:22	scape_	,0xff
11:22	clojurebot	255
11:22	Anderkent	oups
11:22	Anderkent	my bad :P
11:22	scape_	:D
11:23	Anderkent	also I misread, your mlen corresponds to rLength in the code, not len?
11:23	Anderkent	*in the java code
11:23	scape_	yes, the pseudo code above it, or the python and javascript below might be easier to read.
11:24	scape_	mlen is not the right name for it, i should rename that to be honest
11:24	scape_	but the purpose is to determine the mask start position
11:25	scape_	the error is somewhere in the loop i believe
11:25	Anderkent	scape_: what's the type of data? byte array?
11:25	scape_	byte array
11:27	amalloy	scape_: why are you implementing this yourself at all? it sounds like just a websocket protocol implementation, and those exist in clojure already
11:28	scape_	bit of a learning thing, but also I don't know of a clojure websocket library-- it's all from netty from what I've seen; I was interested in integrating ws into my socket server. I'm just about there
11:29	scape_	Anderkent: think I figured it out :)
11:29	Anderkent	scape_: oh, what was it?
11:29	Anderkent	off by one in the array length?
11:29	scape_	I read the buffer at a fixed size and didn't resize the array
11:29	scape_	:D
11:30	Anderkent	ah
11:30	scape_	so it's passing garbage after the content
11:31	Anderkent	right, so pass the length of data or trim the aray; I'd also consider using amap instead of manually looping
11:31	Anderkent	amap / areduce are so nice
11:31	amalloy	scape_: there's a websocket client/server in aleph, and http-kit has one too, though i haven't used that one
11:31	scape_	i'll look in to that, yea the loop is ugly but wanted to figure it out first
11:31	scape_	both use netty i believe
11:32	amalloy	why is that a problem?
11:32	scape_	i'm not against netty, just that I had something for sockets already-- now I have it accepting both raw and websocket clients simultaneously, kinda slick I think
11:33	mikerod	If a fn takes code (as a data structure) and wants to return a callable fn that has that code embedded within; is the only way to do this with an eval?
11:33	mikerod	(not using a macro)
11:34	Anderkent	mikerod: yeah, if oyu take code as data you will have to eval it somewhere
11:34	amalloy	mikerod: yeah, but like...that's generally evil
11:34	mikerod	(defn i-take-code [s-expr] (fn [x y z]
11:34	mikerod	(s-expr)))
11:34	mikerod	Anderkent: that's what I was thinking
11:34	mikerod	amalloy: I figured...
11:34	amalloy	it's hard to imagine that being a good idea, really. either it should be a macro, or it should take a function instead of an expression
11:37	Anderkent	scape_: untested, but might plausibly work: https://www.refheap.com/35945
11:37	Anderkent	I'd say 50-50 that it works as is :P
11:38	mye	Hi. Is an infinite core.async (go (loop ... GC'd when the channel it returns is closed?
11:39	Anderkent	mye: no, the writes will just start returning false afaik
11:40	amalloy	Anderkent: i think his english misled you: he means "when is the channel it returns closed?"
11:41	amalloy	oh
11:41	mye	Anderkent: do you mean 'reads'? Can a go block read from the channel it returns?
11:41	amalloy	no he didn't. sorry, too early for me
11:41	Anderkent	is it? I understood 'will a (go (loop ...)) block automaticall stop and GC when its channel is closed
11:41	mye	amalloy: mostly trying to figure out how not to leak an unlimited number of (go)
11:42	Anderkent	mye: ah, sorry; I misunderstood too. I forgot go blocks return channels
11:43	Anderkent	mye: I'd expect it to collect eventually after it writes a value to the channel it returned. I don't think you have to read it for it to clean itself up, but not sure there.
11:44	Anderkent	[I was thinking of a (go (loop (>! some-third-channel a-value))), in which case you have to look at the result of your write to know if your channel was closed and you should exit]
11:44	mikeyg6754	Hello, I'm working on processing a POST request with liberator. I am also using http-kit and compojure, when I get the response the body is: :body #<BytesInputStream BytesInputStream[len=8897]>. I'm not sure how to get :body as a map.
11:45	mye	My design problem is a service abstraction that keeps taking requests from its queue. To do that I let a (go) spin on the service queue and spawn other (go)'s that handle the requests
11:45	stuartsierra	Go blocks should be GC'd when they are no longer reachable, such as blocked on a channel which has been GC'd. Anything else is a bug.
11:45	chronno	mye: according to a comment from tbaldridge (https://groups.google.com/forum/#!topic/clojure/X6JoXczcRYw) all the go blocks get GC'ed when the channel is
11:45	Anderkent	mikerod: is your request content-type 'application/x-www-form-urlencoded' or something like that?
11:51	mikeyg6754	Anderkent: tried changing the content type. That didn't work.
11:52	Anderkent	mikeyg6754: Hm. Weird, I think there's a ring middleware that handles form-encoded bodies and puts that into params
11:52	mikeyg6754	Anderkent: I have wrap-params and wrap-file-info currently
11:55	Anderkent	mikeyg6754: other than double-checking your content type, no idea. wrap-params should definitely consume the body and assoc :form-params and :params
11:55	Anderkent	see https://github.com/mmcgrana/ring/blob/master/ring-core/src/ring/middleware/params.clj
11:56	mikeyg6754	Anderkent: hmmm, maybe it's something http-kit is doing. This is my first time working with it.
11:57	no7hing	anybody tried accessing a scala Tuple2 from clojure?
11:57	no7hing	(._1 tuple) should work
11:57	Anderkent	mikeyg6754: I'd open the ring jar in <editor-of-choice> and add printouts in relevant functions; see if the middleware is being called at all and if so why it's not doingX
11:58	ptcek	Hello, I am trying to fetch lazily a resultset from a database using clojure.java.jdbc to process huge dataset. I tried to (take 10 (query ...)) but it loads the whole set, even after setting :fetch-size. Anyone know what am I doing wrong? Details here: http://pastebin.com/4XnsEcqE
11:58	mikeyg6754	Anderkent: Thanks, I appreciate your help!
12:00	Anderkent	ptcek: query takes a :result-set-fn function that runs over the entire result set, and is vec / doallb y default
12:00	Anderkent	ptcek: you probably want to supply (parital take 10) as :result-set-fn so that you only get 10 results?
12:01	ptcek	Anderkent: I tried exactly that with this error: SQLRecoverableException Closed Resultset: next oracle.jdbc.driver.OracleResultSetImpl.next (OracleResultSetImpl.java:229)
12:03	Anderkent	ptcek: try (doall (take 10))
12:03	Anderkent	ptcek: I think because take is lazy it's escaping the context where your db connection is open
12:03	Anderkent	thus when it tries to realize later on the conn is closed and you get that exception
12:03	luxbock	Anderkent: I got it working, woo
12:04	Anderkent	luxbock: yay :)
12:08	ptcek	Anderkent: doall takes coll as and argument so I tried :result-set-fn #(doall (take 10 %)) and it's working!
