#clojure – 2016-02-25

2016-02-25

00:03	tolstoy	I imagine the botnet folks have something special.
01:44	adamz	Hi. Is there anyway to improve the performance of instanciating Java classes from clojure? Im calling `(java.util.Data. etime)` > 30k/s. running VisualVM against the clojure VM process shows that `java.lang.Class.forName()` is the most expensive method. To me, that sounds like Clojure is looking up the class on every invocation.
01:45	adamz	sorry, `(java.util.Date etime)`
01:46	jeaye	instantiate
01:52	hiredman	adamz: there are things you could look at to speed that up, but it would be way better to just not allocate all those new data objects, are you sure you can't get by with a long?
02:14	adamz	hiredman: unfortunatly, im passing them onto a libaray (the DataStax Cassandra driver) which /expects/ Date objects — i’ve looked at ways to avoid passing Date and insted just long, but it’s not possible without modifying the driver code :(
02:37	Vishal_	any1 active?
02:40	Vishal_	any1 familiar with atoms and their usage?
02:41	mange	Vishal_: anything in particular?
02:50	paul_muadib666	can someone recommend a clojure web app tutorial that covers datomic and compojure. I found one but its out of date and it uses this compojure-api library that I'm having issues with..
03:10	Vishal_	Link: https://gist.github.com/longdongjohn/5bdbada220b2e4a2bb06 ; I keep getting a Null pointer Exception on the atom variable temp3, which i have initialised... Any suggestions?
03:14	hiredman	Vishal_: this code is poorly formatted, and is full of mistakes
03:14	hiredman	def always creates a global, don't use it inside a function
03:15	hiredman	every time you wrap something in () parens that means a function call
03:15	TEttinger	Vishal_: (prn temp3) returns nil
03:17	TEttinger	and yeah, hiredman's right about def especially, it looks like you're using (def m2_flag 0) where you should use let
03:20	TEttinger	for someone whose nick on github is longdongjohn he doesn't seem to take criticism well...
03:25	somename	nope..... i went for lunch... and the channel pinged out
03:25	somename	thanks for the input anyways
03:27	Vishal_	Thanks for the input btw
03:28	TEttinger	I was wondering what happened! there's a few quick things to try
03:29	TEttinger	def should be changed to let, like this
03:29	Vishal_	i can't use let because I'm trying use that value in another function
03:29	TEttinger	ah, that's bad. that sounds like m1_flag and m2_flag should be atoms if they're changing
03:29	Kneiva	Is there better way to transform [[:a [1 2]] [:b [3 4]] [:c [5 6]]] into {:keys [:a :b :c], :values (1 2 3 4 5 6)} than https://www.refheap.com/115180 ?
03:29	TEttinger	currently they're var, which is kinda mutable
03:30	TEttinger	(gt) on its own seems like it may be a bug
03:31	Vishal_	that's what i tried ...but i get a null pointer exception at the location of temp3 in the code
03:32	Vishal_	https://gist.github.com/longdongjohn/2b274f1b6d786ce07373
03:32	Vishal_	thats the complete fiel
03:33	TEttinger	,(let [coll [[:a [1 2]] [:b [3 4]] [:c [5 6]]] ks (map first coll) vs (mapcat second coll)] {:keys ks :values vs})
03:33	clojurebot	{:keys (:a :b :c), :values (1 2 3 4 5 ...)}
03:33	TEttinger	,(let [coll [[:a [1 2]] [:b [3 4]] [:c [5 6]]] ks (mapv first coll) vs (mapcat second coll)] {:keys ks :values vs}) ; to be exact
03:33	clojurebot	{:keys [:a :b :c], :values (1 2 3 4 5 ...)}
03:35	Kneiva	Ah, right. That's more readable, I think. Thanks.
03:35	TEttinger	I am sure someone more experienced than I could clean it up even more!
03:36	TEttinger	cpu_threshold is undefined, thus nil, Vishal_
03:37	TEttinger	I think the error may be in an odd place and getting bad info
03:37	Vishal_	its part of the includes "globals.clj"
03:37	TEttinger	ah I was wondering what that was
03:39	Vishal_	ill try changing the m1 and m2 flags as atoms and run it... will let you know of the changes
03:40	TEttinger	I wonder if (prn temp3) is the issue, should that be (prn @temp3)
03:41	TEttinger	,(def temp3 (atom 0))
03:41	clojurebot	#'sandbox/temp3
03:41	TEttinger	,(prn temp3)
03:41	clojurebot	#object[clojure.lang.Atom 0xcaf670e {:status :ready, :val 0}]\n
03:42	TEttinger	Vishal_: I think where you have (prn temp3) should actually be (fn [& children] (prn @temp3))
03:43	TEttinger	since in all other cases you have a fn and onlythat case evaluates to nil
03:43	Vishal_	will try it now
03:51	Vishal_	TEttinger: https://gist.github.com/longdongjohn/2b274f1b6d786ce07373 this is the updated code.. Im getting an exception long cannot be cast to IFn on line 42
03:54	Vishal_	somename: asd
03:55	TEttinger	let's see...
03:57	TEttinger	,(doc reset!) ; let me refresh my memory
03:57	clojurebot	"([atom newval]); Sets the value of atom to newval without regard for the current value. Returns newval."
03:58	Vishal_	ok will check it out
04:00	TEttinger	Vishal_: the long to IFn thing usually means that a number is at the start of a parenthesized call, like (1) will throw that
04:00	TEttinger	,(1)
04:00	clojurebot	#error {\n :cause "java.lang.Long cannot be cast to clojure.lang.IFn"\n :via\n [{:type java.lang.ClassCastException\n :message "java.lang.Long cannot be cast to clojure.lang.IFn"\n :at [sandbox$eval47 invokeStatic "NO_SOURCE_FILE" 0]}]\n :trace\n [[sandbox$eval47 invokeStatic "NO_SOURCE_FILE" 0]\n [sandbox$eval47 invoke "NO_SOURCE_FILE" -1]\n [clojure.lang.Compiler eval "Compiler.java" 6927]...
04:01	TEttinger	if it's the other way around, IFn cannot be cast to long, that's usually because you have a fn by name in the middle of a math thing like (+ 1 inc)
04:01	TEttinger	,(+ 1 inc)
04:01	clojurebot	#error {\n :cause "clojure.core$inc cannot be cast to java.lang.Number"\n :via\n [{:type java.lang.ClassCastException\n :message "clojure.core$inc cannot be cast to java.lang.Number"\n :at [clojure.lang.Numbers add "Numbers.java" 128]}]\n :trace\n [[clojure.lang.Numbers add "Numbers.java" 128]\n [clojure.lang.Numbers add "Numbers.java" 3640]\n [sandbox$eval71 invokeStatic "NO_SOURCE_FILE" 0]...
04:01	TEttinger	hm different error, same sort of thing
04:03	TEttinger	ohhhhh
04:03	TEttinger	,(def temp3 (atom 0))
04:03	clojurebot	#'sandbox/temp3
04:03	TEttinger	,(reset! temp3 20)
04:03	clojurebot	20
04:03	TEttinger	those reset calls are returning Long numbers
04:03	TEttinger	you may need to pass a fn to where
04:05	TEttinger	I don't know what "where" does or from which lib it's from
04:06	Vishal_	its part of riemann
04:06	Vishal_	and ill try passing a function to where
04:06	Vishal_	getting into a meeting ...will be afk for 15 mins
04:08	solatis	ok, i'm looking for someone with an opinion on data/type validation in clojure and when to do it
04:08	solatis	i am used to languages with strong types (e.g. Haskell and C++) that validate as early and as often as possble
04:08	solatis	however, I don't think that's an optimal approach for Clojure / LISP
04:09	solatis	i am thinking that the best fit for Clojure would be to validate as late as possible, but let the error "bubble back up", to provide accurate feedback on what/where it went wrong
04:09	solatis	is that a reasonable approach?
04:10	solatis	case in point: using some REST api, do i validate the data before I send it, or do I just let remote server generate an error in case the function was called with incorrect data
04:11	TEttinger	n0lan: are you the typed clojure ninja-in-chief :)
04:11	MJB47	solatis: i like to do both
04:11	MJB47	client side validation is much faster and gives the user a better experience
04:11	MJB47	but cant be trusted
04:12	n0lan	no I was bored looking for "popular" IRC channels on google, lol
04:14	solatis	MJB47: right, but in this case it's for internal code
04:14	solatis	and speed is not really a concern
04:15	solatis	and i don't like to do things twice, that means that when there's a change, i need to change my validation code twice :)
04:15	MJB47	maybe not with reader conditionals
04:17	solatis	let me look that up
04:17	solatis	that looks scary
04:41	adamz	Hi. Is there anyway to improve the performance of instanciating Java classes from clojure? Im calling `(java.util.Date. etime)` > 30k/s. running VisualVM against the clojure VM process shows that `java.lang.Class.forName()` is the most expensive method. To me, that sounds like Clojure is looking up the class on every invocation.