12:09	Anderkent	ptcek: yeah, that's what I meant :)
12:09	ptcek	Anderkent: I played with it the whole day and now you have it in 2 minutes. Thank you very much Anderkent!
12:10	Anderkent	that's what we're here for (other than procrastinating on actual work, that is)
12:13	mikerod	amalloy_: sorry I was in a meeting; I figured the feedback I'd hear would be "it is not a good design/idea" :)
12:14	mikerod	amalloy_: I was thinking in a scenario where I'm parsing a DSL sort of syntax and generating sort a map of data on it. I tend to minimize the use of macros and make most things behave just as simple functions.
12:15	mikerod	It seems natural to me to have a fn that took code-as-data, and wanted to do some processing to it and return some callable code the code-as-data input. Then again, I know this sort of just sounds like describing a macro I guess.
12:15	Anderkent	yep, exactly like a macro :P
12:17	mikerod	Anderkent: yeah... when described like this it does; so I guess I'll try to dig at what I'm missing here.
12:17	mikerod	I appreciate the feedback.
12:19	ambroseb_	tpope: if I figure out how to do it, would you consider a vim-fireplace patch to preserve line/column/source information when Eval'ing with a motion?
12:33	Raynes	Man, the rate of pastes on refheap these days is almost worrisome. :p
12:34	Raynes	I'm might have to make refheap webscale or something!
12:34	Raynes	;)
12:36	Anderkent	clearly got to switch to node & mongodb now
12:36	Anderkent	I've noticed that refheap occasionally takes forever to 'establish secure connection' as google tells me
12:36	Anderkent	though not as common recently
12:39	TimMc	Raynes: You need to start using http://www.supersimplestorageservice.com/
12:39	Raynes	TimMc: Clearly
12:40	RickInAtlanta	llasram: I just sent you an email but I thought it might be faster to reach you here.
12:41	ecounysis	a
12:42	seangrove	It's too bad the programmers aren't allowed to learn a new language after picking up their first one.
12:43	technomancy	seangrove: something I read somewhere "Your second language can be anything, as long as it makes you ask what the hell was wrong with your first one."
12:44	seangrove	technomancy: So, ruby => python then, clearly?
12:44	technomancy	seangrove: all I'm saying is that all these sites teaching JS to new programmers is starting to make sense
12:44	TimMc	technomancy: TI-89 Basic -> Java, checks out.
12:44	TimMc	haha
12:45	seangrove	technomancy: Yeah, I'm certainly fine with it. Just saddened by the reasoning behind this comment https://news.ycombinator.com/item?id=7211623
12:45	technomancy	welp.gif
12:45	seangrove	It's neglecting another option, which is, "You've learned to be a good developer, and you pick up a new language on the job."
12:47	mikerod	seangrove: exactly
12:48	mikerod	However, sounds like JavaScript is the answer to everything
12:48	mikerod	what a perfect language
12:48	technomancy	«I'll take "languages designed in under 10 days" for $200, Alex.»
12:48	seangrove	"How do you take your coffee?" "JavaScript"
12:48	TheBrayn	nah that would be coffeescript, right?
12:49	mikerod	it's so easy to learn JavaScript and JQuery inside and out
12:49	mikerod	and then you are an unstoppable programming force
12:49	locks	technomancy: you're just bummed that Clojure isn't a Lisp
12:49	mikerod	maybe throw in a framework or 2
12:49	technomancy	locks: gonna start a folk of clojure with support for every data type that isn't a list removed. that'll show em!
12:50	locks	ha!
12:50	seangrove	mikerod: Seriously, what can't be accomplished with Javascript, jQuery, and an unlimited amount of energy, time, and blood letting?
12:50	rasmusto	technomancy: folk lore
12:50	mikerod	seangrove: exactly, endless possibilities
12:50	locks	you can't let blood unless you're using ES6 seangrove
12:50	locks	I'll be here all week.
12:50	teslanick	instantrimshot
12:51	mikerod	I wnat to make a language with only abstractions; and no concrete implementations
12:51	mikerod	With this language, you cannot actually do anything at all
12:51	seangrove	mikerod: I hesitate, but... Haskell?
12:52	mikerod	seangrove: eh, maybe close; I guess someone beat me to it
12:52	seangrove	They may have erred on the can-do-a-little-bit side though
12:52	locks	prolog?
12:52	locks	apl?
12:53	mikerod	:P
12:53	seangrove	locks: clojure?
12:53	seangrove	mikerod: You though it was a joke, but actually it's a highly-contested prize in a very competitive field
12:53	locks	seangrove: clojure has implementations, no?
12:54	teslanick	Only when you (doall)
12:54	locks	I mean, it's just Java
12:54	seangrove	locks: Yeah, didn't want to come off as bigoted, so it's appropriate to criticize our own language as well. Even when it's inappropriate.
12:54	locks	;P
12:55	teslanick	"I'm not bigoted, I hate everyone equally!"
12:55	jcrossley3	RickInAtlanta: i'm guessing we're rescheduling tomorrow's meetup, right? ;)
12:55	RickInAtlanta	jcrossley3: yeah, I am trying to reach Marshall now, to figure out when
12:56	RickInAtlanta	expect a notice shortly
12:56	jcrossley3	cool
12:56	mikerod	seangrove: fair enough
12:56	RickInAtlanta	I am going to be AFK for a bit, so if you see a message from him in channel, ask him to reschedule :)
12:56	jcrossley3	RickInAtlanta: will do
13:14	llasram	jcrossley3, RickInAtlanta: the snow isn't supposed to start until midnight, is it?
13:15	jcrossley3	llasram: tonight, i thought
13:16	llasram	I'm just seeing overcast tonight, rain tomorrow, and then snow starting around midnight
13:18	llasram	Bleeergh
13:20	muhoo	meh, cofeescript looks like javascript with enough sugar to make it palatable to ruby programmers
13:22	TimMc	muhoo: Unfortunately, you can't really take full advantage of the syntactic transforms without second-guessing the compiler.
13:22	technomancy	muhoo: dunno, ruby programmers are pretty conditioned to hate whitespace-sensitivity through python flamewars
13:22	TimMc	I keep having to check the output to make sure the nesting is correct.
13:22	RickInAtlanta	llarsam: I thought snow was midnight tonight
13:24	SegFaultAX	technomancy: I hate significant whitespace and I've been doing Python for almost 10 years.
13:24	muhoo	i'm slowly approaching cljs/om from the side, and like what i've seen so far.
13:24	rasmusto	SegFaultAX: ^I^I hate it too
13:25	RickInAtlanta	llarsam: If you want to wait until tomorrow to see if we need to reschedule, that is fine.
13:25	SegFaultAX	My reason is simple: while the code may be superficially pretty, it restricts what the syntax can admit to such an extent that super useful things (multi-line lambdas, Ruby-style blocks, etc.) simply aren't possible.
13:26	muhoo	advantage to significant whitespace is at least it cuts off indenting flamewars at the knees. Indentation styles: There Can Be Only One.
13:26	technomancy	SegFaultAX: being unable to automate re-indentation after a merge conflict soundsn like a nightmare too
13:26	SegFaultAX	Not that Python would have those things if the syntax /could/ support it, but it isn't even an option.
13:26	muhoo	(speaking as someone who has sat in all-day meetings debating c indenting styles)
13:26	SegFaultAX	technomancy: And that, but you get used to it after doing it for a long time.
13:26	rasmusto	SegFaultAX: multi-line list comprehensions are a nightmare
13:26	technomancy	muhoo: doesn't python still have wars over 2-space vs 4-space vs tab?
13:26	SegFaultAX	rasmusto: Yes! Especially when nested.
13:26	melba	muhoo, we can argue 4 spaces vs 2 spaces
13:27	rasmusto	[an eye for an eye]
13:27	SegFaultAX	pep8 says 4
13:27	muhoo	technomancy, melba: hmm, good call. i guess there is no end to flamewars as long as humans are involved.
13:27	rasmusto	I had a wonderful codebase with combination 5 space and ^I
13:27	melba	:D
13:27	SegFaultAX	Most people I think try to at least start at pep8, so that adds some consistency.
13:27	melba	i use 2
13:27	technomancy	muhoo: mostly been able to avoid it in lisp-land thankfully
13:27	rasmusto	I can't tell you how much time I wasted plugging a pep8 linter into my editor and making terrible whitespace commits :(
13:27	SegFaultAX	melba: I prefer 2, but my current company uses [mostly] pep8, so 4 it is.
13:28	muhoo	i should write a language which uses significant smiely-face emojiis
13:28	SegFaultAX	melba: I had to make a special case in my vimrc just to accomodate. :)
13:28	rasmusto	if the language can't handle ggVG= it makes me sad now
13:28	jcrossley3	llasram: according to weather.com, the rain/snow will start tomorrow morning, with a "winter storm watch" commencing about the time you take the podium :)
13:28	SegFaultAX	rasmusto: Then don't use python.
13:29	llasram	jcrossley3: Fiiiiine
13:29	rasmusto	SegFaultAX: I avoid it when I can, and write not-very-nested stuff when I can't
13:29	jcrossley3	llasram: :)
13:29	rasmusto	SegFaultAX: not_very_nested*
13:29	muhoo	technomancy: occasionally i get someone's clojure code that isn't done in clojure-mode in emacs, and i do M-x indent-buffer and all of a sudden the git commit is polluted with a bunch of indenting changes, but not often.
13:30	RickInAtlanta	llasram: jcrossley3: it seems to matter a lot where you are. I checked weather in brookhaven, and I was thinking we would be ok. Flowery Branch, otoh is going to suck.
13:30	technomancy	muhoo: it happens, but most people who do that understand that they're wrong once you explain things =)
13:30	rasmusto	if only vcs could look at code structure and not whitespace :(
13:30	muhoo	~paredit
13:30	clojurebot	paredit is not for everyone, but what you need to understand ís how to become the kind of person that paredit ís for.
13:30	SegFaultAX	rasmusto: Avoiding nesting in general is good, IMHO. Of course it's easier to nest if you have delimited blocks (or some other syntactic form of nesting like sexps) though.
13:31	mynomoto	dakrone: Found an old thread about clojuredocs.org in the group. How it the new version going on?
13:31	llasram	RickInAtlanta: I'm convinced. Let me see if I can get the space next week
13:32	dakrone	mynomoto: not much progress on it currently, clojure/core was going to do some kind of work with the data too (haven't heard any feedback about it yet)
13:32	Rick_AFK	yeah, next week would be great. If not, I suggest delaying till next month, since we do have the evening with Stuart Sierra later this month.
13:32	Rick_AFK	thanks!
13:34	mynomoto	dakrone: I want to contribute, is there something that I can do?
14:00	jowag	,(+ 0.1 0.2)
14:00	clojurebot	0.30000000000000004
14:00	`cbp	,:-)
14:00	clojurebot	:-
14:02	TimMc	,(+ 0.1 0.2 -0.3)
14:02	clojurebot	5.551115123125783E-17
14:03	AimHere	(+ 1.7 0.1)
14:03	AimHere	m(+ 1.7 0.1)
14:03	rplaca	welcome to floating point math! :)
14:03	AimHere	,(+ 1.7 0.1)
14:03	clojurebot	1.8
14:03	AimHere	Boo, was a worse error in my repl
14:03	rasmusto	##(+ 1.7 0.1)
14:03	lazybot	⇒ 1.8
14:22	hombotto	Hi, I'm having problems with midje :autotest, any help?
14:28	rasmusto	~anyone
14:28	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 ..."
14:28	mskou72	Why does this result in out of memory (map println (lazy-seq (file-seq (clojure.java.io/file "/Users/martin")))) when this does not (count (lazy-seq (file-seq (clojure.java.io/file "/Users/martin"))))
14:32	ucb	how can I find out what interfaces/protocols a particular symbol implements?
14:33	stuartsierra	ucb: (ancestors (class foo))
14:33	ucb	stuartsierra: thanks!
14:33	stuartsierra	np
14:33	ucb	stuartsierra: just wondering because of this https://gist.github.com/ulises/8805830
14:36	stuartsierra	ucb: `get` on anything not a map returns `nil`. http://stuartsierra.com/2013/02/04/affordance-concision
14:36	ucb	stuartsierra: that'd explain it, thanks :)
14:36	stuartsierra	you're welcome
14:36	ucb	mind you, I found out in a rather inconvenient way, but otoh yay tests
14:37	AeroNotix	is there any ring middleware which creates a handler for JVM statistics?
14:37	AeroNotix	or some other fancy pants monitoring stuff would be cool, too
14:37	AeroNotix	I want a /health endpoint which my ops dudes can query
14:38	ucb	AeroNotix: shouldn't be that hard to write with codahale's metrics lib
14:39	AeroNotix	ucb: lol, you would be surprised. We looked into codahale's stuff and it seems SOME libs use yammer, some use codahale's, it's a mess
14:39	AeroNotix	the API is practically the same, yet different enough for it to not work together
14:39	AeroNotix	we tried hooking it up to clojure-metrics and a Cassandra client graphite reporter
14:39	AeroNotix	no dice
14:40	AeroNotix	but sure, we're using that already
14:40	AeroNotix	I wanted something which was just (get-jvm-statistics) or something
14:40	ucb	AeroNotix: interesting; I've used clojure-metrics and codahale's lib directly with not much grief. Even started wrapping up the latest version myself.
14:41	AeroNotix	ucb: It was just the integration with cassandra we found annoying, one lib has it, one doesn't, some ring endpoint reporter middleware used the other... seemed a mess
14:42	ucb	AeroNotix: oh, I see. My use case was far simpler :)
14:42	AeroNotix	ucb: no worries
14:45	hombotto	How to I get midje :autotest to rerun my tests, when i change a src/x.clj file in a lein project?
14:46	AeroNotix	hombotto: do you really want that?
14:46	hombotto	I assumed, that I was testing the changes in my src/x.clj files?
14:47	hombotto	Specifically I'm working on this project: https://github.com/iloveponies/p-p-p-pokerface
14:48	hombotto	AeroNotix, I assumed the tests should be rerun against changed functions in src/x.clj
14:49	hombotto	Am I missing something?