04:46	amalloy	adamz: because the clojure compiler doesn't know what type you expect etime to be, and Date has multiple constructors
04:46	amalloy	try (java.util.Date. ^long etime) or whatever type is appropriate
04:46	adamz	cool. i’ll give that a shot. thanks!
05:00	adamz	yup. that certainly made a difference! thanks amalloy :)
05:03	magopian_	hello there :) Can someone tell me what is ^{} ?
05:03	magopian_	eg: ^{:key timestamp}
05:03	magopian_	it looks like a map, but what is the "^" for?
05:05	qsys	it's meta-data
05:07	qsys	http://xahlee.info/clojure/clojure_metadata.html
06:08	edmondo1984	Hello, anyone could help a Clojure newbie to tackle one of the exercise of Clojure for the brave and the true? Quite a basic one
06:13	paul_muadib666	edmondo1984: i maybe can, whatsup?
06:19	magopian_	thanks qsys :) (and sorry for the lag ;)
06:44	edmondo1984	so I need to write a function to update a nested hash
06:44	edmondo1984	something that you can write like this
06:44	edmondo1984	(update-in :key1 :key2 value {} )
06:44	edmondo1984	and would result in { :key1 { :key2 value } }
06:45	TEttinger	typically you'd have keys in a vector
06:45	paul_muadib666	so reimplementing update-in?
06:45	TEttinger	,(update-in {} [:key1 :key2] :value)
06:45	clojurebot	{:key1 {:key2 nil}}
06:45	TEttinger	,(update-in {} [:key1 :key2] (constantly :value))
06:45	clojurebot	{:key1 {:key2 :value}}
06:46	TEttinger	edmondo1984: is the task to reimplement the built-in update-in fn?
06:49	TEttinger	,(defn build-nested [v & ks] (update-in {} ks (constantly v)))
06:49	clojurebot	#'sandbox/build-nested
06:49	TEttinger	,(build-nested 1 :a :b :c)
06:49	clojurebot	{:a {:b {:c 1}}}
06:53	edmondo1984	how do you make the clojure bot evaluate?
06:53	TEttinger	comma prefix
06:53	edmondo1984	, (update-in {} [:key1 :key2] "hello")
06:53	clojurebot	#error {\n :cause "java.lang.String cannot be cast to clojure.lang.IFn"\n :via\n [{:type java.lang.ClassCastException\n :message "java.lang.String cannot be cast to clojure.lang.IFn"\n :at [clojure.core$apply invokeStatic "core.clj" 641]}]\n :trace\n [[clojure.core$apply invokeStatic "core.clj" 641]\n [clojure.core$update_in invokeStatic "core.clj" 5950]\n [clojure.core$update_in doInvoke "c...
06:54	edmondo1984	, (update-in {} [:key1 :key2] (constantly "hello"))
06:54	clojurebot	{:key1 {:key2 "hello"}}
06:54	TEttinger	,(defn build-nested [& args] (update-in {} (butlast args) (constantly (last args))))
06:54	clojurebot	#'sandbox/build-nested
06:54	TEttinger	,(build-nested :a :b :c )
06:54	clojurebot	{:a {:b :c}}
06:54	edmondo1984	ok got it the third parameter is a function
06:54	TEttinger	yep!
06:54	edmondo1984	now the problem is that I have to reimplement updatein
06:54	luma	,(assoc-in {} [:key1 :key2] "hello")
06:54	clojurebot	{:key1 {:key2 "hello"}}
06:54	TEttinger	why though?
06:55	TEttinger	ah
06:55	TEttinger	forgot about that one, (inc luma)
06:55	luma	there's get and get-in, assoc and assoc-in, update and update-in
06:55	edmondo1984	I am learning Clojure and I come from a non lisp background
06:56	edmondo1984	let me find my wannabe-implementation that doesn't work
06:57	TEttinger	I would suspect that being able to reimplement the -in group of fns might be best done later
06:57	TEttinger	I imagine the core implementations of them are a little complex
06:57	edmondo1984	(defn assoc-in "Associate a key inside a nested map" [m [k & ks] v] ((loop [mymap {} ] (if (empty ks) (assoc m k v))) ((recur (assoc-in m ks v) [m ks v]) mymap)) )
06:58	edmondo1984	http://paste.ofcode.org/WdwLuA3cgBdaegQgzPzAg
06:58	edmondo1984	this stuff obviously rdoesn't work
06:58	TEttinger	the two parens seems like a problem
06:59	TEttinger	"((loop"
06:59	TEttinger	that calls the result of loop as if it is a fn
06:59	edmondo1984	and loop is not a fun ?
06:59	edmondo1984	fn
06:59	TEttinger	same with recur, looks like a bug
06:59	edmondo1984	is a macro?
06:59	TEttinger	loop is a macro but it doesn't return a callable thing
07:00	TEttinger	you need one paren to call
07:00	TEttinger	another paren if you want to call what the first call returned
07:00	TEttinger	,(+ 1 2)
07:00	clojurebot	3
07:00	TEttinger	,(constantly (+ 1 2))
07:00	clojurebot	#object[clojure.core$constantly$fn__4614 0x246ee771 "clojure.core$constantly$fn__4614@246ee771"]
07:00	TEttinger	,((constantly (+ 1 2)))
07:01	clojurebot	3
07:01	TEttinger	the first calls +
07:01	TEttinger	the next calls constantly and gives it as an arg 3
07:01	TEttinger	and you can see that prints a messy fn
07:01	TEttinger	since constantly returns a fn
07:01	TEttinger	the last calls the fn constantly returns
07:02	TEttinger	make sense?
07:02	TEttinger	parens don't do grouping in lisps
07:03	TEttinger	you can't just stick more in like in C or Java
07:03	edmondo1984	not sure I understood
07:04	TEttinger	uh hm.
07:04	TEttinger	so you understand that fns are values and can be passed as arguments and returned from other fns and macros, right?
07:05	TEttinger	,(map inc [1 2 3])
07:05	clojurebot	(2 3 4)
07:05	TEttinger	inc is a fn and we can pass it as an arg to map
07:05	edmondo1984	yes
07:05	TEttinger	there's a bunch of fns in clojure's core lib that return other fns
07:06	TEttinger	constantly, comp, and juxt are good examples
07:06	edmondo1984	what do comp and juxt do?
07:06	TEttinger	comp composes multiple fns
07:06	edmondo1984	ah comp is composed
07:06	edmondo1984	right
07:06	luma	,((comp inc :foo) {:foo 3})
07:06	clojurebot	4
07:06	luma	((juxt :foo :bar) {:foo 1, :bar 2})
07:07	luma	,((juxt :foo :bar) {:foo 1, :bar 2})
07:07	clojurebot	[1 2]
07:07	edmondo1984	so juxt take an array of function and applies it
07:07	TEttinger	,(map (comp inc inc) [1 2 3])
07:07	clojurebot	(3 4 5)
07:07	edmondo1984	just a second
07:07	edmondo1984	,((juxt str #(+ 1 %) [ 2 3] ))
07:07	clojurebot	#error {\n :cause "Wrong number of args (0) passed to: sandbox/eval194/fn--195"\n :via\n [{:type clojure.lang.ArityException\n :message "Wrong number of args (0) passed to: sandbox/eval194/fn--195"\n :at [clojure.lang.AFn throwArity "AFn.java" 429]}]\n :trace\n [[clojure.lang.AFn throwArity "AFn.java" 429]\n [clojure.lang.AFn invoke "AFn.java" 28]\n [clojure.core$juxt$fn__4744 invoke "core.c...
07:08	edmondo1984	,((juxt str #(+ 1 %)) [ 2 3] )
07:08	clojurebot	#error {\n :cause "clojure.lang.PersistentVector cannot be cast to java.lang.Number"\n :via\n [{:type java.lang.ClassCastException\n :message "clojure.lang.PersistentVector cannot be cast to java.lang.Number"\n :at [clojure.lang.Numbers add "Numbers.java" 128]}]\n :trace\n [[clojure.lang.Numbers add "Numbers.java" 128]\n [clojure.lang.Numbers add "Numbers.java" 3640]\n [sandbox$eval221$fn__2...
07:08	luma	juxt takes any number of functions and returns a function that calls all those functions on its parameter and returns a vector of the results
07:08	TEttinger	vector, not array, but essentially yes. and it doesn't exactly apply
07:08	edmondo1984	,((juxt str #(+ 1 %)) 2 )
07:08	clojurebot	["2" 3]
07:08	edmondo1984	gooot it!
07:08	edmondo1984	I don't know how I could have survived without Clojure
07:08	TEttinger	there's also partial, which is handy
07:08	luma	so basically ((juxt :foo :bar) param) => [(:foo param) (:bar param)]
07:09	TEttinger	,(map (partial * 2) [1 2 3])
07:09	clojurebot	(2 4 6)
07:09	edmondo1984	what's the benefit of using partial with respect of an anonymous function?
07:09	TEttinger	,(map (juxt identity (comp (partial * 2) inc)) [1 2 3])
07:09	clojurebot	([1 4] [2 6] [3 8])
07:09	edmondo1984	is partial a macro that gives you a better syntax?
07:09	edmondo1984	(partial * 2) is equal to #( * 2 % ) right?
07:09	TEttinger	they're very similar. partial works for any number of arguments
07:10	edmondo1984	got it
07:10	hyPiRion	partial is just a function afaik
07:10	hyPiRion	$source partial
07:10	TEttinger	,((partial *) 1 2 3)
07:10	clojurebot	6
07:10	TEttinger	lazybot ded
07:10	edmondo1984	got it
07:10	hyPiRion	sad times
07:11	edmondo1984	so now what about loop and recurse?
07:12	TEttinger	,(loop [a 0] (if (>= a 5) a (recur (+ a (rand-int 5)))))
07:12	clojurebot	6
07:13	TEttinger	,(loop [a 0] (if (>= a 5) a (recur (+ a 1))))
07:13	clojurebot	5
07:13	luma	recur is a special form that recurses to loop (or the top of the function if there's no loop)
07:13	luma	and loop is a special form that's just like let, but also serves as a recursion point
07:13	edmondo1984	so loop sets the recursion point
07:14	TEttinger	and what you change between recursion
07:14	TEttinger	like I set "a" to 0
07:14	edmondo1984	the a variable gets bound to the argument of recur
07:14	TEttinger	when I recur, I specified I wanted a different value for a, (+ a 1)
07:15	TEttinger	and you can use the name a throughtout the loop body
07:15	edmondo1984	so loop [ variable initialAssignment] (recur (+ variable 1 )
07:15	TEttinger	like it's used when we stop recurring to get the last value
07:15	edmondo1984	it just recur infinitely increasing variable of 1
07:15	edmondo1984	starting from initialAssignment
07:15	TEttinger	yes
07:15	edmondo1984	good
07:15	edmondo1984	now what about the paranthesis around loop?
07:15	edmondo1984	loop return a function right, I need to evaluate it?
07:15	TEttinger	no
07:16	TEttinger	loop doesn't return a fn, it returns whatever the last statement is in the loop that didn't recur
07:16	edmondo1984	so loop evaluates the loop
07:16	edmondo1984	is not lazy somehow
07:16	TEttinger	so in my second loop example, it returned 5
07:16	edmondo1984	got it
07:16	TEttinger	clojure defaults to not lazy
07:16	edmondo1984	now
07:17	edmondo1984	right
07:17	edmondo1984	so now I think I have all the elements except aybe the recursive call
07:18	TEttinger	yeah, loop/recur is a tricky thing and examples help
07:18	TEttinger	it's after 4 AM here, I should head hits keyboard, instantly snoring
07:19	edmondo1984	ok I still can't write the assoc-in
07:20	edmondo1984	(defn assoc-in "Associate a key inside a nested map" [m [k & ks] v] (loop [mymap m ] (if (empty ks) (assoc mymap k v)))
07:21	edmondo1984	this part looks reasonable right? If in the loop I have no other keys (ks), I just associate the current value to the current key and return the map
07:21	luma	right
07:25	edmondo1984	what's the recursive part now
07:26	edmondo1984	can I assign two values in the loop recursion?
07:26	edmondo1984	can I do something like loop [myvar1 initialVal1 myVar2 initialVal2] ?
07:26	luma	sure
07:26	luma	it's just like let
07:27	edmondo1984	what's the syntax?
07:28	edmondo1984	loop [myvar1 myvar2 initialval1 initialval2] or
07:28	edmondo1984	loop [myvar1 initialVal1 myVar2 initialVal2]
07:30	MJB47	second
07:31	MJB47	same with all similar constructs (let, for, doseq etc)
07:35	edmondo1984	MJB47 maybe you can point me to a page where it explain the syntax for all those construct?
07:36	MJB47	i suppose https://clojuredocs.org/clojure.core/let is a good start
07:38	edmondo1984	ok let uses pairs in vector
07:38	edmondo1984	reading the documentation can be very useful
07:39	edmondo1984	now getting back to my recursion problem in assoc-in
07:53	edmondo1984	hello, is anyone here who can help me re-write assoc in?
08:47	Leonidas	edmondo1984: I would check whether there is just one element and then do the assoc
08:59	renl	hi aside from play-clj are there any recommended clojure libraries for doing visualization?
09:35	Empperi	I've got a strange problem with clojure regular expressions
09:36	Empperi	got this regexp (escaped): #"[^\\s()\"]+\\*"
09:36	Empperi	,(re-matches #"[^\\s()\"]+\\*" "foo")
09:36	clojurebot	"foo"
09:36	Empperi	that shouldn't match as far as I understand
09:36	Empperi	and if I run that against plain java that is the case
09:36	Empperi	foo* should match
09:37	Empperi	I'm guessing I should apply some flag to make it work like I want it to but...
09:37	Empperi	oh
09:37	Empperi	remove one backslash before the asterisk
09:38	Empperi	and it works better :)
09:38	Empperi	nevermind!
09:38	Empperi	(complaining in IRC, best way to solve problems)
09:41	mpenet	you can use (re-pattern "") if you insist on escaping stuff
09:41	mpenet	<Empperi> ,(re-matches (re-pattern "[^\\s()\"]+\\*") "foo")
09:41	Empperi	no I don't insist :)
09:41	mpenet	:>
10:17	sdegutis	Do you keep ~/.emacs.d/elpa/ under version control along with its parent directory? If not, how do you ensure that you can quickly and easily get your packages back when cloning your emacs configs?
10:17	sdegutis	Sorry wrong dir.
10:26	sdegutis	Oops I meant wrong channel.
10:26	sdegutis	But I guess it still kind of applies here.
10:26	sdegutis	amalloy_: what's your solution to this? Do you use cask, or do you keep ./elpa/ under version-control, or something else?
10:34	free_beard	sdegutis: just use the package.el api man, :))
10:34	sdegutis	free_beard: thanks bbl
10:46	TimMc	sdegutis: The only thing I gitignore there is .emacs.d/auto-save-list/*
10:51	aurelian	is there a way to go over a list with reduce for example but to know the index? I need to transform the list into another data structure
10:52	aurelian	like map-indexed but with reduce or something
10:52	aurelian	https://www.refheap.com/115194
10:53	aurelian	so, if the index is even add to the first vec, if odd to the second
10:53	sdegutis	TimMc: Hmm my .gitignore has /auto-save-list /backups /autosaves /eshell /tramp /ac-comphist.dat /url/cookies *.elc /ido.last
10:59	TimMc	sdegutis: I suppose if I used those features I would probably ignore those as well. :-)
11:02	sdegutis	ok
11:12	aurelian	partition-by with index would be nice :\|
11:14	ToxicFrog	aurelian: isn't that just (map-indexed vector ...) + (partition-by (comp partitioner second)) ?
11:15	aurelian	hmm
11:15	aurelian	that would go over the coll twice
11:18	aurelian	maybe I just don't have data in best format
11:19	aurelian	the amount of clojure functions with index makes me wonder if this is a good idea
11:53	TimMc	aurelian: Your accumulator could include an index that you increment.
11:53	TimMc	or just use loop :-)
11:54	aurelian	that could also do it
11:54	aurelian	I ended up with something like this: (partition (/ (count data) 2) (map #(-> (:pace %)) (sort-by #(-> (:race %)) data)))
11:54	aurelian	is a list of lists
11:55	aurelian	potentially I can have the data already sorted by race
11:56	aurelian	I'll always have 2 races that's why the partition in the middle
12:00	aurelian	is basically comparing results of bunch of runners over 2 different races, but the sort was by runner, turns out sorting by race simplifies it a lot
12:17	justin_smith	aurelian: resisting the urge to make some sort of "race condition" pun here
12:17	aurelian	xD
12:17	j-pb_	justin_smith: without the context I am not sure wether they are talking about continuous integration stuff or about on which olympic runner to place their sport bets...