14:49	AeroNotix	Dunno, that just seems like a waste of time to me. If you change some minor whitespace do you really want the test suite running?
14:51	hombotto	AeroNotix, well I'm changing the functions implementations in src/x.clj, not just whitespace
14:51	AeroNotix	Forget it, you miss my point.
14:52	hombotto	AeroNotix, Sorry. Am i using midje incorrectly? Do you have experience with using it correctly?
14:53	AeroNotix	hombotto: I'm just saying that perhaps autorunning test suites may sound cool and fancy pants but is probably more effort than it's worth
14:53	AeroNotix	what's wrong with Alt-tab, Up, return?
14:53	AeroNotix	that's my :autotest
14:54	hombotto	AeroNotix, I'm used to autorunning tests in all language, I use :-) As a workaround, I'm autorunning the midje tests using inotify events
14:55	AeroNotix	Then go for it
14:55	LLKCKfan	Is there any natural ways to relieve pain without using herbs or weed? No drugs. Prayer does not help
14:55	hombotto	AeroNotix, But I can't get it to work :-)
14:55	AeroNotix	LLKCKfan: masturbation
14:56	TimMc	technomancy: ^ I'm not sure what LLKCKfan's game is, but they should probably be banned? Dunno.
14:56	LLKCKfan	AeroNotix No
14:56	TimMc	or at least kicked repeatedly
14:56	zerowidth	LLKCKfan: try lisp, let the parentheses embrace you like a hug
14:56	AeroNotix	Lisp really is painless, Lein is mostly painless, Clojure eases you in. Just go for it
14:57	AeroNotix	the pieces just click together
14:57	zerowidth	if it hurts it's probably because you're learning something
14:58	mikeyg6754	Does anyone know of a way to throw an error from a Liberator post! handler? I'm uploading an image and I don't want to return a 201 if it fails.
15:00	technomancy	LLKCKfan: inclined to agree with TimMc; consider this a warning
15:06	JoelMcCracken	what is the name of that chef-like tool in clojure?
15:10	grzm	pallet?
15:10	clojurebot	pallet is https://github.com/pallet and http://palletops.com
15:12	JoelMcCracken	yeah, thx
15:28	dbasch	I'm using emacs in paredit mode in an OSX terminal, and I can't get C - right arrow to slurp forward. Anyone experienced this?
15:30	tbaldridge	dbasch: yeah, it's a well known problem
15:30	tbaldridge	personally I use ALT+) instead, but some people have gotten it to work
15:30	tbaldridge	rkneufeld: didn't you have a how-to on this?
15:32	pbostrom	dbasch: I think you can get it to work with iTerm2 and then set terminal type to xterm
15:33	dbasch	tbaldridge: thanks, I'll try another key binding
15:34	pbostrom	dbasch: actually, I got it to work adding something to my init.el file, let me put it in a gist
15:36	dsrx	C-<Right> works for me, but I use M-) myself
15:37	scape_	,(Integer/toBinaryString(unchecked-byte 129)) ;;expecting 1000 0001
15:37	clojurebot	"11111111111111111111111110000001"
15:40	xuser	emacs key bindings hell
15:40	stuartsierra	dbasch: http://offbytwo.com/2012/01/15/emacs-plus-paredit-under-terminal.html
15:40	xuser	sometimes I think emacs is just meant to be run in GUI mode
15:40	veron	viiiiim
15:41	wink	Notepad.exe ducks and runs
15:42	dbasch	stuartsierra: thanks, I'll try this
16:07	rkneufeld	tbaldridge: Do you mean https://gist.github.com/rkneufeld/5126926#a-special-note-for-mac-users ?
16:07	tbaldridge	yeah, thanks
16:08	tbaldridge	and dbasch is gone.....oh well.
16:13	cmiles74	cemerick: Hello! I'm using Friend for authentication and was wondering if there's an easy way to add an additional header to the interactive form workflow. I'd like to jam a token in there for the client to use later.
16:13	cemerick	cmiles74: you mean, when it's redirecting after a successful auth?
16:14	cmiles74	cemerick: Yep, right then.
16:14	cmiles74	cemerick: I'm thinking maybe I want to write my own interactive form workflow, but thought I'd ask first.
16:14	cemerick	cmiles74: comp a function onto the result of (interactive-form ...) that adds the header in question when its return value is a ring response, and does nothing otherwise
16:15	cmiles74	cemerick: It sounds very reasonable when you say it like that. ;-) Thank you.
16:16	cemerick	(fn [resp] (if (ring-response? resp) (add-header resp ...) resp)) ; modify to use the actual helpers in ring, etc
16:16	cemerick	cmiles74: np :-)
16:17	LLKCKfan	I have some chicken strip that have already been cook and then freeze(where freeze when we go them) how can I heat them up on the stove?
16:19	arohner	hrm, do I /ignore LLKCKfan, or do I wait to see the inevitable kick?
16:19	arohner	:-)
16:20	seangrove	thanks you arohner
16:20	arohner	seangrove: thank technomancy
16:20	seangrove	Well, thank you for bringing it to his attention, anyway ;)
16:25	mdrogalis	cemerick: ping
16:28	saj	Hi all
16:28	saj	If anyone here has used korma
16:29	brehaut	~help
16:29	clojurebot	Nobody can help with "X doesn't work". Please provide context: what you did, what you hoped would happen, and what happened instead. A stack trace is especially helpful, if applicable.
16:29	brehaut	hmm thats not the right one
16:29	brehaut	~question
16:29	clojurebot	Huh?
16:29	brehaut	bah
16:29	saj	is there a way to perform joins on an entity that references itself, like a comment that references a parent or child comment?
16:29	technomancy	brehaut: you're looking for ~anyone
16:30	technomancy	though it's too late =)
16:30	brehaut	technomancy: thanks
16:30	brehaut	yes
16:31	cemerick	mdrogalis: shoot
16:40	mynomoto	saj: I haven't used Korma in a while but I remember seen a thread in the Korma group about how to do that.
16:46	saj	mynomoto: do you remember anything about what it was called? I posted in there a long time ago but got no response
16:57	sdegutis	How's the LT paredit plugin these days?
16:57	sdegutis	I just scrolled in emacs and lost my selection because of it.
16:57	ucb	can anybody explain this to me? https://gist.github.com/ulises/8924971
16:57	arrdem	existant and under active development, but I don't know how good it is.
16:57	ucb	I understand what's going on, but I'd like to understand why this is so
16:58	llasram	ucb: Why keywords are functions?
16:58	ucb	llasram: no, why (let [..] ...) becomes a Runnable
16:58	arrdem	(dec so) ;; still waging war on those two guys.
16:58	lazybot	⇒ -5
16:58	ucb	I assume it's to do with let*
16:58	llasram	ucb: It has nothing to do with the let
16:58	arrdem	ucb: because the value it produces is runnable.
16:58	arrdem	ucb: remember let "returns" its tail expression
16:58	llasram	It's because `a` is bound to `:a` and ##(ifn? :a)
16:59	teslanick	ucb: (let [a :a] a) returns a
16:59	lazybot	⇒ true
16:59	ucb	llasram: oh, right, of course !