12:18	aurelian	second
12:20	j-pb_	I think in the last two olympics there were only black athletes qualified for the 100m sprint, so sorting by race is pretty useless
12:20	justin_smith	aurelian: you may be aware, #(-> (:pace %)) is just :pace but with an arity restriction
12:21	justin_smith	same with #(-> (:race %)) and :race
12:21	aurelian	I'm still learning
12:21	justin_smith	aurelian: np
12:21	justin_smith	just an fyi
12:21	aurelian	what's a better way to do that?
12:22	justin_smith	(map :pace (sort-by :race data))
12:22	aurelian	cheers
12:22	justin_smith	like I said, #(-> (:pace %)) is just :pace with an arity restriction, and the arity restriction didn't help you there :)
12:22	marcfontaine	trying to find a way to use a macro with a def on a set of values
12:22	marcfontaine	(defmacro t [val] `(def ~(symbol val) ~val))
12:22	marcfontaine	user> (t “a”)
12:22	marcfontaine	#'user/a
12:22	marcfontaine	user> a
12:22	marcfontaine	"a"
12:22	marcfontaine	(doseq [v ["b" "c" "d"]] (t v))
12:22	marcfontaine	user> b
12:22	marcfontaine	CompilerException java.lang.RuntimeException: Unable to resolve symbol: b in this context
12:23	aurelian	justin_smith got it. thanks!
12:23	justin_smith	marcfontaine: macros cannot use data that is not present when the form is compiled
12:24	justin_smith	that doseq hides the argument from the macro, the macro just sees "v"
12:24	justin_smith	err, just sees v, the symbol
12:25	justin_smith	at the time of compilation it does not run the macro on the args it would see at runtime - just the literal args in the form
12:25	justin_smith	marcfontaine: a macro is a function that takes a form as an argument and returns a new form. That form is compiled. Only later is anything run.
12:26	marcfontaine	I don’t follow, if I do (t “a”) that works if I do (doseq [v [“a”]] (t v)) that doesn’t work the only difference is that it’s in a doseq
12:26	justin_smith	marcfontaine: if you use intern instead of def you don't need to mess with macros, it's much simpler
12:26	justin_smith	marcfontaine: the difference is that the argument is v not "a"
12:27	justin_smith	marcfontaine: macros only run before compilation, and they are functions that transforms forms, macros do not see runtime data.
12:27	justin_smith	if you use a function instead of a macro, and the function uses intern, this is much easier. Just use intern instead of def.
12:28	justin_smith	,(doc intern)
12:28	clojurebot	"([ns name] [ns name val]); Finds or creates a var named by the symbol name in the namespace ns (which can be a symbol or a namespace), setting its root binding to val if supplied. The namespace must exist. The var will adopt any metadata from the name symbol. Returns the var."
12:29	justin_smith	,(intern 'sandbox 'a "foo")
12:29	clojurebot	#'sandbox/a
12:29	justin_smith	,a
12:29	clojurebot	"foo"
12:32	justin_smith	,(doseq [v ["b" "c" "d"]] (intern 'sandbox (symbol v) v))
12:32	clojurebot	nil
12:32	justin_smith	,b
12:32	clojurebot	"b"
12:32	justin_smith	,c
12:32	clojurebot	"c"
12:32	justin_smith	,d
12:32	clojurebot	"d"
12:48	rhg135	I love cljs
12:49	marcfontaine	justin_smith I need this for clojurescript and intern is not implemented
12:56	justin_smith	marcfontaine: oh, that's rough, because that also rules out eval (the only way to use def with a name only known at runtime)
12:56	justin_smith	marcfontaine: what about using a hash-map with keys instead of defs in a namespace?
12:59	justin_smith	marcfontaine: oh, the other option is the fact that "def" doesn't create vars in js, it just creates js variables. So you could just use interop on the namespace or on the window itself
13:00	marcfontaine	it’s for using bootstrap classes from reagent https://gist.github.com/marcandrefontaine/c32b566b2e981d867967 right now I have to do them one by one, I am trying to rewrite the macro by passing it all the class names but I can’t figure it out.
13:03	justin_smith	marcfontaine: yeah, macros can't use runtime data. What about just using aset? aset does in cljs what intern does in clj, considering that vars don't exist in cljs.
13:04	justin_smith	otherwise you could make a macro that expands to all the def calls in one body - eg. you would pass all the names to the macro, rather than calling the macro inside a doseq, the macro would have a for or map inside it to do all the defs
13:05	marcfontaine	yes that’s what I was trying to do with a doseq in the macro which didn’t work at all I’ll try with the for/map
13:07	justin_smith	marcfontaine: yeah, doseq is totally useless in a macro - a macro must return the form you want to evaluate
13:09	justin_smith	,(defmacro defs [& bindings] (cons 'do (for [[sym value] (partition 2 bindings)] (list 'def sym value)))) ; marcfontaine
13:09	clojurebot	#'sandbox/defs
13:09	justin_smith	,(defs a "a" b "b" c "c" z "z")
13:09	clojurebot	#'sandbox/z
13:09	justin_smith	,[a b c z]
13:09	clojurebot	["a" "b" "c" "z"]
13:10	justin_smith	marcfontaine: the above should work in cljs
13:12	marcfontaine	beautiful
13:16	rhg135	Actually, I'm pretty sure vars exist in some form on cljs so def would be the way to go
13:28	justin_smith	rhg135: no, they explicitly do not exist in cljs
13:29	justin_smith	rhg135: unless the clojurescript wiki is severely out of date
13:29	justin_smith	https://github.com/clojure/clojurescript/wiki/Differences-from-Clojure#vars-and-the-global-environment
13:32	justin_smith	rhg135: oh, my bad, the wiki is out of data
13:32	justin_smith	*date
13:32	sdegutis	Is there a more convenient way to use clojure.data.csv than (write-csv) with a temporary java.io.StringWriter which you just call (str) on afterwards in order to get the CSV as a string?
13:32	sdegutis	Please tell me the answer is yes, because I'm certain it'sn o.
13:32	justin_smith	rhg135: if vars exist, then it should have intern :(
13:34	rhg135	They usually don't at runtime
13:35	justin_smith	rhg135: ahh, so vars are a thing but only in the compiler, not in the runtime
13:35	justin_smith	ick
13:35	rhg135	sdegutis: with-out-str
13:35	justin_smith	rhg135: it doesn't print anything interesting
13:36	sdegutis	rhg135: Oh right! Why didn't I think of that? Actually, I remember thinking of that, and it for some reason not being viable in this usage.
13:36	rhg135	Oh, then no
13:36	sdegutis	I can't remember why I thought so. Hmm.
13:36	justin_smith	rhg135: the issue is that instead of returning or printing a string, it wants an arg it can write to, which means making a stringwriter
13:36	matthavener	anyone know how to get leiningen to copy CSS files from a cljsjs package into my resources dir?
13:36	sdegutis	Oh right, that's why.
13:36	sdegutis	I should probably just sign the dang CSV or whatever it's called, and send a PR through Jira.
13:37	puredanger	sdegutis: http://clojure.org/community/contributing
13:39	sdegutis	Yeah that.
13:40	sdegutis	puredanger: if I go through all this hassle do you think there's a chance that I could get clojure.data.csv fixed to not be so inconvenient?
13:40	puredanger	sure - what's the issue again?
13:40	sdegutis	(The beautiful thing about Github Issues and PRs is that it's really quick/easy to figure that question out.)
13:41	sdegutis	puredanger: write-csv is inconvenient when all you want is a string
13:41	sdegutis	puredanger: you have to create your own StringWriter and then call (str) on it and discard it. 3 lines to do something that should take 1.
13:41	rhg135	(defn as-str [f] (str (doto (StringWriter.) (f))))
13:41	sdegutis	puredanger: not the end of the world obviously, but not as convenient as it should be.
13:41	sdegutis	rhg135: yes I have that now, and I only use it for this one use-case and would like to get rid of it
13:41	puredanger	what would you propose?
13:42	sdegutis	puredanger: this function be inlined as a convenience function called csv-to-string or something that just wraps write-csv with a StringWriter and returns the inner string
13:43	puredanger	seems like a reasonable request to me
13:43	sdegutis	puredanger: (defn csv-to-string [x] (let [out (java.io.StringWriter.)] (write-csv out x) (str out)))
13:43	sdegutis	Okay. Then I'll do it!
13:43	puredanger	it's an e-form, no scanner required
13:44	sdegutis	Woo!