16:59	teslanick	er :A
16:59	ucb	I was puzzled by a piece of code along the lines of (defn foo [^Runnable f] ... do stuff with f) and (foo (with-open [bindings here] ...))
16:59	ucb	and with-open results in a let construct, so I was puzzled
17:04	effy	i'm reading the book Programming clojure, and in chapter 7 macros, the author describe the defstruct macro, and say: "This macro looks so simple that you may be tempted to try to write it as a function. You won’t be able to, because def is a special form. You must generate def at macro time; you cannot make “dynamic” calls to def at runtime."
17:04	effy	i fail to understand how calling def inside a function doesn't work, could someone show me the light ?
17:04	gfredericks	does reducers work well on sets?
17:05	brehaut	effy: it takes a symbol literal as its first arg right? if you pass a list literal in, its not going to be happy
17:05	arrdem	effy: so def is a special form, it isn't truly a function. The compiler has to deal with a def, which means invoking (eval `(def ...))
17:06	arrdem	effy: normal Clojure code never hits eval. it doesn't need to.
17:06	gfredericks	my guess based on crude benchmarks is no
17:06	gfredericks	is reducers mostly only useful for vectors then?
17:06	hyPiRion	effy: Calling def works inside a function, but what would (defn my-def [sym val] (def sym val)) do? It'll just redefine "sym" all the time
17:07	arrdem	,(defn f [x] (def myfoo x))
17:07	arrdem	,(f 3)
17:07	clojurebot	#'sandbox/f
17:07	clojurebot	#'sandbox/myfoo
17:07	arrdem	0.o
17:07	arrdem	,myfoo
17:07	clojurebot	3
17:07	effy	hyPiRion: oh ok i see with the exemple but then we could still emulate this by passing a (my-defn '('foo 'bar)) no ?
17:08	hyPiRion	effy: nope, because you define sym, not 'foo.
17:09	hyPiRion	,(defn my-def [sym val] (def sym val))
17:09	clojurebot	#'sandbox/my-def
17:09	hyPiRion	,(my-def 'foo 'bar)
17:09	clojurebot	#'sandbox/sym
17:09	hyPiRion	,(my-def 'baz 'bar)
17:09	clojurebot	#'sandbox/sym
17:10	effy	(defn my-def [foo] (def (first foo) (second foo)) (my-def '('a 'b))
17:10	effy	does it make sense ?
17:10	Bronsa	no
17:10	Bronsa	def is a special form, it has special evaluation semantics
17:10	effy	meaning i cannot pass a (first foo) as an argument ?
17:10	Bronsa	right
17:11	hyPiRion	exactly, it has to be a symbol
17:11	effy	ok, that's way clearer now :) thanks you
17:12	hyPiRion	np, enjoy the rest of the book :)
17:12	Wild_Cat	wait, so calling def anywhere but at the toplevel of a namespace doesn't alter the namespace?
17:13	llasram	Wild_Cat: It does. `def` interns a name in a namespace, creating a var if necessary
17:13	hyPiRion	Wild_Cat: sure does
17:13	Wild_Cat	OK, nevermind, I misread your explanation.
17:30	arrdem	is there an existing tookit for dealing with regexes as abstract patterns?
17:32	broquaint	arrdem: Like this? https://github.com/jclaggett/seqex
17:32	arrdem	broquaint: ooh shiny
17:33	broquaint	It is rather, haven't found an excuse to use it yet :)
17:33	seangrove	Pretty interesting about Pedestal. Makes me happy they're not averse to adopting React if it works out for the best
17:33	arrdem	:P it looks awesome, but I was really looking for a toolkit that did regex string to abstract state machine translation :P
17:33	arrdem	I'll probably just have to roll my own because it's so specific...
17:34	arrdem	(dec so) ;; this guy...
17:34	lazybot	⇒ -6
17:34	AimHere	Well clojure being clojure, all it does is use the underlying Java regexes, I'm led to believe
17:34	arrdem	AimHere: correct it does. I'm pondering a project where I want to be able to disassemble regular patterns.
17:47	srruby	How do I get (1 2 1 2 1 3) from (1 2) ?
17:48	srruby	I meant (1 2 1 2 1 2) from (1 2)
17:49	dnolen_	,(take 3 (cycle '(1 2)))
17:49	clojurebot	(1 2 1)
17:49	dnolen_	,(take 6 (cycle '(1 2)))
17:49	`cbp	,(take 6 (cycle '(1 2)))
17:49	clojurebot	(1 2 1 2 1 ...)
17:49	clojurebot	(1 2 1 2 1 ...)
17:49	srruby	thanks!!
17:49	srruby	I forgot cycle
17:49	arrdem	dnolen_: can core.logic use dynamically emitted constraints?
17:51	dnolen_	arrdem: constraints are reify instances, so you can construct interesting constraints on the fly, but it's going to be tricky to something super dynamic.
17:51	dnolen_	"tricky to do"
17:54	arrdem	dnolen_: okay, thanks. sounds like that's probably a no then.
17:56	dnolen_	arrdem: the API around constraints isn't well documented because I'm still unsure exactly what it should look like - there are lot of tricky things around designing constraints that work well together. It's surprisingly easy to have unsound combinations - especially arbitrary predicates (which is supported but again haven't really talked about much because
17:56	dnolen_	of the issues)
17:59	arrdem	dnolen_: sure. Basically what I'm looking at is an AI search problem in a game with dynamic rules. So some things which would normally be valid options may become illegal due to special rules in play. I was thinking that rather than build an abort based system with dynamic predicates that I could just rewrite the search constraint(s) based on rules in play and hopefully perform better as a result.
18:00	arrdem	the alternative being "try anything and if a dynamic rule predicate fails, it isn't legal abort".
18:00	arrdem	s/anything/everything/
18:03	dnolen_	arrdem: sounds interesting, but yeah hard for me to say if core.logic is a good fit - clara-rules might interesting here?
18:04	dnolen_	s/might/might be
18:05	arrdem	dnolen_: clara-rules is definitely along the same lines for rule composition. may or may not be useful.
18:06	arrdem	anyhow. thanks!
18:53	karlgrz	Hey all, I'm a newb clojure'r hacking on overtone and trying to get live coding working in vim. Anyone have experience setting up vim-clojure-static and vim-fireplace?
18:54	karlgrz	I keep seeing this error when I try to run :Eval
18:54	karlgrz	IllegalAccessError index-buffer does not exist clojure.core/refer (core.clj:384
18:54	karlgrz	9)
18:54	arrdem	karlgrz: sounds like your fireplace is working just fine and your code is wrong...
18:55	karlgrz	arrdem: that surprises me very little ;-)
18:55	arrdem	karlgrz: paste? http://refheap.com
18:56	karlgrz	arrdem: https://www.refheap.com/36143
18:56	karlgrz	super simple
18:57	karlgrz	I'm on OS X mavericks, running iTerm with two tabs, one running "leon repo" from my project's root
18:57	arrdem	urgh. this looks like an error in overtone....