13:45	puredanger	see http://dev.clojure.org/display/community/Developing+Patches for making patches
13:45	puredanger	I need to move that stuff onto clojure.org
13:45	sdegutis	puredanger: Am I right to understand that it's common to not need to fill out the top three fields (and in fact you can't) in the e-form? Or am I totally misunderstanding this thing?
13:46	puredanger	yeah, those populate based on later fields
13:46	puredanger	which is weird, but I can't change it
13:47	sdegutis	Ok just checking thanks alex :)
13:48	sdegutis	puredanger: woo signed and verified! thanks for the link
13:48	puredanger	I see it! :)
13:49	sdegutis	:D
13:50	puredanger	http://dev.clojure.org/jira/secure/Signup!default.jspa
13:50	TimMc	Oh, how the process has changed.
13:50	TimMc	A couple years back I would not have seen a ":D" anywhere near a discussion around the contributer agreement. :-P
13:51	justin_smith	haha
13:51	justin_smith	TimMc: I had to walk uphill both ways in the snow to get to the snail mail
13:51	TimMc	I had to hand crank the fax machine all night in case something came through.
13:51	puredanger	hey, there's still a filing cabinet somewhere with printed versions of all these :)
13:51	justin_smith	maybe I shouldn't ever quote a cosby joke ever again, now that I consider it
13:53	TimMc	justin_smith: Because stuff?
13:53	justin_smith	TimMc: stuff pretty much - I don't like the idea of repeating his jokes any more, after hte other stuff he did. I dunno, maybe it's no big deal.
13:55	TimMc	No one is all good or all bad. I think it's fine.
13:56	TimMc	but I hear you
14:04	sdegutis	Tests pass!
14:07	puredanger	I hope so. :) CI runs them all the time so it would be surprising if they did not.
14:11	sdegutis	puredanger: so, the instructions I'm following say to include the JIRA ticket's number in the commit, so I'm creating a JIRA ticket before committing.. but the JIRA ticket says it requires a patch, and I have no idea how to create patch files but I have a commit prepared (all except the name). I feel kind of stuck on how to move forward.
14:11	puredanger	create the ticket. create the patch. add the patch to the ticket.
14:11	rcassidy	sdegutis: git format-patch -1 HEAD
14:11	rcassidy	If HEAD is the commit with what you wanna make a patch based on
14:11	sdegutis	rcassidy: thanks!
14:11	rcassidy	np
14:11	rcassidy	and that's a 1 (one) not L
14:11	sdegutis	I just copy/pasted it ;)
14:11	sdegutis	Wow, I have a newfound appreciation for GitHub's Pull Request based workflow.
14:11	puredanger	if you do it like 20 times a day for a few years, it gets really easy
14:11	sdegutis	Haha.
14:11	ToxicFrog	sdegutis: yeah, this is why I never submit patches to clojure and only rarely to bup
14:12	sdegutis	Creatd!
14:12	sdegutis	Woo!
14:14	sdegutis	puredanger: are the appropriate people notified when I create a ticket including a code/test patch like I just did?
14:14	puredanger	usually, although you created it in the wrong project (CLJ not DCSV)
14:14	puredanger	I moved it
14:15	sdegutis	Dang, sorry. Thanks puredanger.
14:16	puredanger	Jonas Enlund is the project owner so you could ping him too
14:16	sdegutis	puredanger: Okay now I see how to find the right project and create it there.
14:16	puredanger	I think it should be set up so that happened on creation, but not positive (not all of the contribs are set up the same)
14:18	sdegutis	Wow Jira is something else.
14:18	sdegutis	How do you @ping someone in a comment, the @at syntax isn't pulling up a dropdown list of names?
14:19	puredanger	Jonas is "jonas" on slack if you're ever on there or @jonasenlund on twitter
14:20	puredanger	in jira, I forget - you can look it up in the formatting instructions
14:21	sdegutis	Phew glad that's done with.
14:22	TimMc	HAHAHA it has only just begun
14:22	sdegutis	Thanks puredanger for all your help with this process, and you have my condolences that you have to go through that every day.
14:22	puredanger	it's really not that bad
14:22	puredanger	if you don't hear from Jonas in the next week, ping me
14:23	sdegutis	puredanger: nah don't worry about it, he'll see it when he sees it
14:41	sdegutis	justin_smith: I was pretty convinced you were right about the with-out-str thing not working, but now I don't remember why, and it's the thing Jonas is proposing as to why the helper function isn't necessary.
14:42	sdegutis	Hmm. It does in fact work.
14:42	justin_smith	sdegutis: oh does it like use out by default and I forgot it did that?
14:42	sdegutis	justin_smith: yeah
14:42	sdegutis	justin_smith: wait the CSV lib? no
14:42	justin_smith	sdegutis: oh so you pass in out and surround that with with-out-string ?
14:42	sdegutis	,(do (require 'clojure.data.csv) (with-out-str (clojure.data.csv/write-csv out [["a" "b"] ["c" "d"]])))
14:42	clojurebot	#error {\n :cause "Could not locate clojure/data/csv__init.class or clojure/data/csv.clj on classpath."\n :via\n [{:type java.io.FileNotFoundException\n :message "Could not locate clojure/data/csv__init.class or clojure/data/csv.clj on classpath."\n :at [clojure.lang.RT load "RT.java" 456]}]\n :trace\n [[clojure.lang.RT load "RT.java" 456]\n [clojure.lang.RT load "RT.java" 419]\n [clojure.co...
14:42	justin_smith	OK right
14:43	sdegutis	However, I thought one was a printer and one was a writer.
14:43	sdegutis	So I thought it just simply wouldn't work.
14:43	justin_smith	sdegutis: is that really that much better than a StringWriter?
14:43	sdegutis	No, I personally think it's still worth a patch.
14:43	sdegutis	But it's a harder argument to make.
14:47	amalloy	i think c.d.csv does the right thing already, by working with streams (or writers, i forget) and not encouraging you to turn a csv into one giant string
14:48	amalloy	you already can do it if you want to, quite easily, with with-out-str, but you usually shouldn't, so there's no convenient shortcut for it. just like there's no dissoc for vectors because it's expensive to remove elements from the middle
14:48	justin_smith	that's a good point
14:48	sdegutis	amalloy: very good point
14:49	sdegutis	amalloy: how do you return a CSV from a Ring handler function without turning it into a string first?
14:49	rhg135	,(dissoc '[a b c] 1)
14:49	clojurebot	#error {\n :cause "clojure.lang.PersistentVector cannot be cast to clojure.lang.IPersistentMap"\n :via\n [{:type java.lang.ClassCastException\n :message "clojure.lang.PersistentVector cannot be cast to clojure.lang.IPersistentMap"\n :at [clojure.lang.RT dissoc "RT.java" 848]}]\n :trace\n [[clojure.lang.RT dissoc "RT.java" 848]\n [clojure.core$dissoc invokeStatic "core.clj" 1462]\n [clojure.c...
14:49	rhg135	cool
14:49	amalloy	sdegutis: ring accepts a lot of things other than strings
14:50	amalloy	notably, a lazy seq of strings
14:50	amalloy	so you can stream the response
14:50	sdegutis	amalloy: But still, what do you create and use as the out-param?
14:51	sdegutis	I've closed the issue so this isn't an argument in favor of the stringifying function. I'm just curious what's a better output writer than StringWriter.
14:51	amalloy	ring will also accept an InputStream, i believe; one option would be to use a pipedinputstream/pipedoutputstream pair
14:51	amalloy	but that's a bit clunky
14:52	amalloy	does c.d.csv have a row->string function or something?
14:52	amalloy	unhelpfully, it has that function but marks it private
14:54	hiredman	~private
14:54	clojurebot	not even once
14:54	justin_smith	yeah, I think the piped stream combo is actually the whizbang thing
14:54	hiredman	~o/
14:54	clojurebot	\o ... High five!
14:54	amalloy	personally i would just not use c.d.csv because it's just getting in the way. {:status 200, :body (interpose "\n" (for [row csv] (interpose "," row)))}
14:54	amalloy	this of course has the problem that it doesn't handle cell data with commas
14:55	amalloy	which would be easy to fix if c.d.csv exposed its functions for working with data smaller than a whole file
14:55	sdegutis	amalloy: right quoting is a nice behavior of this lib
14:55	sdegutis	amalloy: you know, I was actually thinking "why don't I just write this myself, it's not that fricken hard of a data format" but then I was like "no, no, I should be using the appropriate libs for things like this, that's the Right Way™" and decided to keep using this lib
14:56	amalloy	yeah, i agree with you on that, it really sucks to reimplement it
14:56	sdegutis	But you're right. In this case, I only use the lib in exactly one route, and in that case, there are no quoting issues and never will be.
14:56	amalloy	i think a much better PR to c.d.csv would be to make its private functions mostly public
14:56	amalloy	and then you could actually use it to stream to ring pretty easily
14:56	amalloy	or, as i said, you could dedicate a thread to streaming the response using piped io and the existing api
14:56	sdegutis	Meh, I'd actually much prefer a CSV API similar to clojure.pprint/print-table
14:57	justin_smith	,(let [o (java.io.PipedOutputStream.) i (java.io.PipedInputStream. o)] (.write o (.getBytes "hello\n")) (.close o) (slurp i))
14:57	clojurebot	"hello\n"
14:57	sdegutis	Where you give it the keys and a seq of maps and it just spits out the right CSV. That's a sweeter API.
14:57	sdegutis	,(clojure.pprint/print-table [{:a 1 :b 2} {:a 3 :b 4}])
14:57	clojurebot	#error {\n :cause "clojure.pprint"\n :via\n [{:type java.lang.ClassNotFoundException\n :message "clojure.pprint"\n :at [java.net.URLClassLoader$1 run "URLClassLoader.java" 366]}]\n :trace\n [[java.net.URLClassLoader$1 run "URLClassLoader.java" 366]\n [java.net.URLClassLoader$1 run "URLClassLoader.java" 355]\n [java.security.AccessController doPrivileged "AccessController.java" -2]\n [java.n...
14:57	amalloy	actually, something i just remembered: compojure (or maybe it's ring?) has a helper function for creating a piped io pair out of a function
14:57	amalloy	makes it super easy
14:57	justin_smith	amalloy: oh, nice
14:57	sdegutis	,(do(require'clojure.pprint)(clojure.pprint/print-table [{:a 1 :b 2} {:a 3 :b 4}]))
14:57	clojurebot	#error {\n :cause "require'clojure.pprint"\n :via\n [{:type clojure.lang.Compiler$CompilerException\n :message "java.lang.ClassNotFoundException: require'clojure.pprint, compiling:(NO_SOURCE_PATH:0:0)"\n :at [clojure.lang.Compiler analyze "Compiler.java" 6688]}\n {:type java.lang.ClassNotFoundException\n :message "require'clojure.pprint"\n :at [java.net.URLClassLoader$1 run "URLClassLoade...
14:58	sdegutis	,(do(require 'clojure.pprint)(clojure.pprint/print-table [{:a 1 :b 2} {:a 3 :b 4}]))
14:58	clojurebot	\n\| :a \| :b \|\n\|----+----\|\n\| 1 \| 2 \|\n\| 3 \| 4 \|\n
14:58	justin_smith	amalloy: that would also be easy to implement
14:58	amalloy	https://ring-clojure.github.io/ring/ring.util.io.html#var-piped-input-stream
14:58	amalloy	so it would be like: {:status 200 :body (piped-input-stream (fn [o] (write-csv o my-csv))))
14:58	amalloy	modulo mismatched delimiters
14:59	hiredman	how perfect would it have been if that was also a private function
14:59	amalloy	weavejester knows better though
15:01	sdegutis	amalloy: really!?
15:01	sdegutis	sweet!
15:02	sdegutis	amalloy: it's kind of ugly though because you have to write a helper function, but still it's mostly nicer
15:02	amalloy	what helper function?
15:02	sdegutis	amalloy: on your example the (fn [o] ...)
15:02	sdegutis	I also like that Compojure allows you to return a variety of formats.
15:02	sdegutis	https://github.com/weavejester/compojure/blob/master/src/compojure/response.clj
15:03	amalloy	i think "helper function" generally means something quite different from that lambda
15:03	sdegutis	amalloy: wrapper function then
15:30	sdegutis	Every time I need to use defrecord and definterface I have to look them up to see which one of them takes a "self" first-param and which doesn't.
15:30	clojurebot	Gabh mo leithscéal?
15:30	sdegutis	Is there a good mnemonic for this?
15:31	ystael	sdegutis: "conj.io"
15:31	ystael	:p
15:31	sdegutis	???
15:31	sdegutis	Oh https://www.conj.io/
15:31	sdegutis	ystael: is that your site?
15:31	ystael	which doesn't, at all, address your issue
15:31	ystael	namely having to look them up :)
15:31	ystael	no, I forget who runs it, it isn't me
15:33	Kamuela	Which is the correct answer? Which requires self?
15:34	sdegutis	Kamuela: defrecord apparently
15:34	sdegutis	(doc defrecord)
15:34	clojurebot	"([name [& fields] & opts+specs]); (defrecord name [fields] options specs*) Options are expressed as sequential keywords and arguments (in any order). Supported options: :load-ns - if true, importing the record class will cause the namespace in which the record was defined to be loaded. Defaults to false. Each spec consists of a protocol or interface name followed by zero or more method bodies: protocol-or-interface-or-Object (metho
15:35	sdegutis	"Note that a parameter must be supplied to correspond to the target object ('this' in Java parlance). Thus methods for interfaces will take one more argument than do the interface declarations.”
15:35	Kamuela	I would just repeat record yourself over and over
15:37	sdegutis	Kamuela: that didn't work for hippocampus in college
15:37	sdegutis	Kamuela: when I try to remember what hippocampus means, I think of a hippo standing on top of the campus building shouting
15:37	sdegutis	Kamuela: I have absolutely no idea what it's supposed to represent
15:38	Kamuela	Memory
15:38	sdegutis	But how does that relate to its function?
15:38	ystael	sdegutis: "horse area"
15:38	sdegutis	stupid mnemonic
15:38	Kamuela	Amygdala emotion
15:39	Kamuela	Cerebellum balance? Not actually sure anymore
15:42	sdegutis	see?
15:42	sdegutis	useless
16:05	amgarchIn9	http://pastebin.com/P8T44STu
16:06	amgarchIn9	hi, is there something wrong with snag-quotes? I see some deadlock for n = 2048. New to clojure and dont know many debug techniques.
16:07	justin_smith	amgarchIn9: "launching" is done by calling future, you don't need to mapv deref, you just need (doall q)
16:07	justin_smith	amgarchIn9: what happens is the next future is not even started until the previous completes, because deref blocks and for is lazy
16:08	justin_smith	alternatively you could do the silly (mapv identity q), but just (doall q) instead
16:09	amgarchIn9	(mapv deref (vec q)) I forced the q and the performace degraded. Chunks of 32 seem to work for arbitrary n
16:09	justin_smith	amgarchIn9: when you map deref you only start at most a chunks worth at a time, (doall q) makes all the futures start immediately
16:10	justin_smith	if you want only N active at a time, use a proper thread pool instead of futures
16:11	hiredman	don't do io in a go block
16:11	amgarchIn9	well, snag-futures was there as a reference. The actual problem is with snag-quotes
16:12	hiredman	what does that have to do with anything?
16:13	hiredman	don't do io in a go block
16:14	hiredman	you most likely should be using pipeline-blocking
16:14	hiredman	https://github.com/clojure/core.async/blob/master/src/main/clojure/clojure/core/async.clj#L546-L550
16:17	amgarchIn9	thatnk, reading. (show s) was not there when the problem first occured. Without IO which (wget) also does there is no point in the excersize.
16:19	hiredman	right, so do your io, but don't use a go block
16:19	hiredman	pipeline-blocking runs each blocking task on a real thread https://github.com/clojure/core.async/blob/master/src/main/clojure/clojure/core/async.clj#L497-L500
16:39	sdegutis	Oh man I already forgot when to use defprotocol vs definterface :/
16:39	sdegutis	Oh right. I want defprotocol.
16:40	amalloy	also don't start an unlimited number of threads to perform IO at once. you'll just DoS your system when it reaches capacity
16:40	hiredman	which is why pipeline makes you pass in a `n`
16:41	hiredman	on a virtual machine a useful 'n' for io is going to be a relatively small number
17:02	amgarchIn9	that why I thout go blocks and "the state machine" on a predefined fixed size thread pool is fine. I have "netstat -aon \| grep <the ip here> \| wc" about 60. Fore some reason all "takes" from the channel are "delayed" (exclamation singns here http://pastebin.com/B05czQTB )
17:03	justin_smith	IO in go blocks just starves core.async's fixed size pool, use pipeline-blocking instead
17:04	ystael	justin_smith: i'm not familiar with core.async at all, is this analogous to other situations where you are doing blocking operations on a thing that needs to be a nonblocking asynchrony reactor?