18:57	karlgrz	the other with a source window
18:57	karlgrz	ah
18:57	arrdem	karlgrz: is there a longer stacktrace you can show me?
18:57	akurilin2	howdy
18:57	karlgrz	let's see...that's all I got from the vim windows
18:57	karlgrz	*window
18:57	akurilin2	I'd like to move out the peristence layer of a project of mine into its own library so that I can share it across multiple web apps.
18:58	akurilin2	What's the best way for me to extract that code and move it into a local library?
18:58	akurilin2	I'm sure there's something in lein for this :P
18:59	akurilin2	Ideally I'd be able to run everything off of source so that I can edit both the projects and the dependencies without having to rebuild each time
18:59	karlgrz	arrdem: to note, when I just go to the repo window and type commands in they work as expected
18:59	karlgrz	arrdem: for example:
18:59	karlgrz	user=> (use 'overtone.live)
18:59	karlgrz	nil
18:59	karlgrz	user=> (demo (sin-osc))
18:59	karlgrz	#<synth-node[loading]: user/audition-synth 32>
18:59	karlgrz	and I hear the desired beep
19:00	karlgrz	I installed SuperCollider properly (I think)
19:03	dbasch	If it's helpful to anyone, the reason Control arrows didn't work for me on Emacs under OSX was that Mission Control intercepts those keys. I disabled them and now it works.
19:05	dsrx	a leiningen question (which might be more of a maven question): if my clojurescript project is intended to be used as a library, is it best practice to depend on as old a version as possible?
19:06	dobry-den	Question about noir.util.crypt: (encrypt (BCrypt/gensalt) "secret-password") generates a salted password. But how does it know what the salt is later when you go to authenticate a password?
19:07	Cr8	its embedded in the output
19:08	Cr8	the thing "encrypt" returns contains both the bcrypt hash and the salt
19:08	dobry-den	The underlying Java is here http://www.mindrot.org/files/jBCrypt/jBCrypt-0.2-doc/BCrypt.html - If (BCrypt/checkpw hashed-and-salted-pw raw-pw) can extract the salt, then what does a salt add at all?
19:08	Cr8	that work done against one hash isn't valid against another
19:08	Cr8	that is, no work you do cracking user A's password hash is useful against user B's
19:09	Cr8	salts aren't secrets
19:09	Cr8	they just exist to perturb the hash and add a random factor
19:09	karlgrz	arrdem: if I do a :lopen I see what appears to be some stack traces, but nothing much related to the error, I think
19:10	arrdem	karlgrz: you'd be surprised... paste please?
19:10	dobry-den	Cr8: so checkpw can't split the unsalted hash from the salt. it can only see what the salt is?
19:11	Cr8	checkpw grabs the salt, runs encrypt against the user input with the same salt, and compares against the hash
19:11	Cr8	that's it
19:12	Cr8	you could easily split out the unsalted hash if you wanted to
19:12	Cr8	but that wouldn't be useful
19:12	karlgrz	arrdem: https://www.refheap.com/36147
19:12	Cr8	as it's not "unsalted"
19:13	Cr8	you still need to know the salt value to check against it
19:13	arrdem	karlgrz: okay yeah that stacktrace is pretty useless :P
19:14	karlgrz	arrdem: ;-) ohhhh yea
19:14	karlgrz	arrdem: the only plugins I have installed are vim-fireplace, vim-clojure-static, and vim-classpath
19:15	dobry-den	Cr8: so if the pwd is "secret", the salt is "abc", and the hash is "q1w2e3", and together it's "abc-q1w2e3" in the database, then q1w2e3 is in your rainbow table
19:15	karlgrz	arrdem: SORRY...also nerdtree
19:15	Cr8	at any rate, correct usage is to store the exact output of the encrypt fn, and check it with the checkpw fn
19:15	dobry-den	oh duh, they are hashed together
19:15	Cr8	yeah
19:15	dobry-den	so it'd be "abc" + hash("abc" + "q1w2e3")
19:15	dobry-den	damn
19:15	Cr8	sort of
19:15	Cr8	bcrypt is a bit more complex than that
19:15	dobry-den	right
19:16	dobry-den	thanks.
19:16	Cr8	does lots of iterations and stuff to be more expensive
19:16	Cr8	but the general concept of the salt being used as early input to the function is accurate
19:16	Cr8	which is why the salt doesn't need to be a secret, it's already done its job :)
19:16	karlgrz	arrdem: I just tried opening from the project root instead of the src folder...no change.
19:16	dobry-den	(encrypt "secret") and then (checkpw "secret" (:digest user)) just felt too simple
19:17	dobry-den	i see
19:17	Cr8	\(and also why you use a new salt for each thing you hash. which some folks get wrong)
19:17	arrdem	karlgrz: sorry I can't help much here. I transitioned from Vim to Emacs for Clojure work before fireplace arrived :P
19:17	Cr8	(encrypt "secret") is probably the correct usage, that way you can't make that mistake
19:17	Cr8	it'll always generate you a brand new salt
19:17	Cr8	which is what you want
19:17	arrdem	karlgrz: I was looking for an obvious Clojure side issue and failed.
19:17	karlgrz	arrdem: all good, thanks for trying!
19:18	karlgrz	arrdem: I'll keep hacking at it...might need to reach out to the debs of those plugins
19:21	karlgrz	arrdem: success!
19:21	karlgrz	arrdem: in vim, I had to type cqq to get my code into the repl
19:21	karlgrz	arrdem: then it threw that exception to me, but it played the tone
19:22	arrdem	:o
19:22	sdegutis_	So Hy is the new Clojure it turns out?
19:22	dobry-den	I have a simple app where kids drop [url=example.com/view/q1w2e3][img=example.com/image/q1w2e3.gif][/url] in their forum signature and the app tracks unique requests to /view/q1w2e3 and /image/q1w2e3.gif. Is logging IP address and ensuring referrer enough to get a rough idea of unique visits?
19:22	karlgrz	arrdem: weird shit...I'm gonna have to figure that out
19:22	karlgrz	arrdem: but at least it's working kind of
19:22	dobry-den	sdegutis_: more impressive: https://github.com/halgari/clojure-py
19:22	arrdem	yeah.. the "working kinda" is what ultimately drove me to Emacs...
19:22	sdegutis_	dobry-den: I heard it died.
19:22	holo	i'm feeling hy, so i'm going to port some clojure fn
19:22	arrdem	the "kinda" days get old. fast.
19:23	dobry-den	sdegutis_: for fun i tried to make some clojure functions for Hy https://github.com/danneu/hyclops/blob/master/hyclops.hy
19:23	sdegutis_	arrdem: I scrolled in emacs today and it ruined my selection.
19:23	sdegutis_	Emacs can't handle the concept of your cursor not being visible on the screen. It just can't wrap its head around that idea.
19:24	arrdem	nope. that was an open glare.