17:05	justin_smith	ystael: pretty much - core.async has a state machine and runs multiple go blocks in a specialized thread pool, doing blocking IO in those go blocks subverts that design (and can lead to bad failures)
17:06	ystael	got it, i remember a similar thing from working with rxjava
17:06	justin_smith	ystael: simplified version is <! turns into a parked state meaning some other code can be run in your thread >! similarly parks but also wakes up whoever reads from the chan if they were parked
17:07	ystael	"open mouth" and "feed toddler"? :)
17:07	justin_smith	so it's a combination of coroutines and actual threading, copied from golangs goroutines
17:08	justin_smith	ystael: something like that - the way I remember it is (let [x (<! chan)] ...) the data is moving from chan into x
17:09	ystael	justin_smith: i tend to conceptualize all other areas of my life in terms of toddler management tasks at the moment :)
17:12	amalloy	justin_smith: <! and >! both have the same behavior re waking up other code, right?
17:13	amalloy	<! will wake up anyone who was parked trying to write, and >! will wake up anyone who was parked waiting to read
17:13	hiredman	amgarchIn9: mixing blocking tasks and non-blocking tasks in the same threadpool is problematic, the blocking tasks will starve the non-blocking tasks, and the level of parallelism to get the best performance out of each type of task tends to be different
17:14	hiredman	amgarchIn9: the threadpool core.async uses for go blocks is a global resource, any code libraries, whatever, you load could be using it
17:15	hiredman	amgarchIn9: as a shared resource it needs to be designated for blocking or non-blocking tasks (so different people don't use it for different kinds of tasks)
17:15	hiredman	it has been designated for non-blocking tasks
17:18	hiredman	and, more specifically, the core.async threadpool is a forkjoin threadpool, which is designed for compute bound tasks (https://stackoverflow.com/questions/8206318/is-javas-fork-and-join-thread-pool-is-good-for-executing-io-bound-task)
17:33	amgarchIn9	thank you all. Apparently posting to a channel does not necessarily wake up the reader. But other writers in my case. I got you message, just wanted to understand what is going on here.
17:49	ashnur	on an non unrelated but distant note... check this out: https://www.youtube.com/watch?v=K4vyRvMASPU
17:55	justin_smith	amalloy: oh yes, right
18:08	sdegutis	Hello. Good afternoon and I hope you're all doing well. What is the standard way of creating a Date that's precisely 300 seconds from now, without requiring third party libraries?
18:09	justin_smith	,(java.util.Date. (+ (.getTime (java.util.Date.)) (+ 1000 300)))
18:09	clojurebot	#inst "2016-02-25T23:09:07.031-00:00"
18:09	justin_smith	,(java.util.Date. (+ (.getTime (java.util.Date.)) (* 1000 300))) ; OOPS
18:09	clojurebot	#inst "2016-02-25T23:14:18.314-00:00"
18:09	sdegutis	justin_smith: Thank you. Do you happen to know why some people recommend using Calendar instead of that small simple solution?
18:09	justin_smith	sdegutis: the chance of a bug like using + instead of * ?
18:10	sdegutis	justin_smith: no, they still suggest using *, so the chance is still there
18:10	justin_smith	haha, no idea then
18:12	sdegutis	justin_smith: Also why not use this?
18:12	sdegutis	,(System/currentTimeMillis)
18:12	clojurebot	1456441941419
18:12	justin_smith	sdegutis: because I keep forgetting that exists
18:12	sdegutis	,(.getTime (java.util.Date.))
18:12	clojurebot	1456441955900
18:12	sdegutis	Well they seem to produce different values.
18:12	justin_smith	sdegutis: time has this habit of doing things
18:12	sdegutis	Several thousand units apart, too.
18:12	justin_smith	very mutable, that time thing is
18:13	sdegutis	,(= (.getTime (java.util.Date.)) (System/currentTimeMillis))
18:13	clojurebot	true
18:13	justin_smith	,(- (.getTime (java.util.Date.)) (System/currentTimeMillis))
18:13	clojurebot	0
18:13	justin_smith	well, then
18:14	sdegutis	So 1000 milliseconds = 1 seconds. Okay, will remember htis.
18:14	justin_smith	sdegutis: if our country used metric you would know by the milli prefix
18:14	justin_smith	maybe we need to use hogsheadseconds or something
18:15	amalloy	temporal ounces
18:17	TEttinger	time is measured by the accumulation of cheeto dust in the US department of weights and measures' trailer
18:19	rhg135	time is only an abstraction we use to explain how values succeed one another
18:20	sdegutis	,(java.util.Date/from (.. (java.time.LocalDateTime/now) (plusMinutes 5) (atZone (java.time.ZoneId/systemDefault)) (toInstant)))
18:20	clojurebot	#error {\n :cause "java.time.LocalDateTime"\n :via\n [{:type clojure.lang.Compiler$CompilerException\n :message "java.lang.ClassNotFoundException: java.time.LocalDateTime, compiling:(NO_SOURCE_PATH:0:0)"\n :at [clojure.lang.Compiler analyzeSeq "Compiler.java" 6875]}\n {:type java.lang.ClassNotFoundException\n :message "java.time.LocalDateTime"\n :at [java.net.URLClassLoader$1 run "URLClas...
18:20	sdegutis	,(import 'java.time.LocalDateTime)
18:20	clojurebot	#error {\n :cause "java.time.LocalDateTime"\n :via\n [{:type java.lang.ClassNotFoundException\n :message "java.time.LocalDateTime"\n :at [java.net.URLClassLoader$1 run "URLClassLoader.java" 366]}]\n :trace\n [[java.net.URLClassLoader$1 run "URLClassLoader.java" 366]\n [java.net.URLClassLoader$1 run "URLClassLoader.java" 355]\n [java.security.AccessController doPrivileged "AccessController.ja...
18:20	sdegutis	,(System/getProperty "java.version")
18:20	clojurebot	#error {\n :cause "denied"\n :via\n [{:type java.lang.SecurityException\n :message "denied"\n :at [clojurebot.sandbox$enable_security_manager$fn__835 invoke "sandbox.clj" 69]}]\n :trace\n [[clojurebot.sandbox$enable_security_manager$fn__835 invoke "sandbox.clj" 69]\n [clojurebot.sandbox.proxy$java.lang.SecurityManager$Door$f500ea40 checkPropertyAccess nil -1]\n [java.lang.System getProperty ...
18:20	sdegutis	haha stupid clojurebot
18:21	sdegutis	oh that lovable rascal
18:24	sdegutis	That said, I like this solution best: (-> (java.time.LocalDateTime/now) (.plusMinutes 5) (.atZone (java.time.ZoneId/systemDefault)) (.toInstant) (java.util.Date/from))
18:28	rhg135	why do you need to know about tz to add 5 minutes to your time
18:28	rhg135	unless minutes depend on zone
18:28	rcassidy	,(-> (java.time.LocalDateTime/now) (.plusMinutes 5) (.atZone (java.time.ZoneId/systemDefault)) (.toInstant) (java.util.Date/from))
18:28	clojurebot	#error {\n :cause "java.time.LocalDateTime"\n :via\n [{:type clojure.lang.Compiler$CompilerException\n :message "java.lang.ClassNotFoundException: java.time.LocalDateTime, compiling:(NO_SOURCE_PATH:0:0)"\n :at [clojure.lang.Compiler analyzeSeq "Compiler.java" 6875]}\n {:type java.lang.ClassNotFoundException\n :message "java.time.LocalDateTime"\n :at [java.net.URLClassLoader$1 run "URLClas...
18:28	rcassidy	:(
18:31	sevvie	we should use french metric time.
18:32	rcassidy	also, this is confusing -- i am using a script that uses the sum of the ascii values of a IRC nick to choose chat name colors
18:32	justin_smith	rhg135: clearly, you want a different result if there's a DST change between now and 5 minutes from now
18:32	rcassidy	,(mod (apply + (map int (seq "clojurebot"))) 256)
18:33	clojurebot	57
18:33	rcassidy	,(mod (apply + (map int (seq "rhg135"))) 256)
18:33	clojurebot	218
18:33	rcassidy	,(mod (apply + (map int (seq "sdegutils"))) 256)
18:33	clojurebot	212
18:33	rcassidy	hmm
18:33	rcassidy	you all look the same color of red to me
18:34	sdegutis	rhg135: I have no idea?
18:34	sdegutis	rhg135: I just know you can't get an #inst without doing that
18:34	justin_smith	rhg135: minor thing, that seq call is not needed, map already calls seq
18:34	justin_smith	err
18:34	justin_smith	rcassidy: ^
18:35	rcassidy	thanks justin_smith ! still lerning that stuff
18:35	rcassidy	*learning
18:36	rhg135	(inc justin_smith) ; I totally forgot about DST
18:36	justin_smith	(map int "lazybot is gone now :(")
18:36	justin_smith	,(map int "lazybot is gone now :(")
18:36	clojurebot	(108 97 122 121 98 ...)
18:36	rhg135	I might run that code a few minutes before the time shift
18:36	rcassidy	and I figured out it's because I'm not mod'ing by 256 in my actual irssi plugin, it's just mod'd over how many colors exist in the plugin, so maybe it isn't very many
18:36	rcassidy	too many collisions though
18:37	sevvie	,(mod (apply + (map int (seq "sevvie"))) 256) ; vanity thy name is...
18:37	clojurebot	146
18:38	justin_smith	,(mod (transduce (map int) + "justin_smith") 256) ; yet more vanity
18:38	clojurebot	33
18:38	sdegutis	,(let [a (java.util.Date.) b (java.util.Date.)] [(= a b) a b])
18:38	clojurebot	[true #inst "2016-02-25T23:38:58.227-00:00" #inst "2016-02-25T23:38:58.227-00:00"]
18:39	sdegutis	,(let [a (java.util.Date.) b (java.util.Date.)] [(= a b) a b])
18:39	clojurebot	[true #inst "2016-02-25T23:39:02.516-00:00" #inst "2016-02-25T23:39:02.516-00:00"]
18:39	sdegutis	Haha silly clojurebot.
18:39	justin_smith	rcassidy: fwiw I showed how to combine the apply and map into a single thing with transduce above
18:40	rcassidy	ah that's neat
18:40	rcassidy	,(mod (apply + (map int (seq "sdegutils"))) 15)
18:40	clojurebot	5
18:40	rcassidy	,(mod (apply + (map int (seq "rhg135"))) 15)
18:40	clojurebot	9
18:40	rcassidy	grrr
18:41	rcassidy	,(mod (apply + (map int (seq "clojurebot"))) 15)
18:41	clojurebot	1
18:41	rcassidy	this doesn't make sense!
18:41	justin_smith	,(mod (transduce (map int) + 0 "clojurebot") 15)
18:41	clojurebot	1
18:41	justin_smith	,(mod (transduce (map int) + "clojurebot") 15) ; aha, the 0 is optional
18:41	clojurebot	1
18:44	rcassidy	i wonder how i can make irssi tell me how many colors it's choosing from
18:44	sdegutis	Of course you can't.
18:45	rcassidy	:(
18:45	rhg135	`tput colors` can show you how many it should think to use
18:47	rhg135	terminals are weird with colors
18:48	rcassidy	yeah, tput colors gives me 256
18:48	rcassidy	and i'm reading my irssi script, but i don't think it picks from terminal colors
18:48	rcassidy	looks like it picks from a list 11 items long
18:49	rcassidy	but my own tests with mod 11 / 10 don't work
18:49	rhg135	it's a handy feature for sure
18:49	rhg135	you are yellow-orange
19:26	TEttinger	rcassidy: if you have an xterm-based terminal, it probably supports a certain escape code for 256-color selection, but almost all terminals support 16-color using a different escape code (notably windows cmd.exe doesn't support ansi escape codes)
19:28	TEttinger	https://github.com/shabble/irssi-docs/wiki/Notes-256-Colour
19:29	{blake}	Can laziness "get" you in React? Like, if your rendering routine contains a map or for or whatevs, will that potentially not be realized without forcing it?
19:30	{blake}	'cause I got a situation where I'm changing the underlying data atom and it's not updating the screen.
19:31	{blake}	But the rendering function is being called.
19:32	sdegutis	Wooo!
19:32	{blake}	Who?
19:32	sdegutis	Just replaced all our clj-aws-s3 usages with plain old aws-java-sdk Java stuff!
19:33	sdegutis	That was one situation where using the Java-based SDK actually resulted in fewer lines of code than using the Clojure wrapper around it.
19:34	TEttinger	wow
19:34	TEttinger	clojure does have durn good interop
19:35	sdegutis	Well in this case it's mainly because of the design of the wrapper.
19:36	TEttinger	yeah, I just think it's nice that wrappers aren't always needed
19:37	TEttinger	and you can just use the java one "just like that"
19:37	sdegutis	:)
19:42	TEttinger	hm, just curious: is cfleming or any other cursive expert around? I'm not sure if I'm experiencing a bug
19:44	TEttinger	this is my project.clj, very small, and cursive (but not lein) can't seem to see the deps. https://github.com/tommyettinger/squidlib-clj/blob/master/project.clj
19:44	TEttinger	lein runs just fine
19:54	{blake}	TEttinger: I have occasional problems with deps in cursive, also.
19:54	{blake}	Sometimes I just restart IJ.
20:11	Kamuela	I actually think cider and paredit are working with Emacs. I'm just not doing well with it yet
20:22	TEttinger	{blake}, hm, I've restarted at least 3 times
20:22	{blake}	TEttinger: I'm sorry, I didn't mean to imply that it actually worked, just that I did it. =P
20:23	TEttinger	oh haha
20:23	TEttinger	it's odd that there's no leiningen context menu like there is for maven and gradle
20:24	{blake}	Hey, did you "lein clean"?
20:24	TEttinger	hm, no
20:25	{blake}	I've been having to do that a ton lately. Not sure what evil I'm perpetuating.
20:26	{blake}	I'm thinking I don't have a handle on this whole react/reagent thing. My desire to set a defaultValue launches me into this whole world of "controlled components" where I'm now manually updating the app state.
20:26	TEttinger	also my "External Libraries" in cursive just shows leiningen and Java 8
20:27	{blake}	TEttinger: Where is that? It's not "Libraries", right?
20:29	TEttinger	not sure. http://i.imgur.com/wDJ0pOm.png
20:30	{blake}	Oh, in the project navigator thingy.
20:30	{blake}	Yeah, that doesn't seem right at all.
20:30	{blake}	I get carpal tunnel trying to go trhough my External Libraries.
20:30	TEttinger	I agree
20:32	TEttinger	it's like it isn't really a clojure project
20:32	{blake}	Oh, could that be it?
20:32	TEttinger	no idea
20:33	{blake}	I've been using Cursive for so long, I haven't checked lately as to its project generation. I always do a "lein new" from the command line.
20:33	TEttinger	there's no clojure facet I can add. I think I did lein new as well
20:33	{blake}	But trying to create a new Clojure project =inside= IJ has never worked out well for me.
20:34	TEttinger	I don't know how I made it, but I think it was "lein new"
20:35	{blake}	When I do that, I get the 1.8 JDK, clojure-complete, clojure 1.8 and nrepl, and that's it for exterinal libraries.
20:37	TEttinger	yeah, looks like that for me
20:38	TEttinger	how do you add deps? does adding them to project.clj automatically pop those suggestions up in sources when it's working?
20:39	{blake}	TEttinger: Yes, eventually. Sometimes it's real smooth. Other times, not so much.
20:39	{blake}	Like I just added in clj.time, and it's not seeing that.
20:40	TEttinger	hm, ok
20:40	{blake}	Oh...I may also do a "lein deps" from the CL.
20:41	{blake}	But I don't think that's necessary any more.
20:42	{blake}	Be worthwhile to ask cfleming about: how deps get into the environment and what we're doing wrong.
20:42	TEttinger	it works in the repl in cursive but not in the editor, oddly
20:42	{blake}	Like, I just added clj-time and it didn't show up. I restarted twice, now it's in the REPL.
20:43	{blake}	I've seen that, too. I've gone cuckoo trying to figure out that one.
20:44	{blake}	I've got that now with my simple clj-time example. Works in REPL, editor doesn't see it.
20:49	{blake}	Yeah, I don't get this thing. I've got this atom with my app data in it. I've changed it, which triggers a redraw. "Controlled components" have onChange events which...change the data in the atom, which (it seems to me) would trigger the redraw again.
20:53	TEttinger	it's definitely odd
20:53	TEttinger	thanks for confirming that this is not isolated to my computer, {blake}
20:53	TEttinger	or my miiiiiind
20:54	{blake}	It's AAALLL in your MIIIIND!
20:55	{blake}	Yeah, love Cursive, but this is a weak spot.
20:57	TEttinger	I mean I'm trying to make a clojure wrapper for this java lib, and not being able to see the lib I'm wrapping is trouble
22:10	sdegutis	TIL you can't merely (import) a namespace you created containing a defprotocol, you have to (require) it first.
22:45	rhg135	Does that work with a non-aot definterface?