19:24	dobry-den	holo: i meant to ping you https://github.com/danneu/hyclops/blob/master/hyclops.hy
19:25	holo	dobry-den, no worries, i was reading it anyway.. looking for interleave and couldn't find it. did you think about porting it? :>
19:25	arrdem	sdegutis_: see... there is some shit I'm willing to put up with, and as I'm slowly evolving away from my mouse for editing, I'm OK with Emacs having shall we say unfortunate mouse behavior.
19:25	karlgrz	arrdem: :-) if it means I don't need to re-learn my text editor just to get some metronome beats for my guitar playing, that's a load of win in my book ;-)
19:25	arrdem	sdegutis_: because really it was never designed for mouse navigation.
19:25	karlgrz	arrdem: thanks again for the nudges in the right direction
19:25	arrdem	$google all I do is win youtube
19:25	lazybot	[DJ Khaled "All I Do Is Win" feat. Ludacris, Rick Ross, T ... - YouTube] http://www.youtube.com/watch?v=GGXzlRoNtHU
19:26	dobry-den	holo: i thought it was a kinda funny project i wrote in an afternoon but the reception it got in #hy actually put me off of it
19:27	holo	dobry-den why?
19:27	karlgrz	arrdem: \m/
19:28	dobry-den	holo: i think everyone that was on at the moment was just having a bad day.
19:28	holo	dobry-den, why not try again? maybe today is better
19:29	dobry-den	i create a lot of dumb projects. it's not something that had any real follow through.
19:29	dobry-den	but since you're interested in Hy, there might be some low hanging fruit you can grab from it
19:30	dobry-den	i learned pretty quickly how easy clojure's seq abstraction makes this language
19:30	dobry-den	and my efforts to add clojure fns to hy and elisp immediately miss that magical fruit
19:33	holo	i see
19:33	holo	yes, i'll keep your file, thanks
19:44	dobry-den	Would it be a bad fit for Datomic to track visits to a resource? {:request/ip "1.2.3.4", :request/referer "http://example.com/hello", :request/target <ResourceRef>}?
19:45	sdegutis_	arrdem: and 30 years later they haven't caught onto the idea
19:45	arrdem	sdegutis_: eh... still OK with it.
19:46	dobry-den	I've been using Datomic as my general purpose database for so long that even trying to use Redis on a recent project slowed me way down.
20:20	greg`	newbie question, best resource for learning clojure?
20:22	technomancy	greg`: clojurebook.com is a great place to start
20:29	dobry-den	greg`: http://www.braveclojure.com/ is a real lightweight resource
20:32	dobry-den	what's a rule of thumb when deciding to create a :created "column" in datomic vs. just leaning on the tx time?
20:32	dobry-den	the obvious one of course would be if you actually intend to every change it
20:32	greg`	thanks
20:42	dobry-den	I know this isn't a good hour for questions, but I'll try one more. I have a bunch of "Head" images, "Body" images, "Legs" and "Arms". I'd like to write an abstraction that lets me set, say, and "anchor point" on the body for the head, armL, armR, legL, legR (and an anchor point on each body part) that will compose them into an image.
20:42	dobry-den	What image library would you use for this kind of thing?
20:42	dobry-den	I'd even use a Python or Ruby lib/wrapper if it's easier
20:58	`szx	dobry-den: I've used libgd in nodejs/ruby and VIPS in ruby to compose images, not sure if either has a java binding
21:13	dobry-den	`szx: thanks. i'm having a go using Java's ImageIO and i'll check those out if i can't figure out my issue
21:23	bob2	so, I have a global (atom map) that I'm using to fake out a db for the moment. in my test-fixture I'd like to empty the map out after each test, but I can't seem to figure out how to neatly delete all the keys from map. is there a simple way? is this a terrible idea?
21:24	bob2	(I can do it with apply and dissoc)
21:25	muhoo	dobry-den: it's write constrained due to the transactr
21:26	muhoo	i wouldn't use it for logging responses or anything write-bound
21:26	bob2	hm, I guess I just return {} from the function swap! invoked
21:32	noto2	bob2, reset!
21:38	dobry-den	muhoo: how would you personally log a datastructure that looks like {:visitor/ip "1.2.3.4", :visitor/referer "http://example.com", :visitor/target <Resource ID>}?
21:43	rhg135	TEttinger, side effecting fns make y'all sound excited
21:44	TEttinger	ha
21:44	rhg135	it adds spice
21:44	arrdem	rhg135: technically they are STM fns. normal fns with side-effects don't get the !, only STM ones.
21:45	arrdem	I'ts A Feature™
21:45	rhg135	rly?
21:45	arrdem	yarly
21:46	bob2	TEttinger, hm - (reset! x (atom {})) ?
21:46	TEttinger	no, no atom
21:47	TEttinger	(reset! x {})
21:49	bob2	hmm
21:53	muhoo	dobry-den: the cheap and dirty thing would be just to stream it to disk as edn
21:54	muhoo	dobry-den: or use some other database that's optimized for write speed. but i've been able to make datomic choke on heavy writes
21:56	john2x	hello. why is my font having "extra" space at the bottom here? http://i.imgur.com/H85K1jR.png in other programs it doesn't.
21:57	john2x	(see how the blocks on the parens go beyond the parens shape)
21:59	brehaut	john2x: sure you've got the right channel?
22:00	john2x	oh wow. sorry
22:01	brehaut	john2x: fwiw it looks like a foible of the particular monospace font you are using
22:01	brehaut	john2x: that line of text has no descenders (tails of eg lower case g, p, q, y)
22:02	brehaut	john2x: but if it did, they would use most of that space below the baseline
22:03	systemfault	Let's say I would like trying to write a RESTful/ish web service using clojure, what framework/lib would you suggest me to use?
22:05	xuser	systemfault: I though Liberator was the clojure thing for that :)
22:05	systemfault	xuser: I saw a video on it recently, looks too amazing to be real.
22:05	systemfault	So.. I wanted to know what people are actually using :)
22:06	bob2	systemfault, liberator and compojure is going ok for me so far
22:07	bob2	it's a bit uh detailed if you let it be
22:07	systemfault	Ok good :)
22:08	systemfault	And from what I understand... there are a few "ring-compatible" libraries, which one should I use? Read that Httpkit is ring-compatible
22:08	systemfault	(I'm a total noob.. just reading... and reading..)
22:09	bob2	for serving? I'm just using jetty for now
22:09	systemfault	Ok
22:09	akhudek	systemfault: depends on your requirements. Jetty is fine for a lot of things. Don't have much experience with http-kit, but it seems to fair well in benchmarks and is based on an async model, and immutant is there if you need HA and other serious features.
22:10	systemfault	akhudek: Ah, good to know, thank you
22:12	akurilin2	Lein question: does anybody have a sample project.clj where you use a local repository to link against a custom clj library that's not in clojars?
22:12	systemfault	Last question... I'm a fan of static typing, if I'm a beginner, should I bother using core.typed?
22:12	arrdem	akurilin2: it's invisible
22:12	akurilin2	or at the very least the snippet from that project.clj
22:12	arrdem	akurilin2: just do a local build, and `lein install`
22:12	bob2	systemfault, I'd say no
22:13	arrdem	akurilin2: the lein project looks identical.
22:13	ambrosebs	systemfault: probably not if you're just starting out
22:13	akurilin2	arrdem: ok let me investigate lein install
22:13	systemfault	Ok guys, your opinions are invaluable to me. Thank you so much
22:13	arrdem	akurilin2: for arbitrary jars you have to muck with maven :c
22:13	akurilin2	arrdem: basically what you're saying is that lein install will add the library to the current user's m2 repo and will check against that, right?
22:13	arrdem	yarp
22:13	ambrosebs	systemfault: if you're looking for better errors, dynalint might help
22:14	akurilin2	that's pretty damn
22:14	akurilin2	convenient
22:14	bob2	dynalint seems to not want to install for me :/
22:14	systemfault	Ah :)
22:15	ambrosebs	bob2: oh, what went wrong?
22:15	bob2	java.io.FileNotFoundException: Could not locate dynalint/lint__init.class or dynalint/lint.clj on classpath:
22:16	bob2	$ grep dynalin project.clj
22:16	bob2	[lein-dynalint "0.1.3"]]
22:16	ambrosebs	bob2: ah there's a new requirement. You need to add a version of com.ambrosebs/dynalint to your profiles
22:17	ambrosebs	$ grep com.ambrosebs/dynalint project.clj
22:17	ambrosebs	:P
22:17	ambrosebs	check the lein-dynalint readme
22:17	bob2	ah, I did not read carefully enough, sorry!
22:18	ambrosebs	bob2: I'll update lein-dynalint to give a better error
22:18	ambrosebs	it's confusing, but means I don't need to release lein-dynalint just because there's a new dynalint
22:18	bob2	ohhh right
22:20	bob2	ambrosebs, works now - thanks!
22:20	ambrosebs	bob2: great
22:21	akurilin2	OK I'm going to ask a silly one: if I want to group my libraries under a prefix like "myorg.", how do I still get lein new to spit out core.clj by default? Right now if I do lein new myorg.foo, it will generate src/myorg/foo.clj
22:21	akurilin2	Is this just one of those cases where I have to do it by hand?
22:22	akurilin2	As in, I'd like that ns to be called "myorg.foo.core"
22:28	akurilin2	Oooh leiningen won't install new versions of my lib until I change version numbers.
22:28	akurilin2	Sad.
22:28	akurilin2	And makes total sense.
22:31	akurilin2	Hm does anybody know how to make this a bit snappier?
22:31	brehaut	Mechanical Bull. Game Changer. Innovate
22:31	arrdem	brehaut: lolz
22:31	arrdem	akurilin2: not really. lein already does some stuff to make JVM startup fast-ish
22:32	arrdem	akurilin2: lein ancient is your friend
22:34	akurilin2	arrdem: So if I'm understanding correctly, if I change a library and run lein install, I HAVE to restart the process to pick up the newer jar?
22:35	arrdem	yes
22:35	akurilin2	ok
22:35	akurilin2	arrdem: should my boot times be sped up if a giant chunk of my existing codebase suddenly is crammed into a jar?
22:36	arrdem	not really...
22:36	arrdem	afaik
22:36	akurilin2	ok
22:37	lgas	Hi. Is there an idiom for handling the namespace refers if you want to library that takes advantage of another library (e.g. prismatic's schema) if it's installed, but installs just fine without it?
22:37	lgas	*namespace references
22:38	akurilin2	Seems like a pretty frequent pattern you encounter when working with storage is implementing the persistence layer once and then wanting to use it across a variety of applications
22:38	akurilin2	So I'm moving that chunk into its own lib
22:46	dsrx	is it a known bug that (.start (Thread. #(println "hello world"))) does not print to the repl buffer in cider?
22:46	dsrx	seems like futures work ok
22:53	brehaut	dsrx: you probably have to flush the buffer
22:53	brehaut	(doc flush)
22:53	clojurebot	"([]); Flushes the output stream that is the current value of out"
22:54	brehaut	or, alternatively, make sure that out is the same out as in the main thread
22:54	dsrx	brehaut: yeah, all of my printed messages are going to nrepl-server
22:55	dsrx	so i guess it's just a matter of rebinding out then
22:56	dsrx	ahh nice, thanks brehaut
22:56	dsrx	(inc brehaut)
22:56	lazybot	⇒ 22
22:56	brehaut	np
22:56	brehaut	consider it some lucky guessing
22:57	dsrx	:)
23:03	akurilin2	arrdem: btw thanks for explaining the lein install stuff, super helpful.
23:03	arrdem	akurilin2: np :D
23:29	munderwo	Hi all, so I played around with the clojure Koans and they had a nifty thing with their unit tests where they auto-re-ran… does anybody know how to get that? I just want to avoid the repl reload
23:34	dobry-den	munderwo: there's a lein plugin for every testing framework that you can add to ~/.lein/profiles.clj that lets you run some variant of 'lein autotest'
23:35	dobry-den	munderwo: for example, i use http://jayfields.com/expectations/ and it has this plugin https://github.com/jakemcc/lein-autoexpect
23:35	dobry-den	lein autoexpect
23:35	dobry-den	there's a similar plugin for clojure's standard test lib and https://github.com/marick/Midje (another test lib)
23:36	munderwo	dobry-den: so I'm using just the standard clojure.test … any idea what the plugin for lein is for that?
23:37	dobry-den	munderwo: looks like this https://github.com/jakemcc/lein-test-refresh
23:37	dobry-den	(found it here https://github.com/technomancy/leiningen/wiki/Plugins)
23:37	munderwo	dobry-den: so I think it might also be lein-prism?
23:38	dobry-den	id just roll with the one with the most stars and/or most recent activity
23:39	dobry-den	looks like https://github.com/jakemcc/lein-test-refresh
23:40	akurilin2	Korma time: I heard there's a version of the API where I get to pass in the DB map rather than using the dynamically scoped var
23:40	akurilin2	is that a false rumor?
23:44	akurilin2	Or is it all defdb or nothing, huh?
23:45	technomancy	akurilin2: re: restarting when your dependencies change, take a look at lein's checkouts feature
23:45	technomancy	it's meant to help avoid restarts
23:47	dbell	(doc require)
23:47	clojurebot	"([& args]); Loads libs, skipping any that are already loaded. Each argument is either a libspec that identifies a lib, a prefix list that identifies multiple libs whose names share a common prefix, or a flag that modifies how all the identified libs are loaded. Use :require in the ns macro in preference to calling this directly. Libs A 'lib' is a named set of resources in classpath whose contents define a library of
23:48	akurilin2	technomancy: aaaaah that sounds awesome <3
23:48	akurilin2	I thought you guys were seriuosly going to make me lein install for my life ;)
23:50	technomancy	akurilin2: nope; I lived through like four months of doing that with maven, and that's pat of what made me start working on lein
23:51	akurilin2	Thank you so much :P
23:52	dsrx	lein checkouts are great
23:52	seubert	are they dsrx ???
23:52	lazybot	seubert: Yes, 100% for sure.
23:53	seubert	lazybot too good at his/her job
23:53	technomancy	lazybot: botsnack
23:53	lazybot	technomancy: Thanks! Om nom nom!!