#clojure – 2012-12-20

2012-12-20

00:06	mehwork	apt-get install clojure installed 1.1, but i want clojure 1.4. what's the best way to get it? can i just add 1.4 to my leiningen's :dependencies?
00:08	tpope	yes
00:08	tpope	don't bother with apt for anything other than maybe java itself
00:08	Raynes	Never use a package manager for anything Clojure related besides Java.
00:08	Raynes	You'll wake up naked and covered in blood.
00:09	mehwork	thanks
00:10	Raynes	Leiningen is totally independant of anything else. It downloads Clojure versions as necessary and manages all the pain for you to keep you waking up clothed, or at least clean with no blood.
00:10	mehwork	it's insane not having basic string splitting functions and stuff in 1.1
00:10	Raynes	There is stuff like that in Java itself that you used in 1.1
00:10	Raynes	But yeah, 1.1 is way too old.
00:10	Raynes	You definitely don't want that.
00:10	mehwork	i wanna stay as far away from java as possible
00:11	Raynes	Then you might be using the wrong languag.
00:11	Raynes	language*
00:11	mehwork	probably
00:11	Raynes	Everything calls out to Java somewhere. There are a lot of Clojure libraries around, but there are lots of Java libraries too, and chances are you'll find yourself wanting one eventually.
00:12	mehwork	then i'll jus wrap them
00:12	Raynes	Thought you wanted to stay away from Java? ;)
00:12	mehwork	as much as possible
00:13	Raynes	Words, words.
00:51	bbloom	when viewed with the experience of values and identities, the browser dom is completely fucking hopeless
01:13	tomoj	it's hopeless because if you build something decent on top of it, it will be too slow?
01:13	tomoj	or because you just can't build anything decent on top of it
01:13	bbloom	tomoj: both. heh
01:14	bbloom	i'm tempted to just write my own complete gui toolkit on canvas
01:14	bbloom	apparently even mozilla is toying with that idea: https://wiki.mozilla.org/Labs/Thunderhead
01:16	Raynes	Obviously stealing your idea.
01:16	Raynes	Also, my word. They need to quit with the names.
01:16	Raynes	What's next? ElectricStormParakeet
01:17	brainproxy	is the datomic crew more or less moving away from the ".dtm" filename extension convention in favor of ".edn"?
01:18	tomoj	why is using canvas underneath much easier?
01:19	bbloom	tomoj: the main problem is one of shit changing under you
01:19	bbloom	there are basically two approaches: play the DOM game with css/html/javascript, or own everything
01:20	bbloom	angular.js plays the DOM game
01:20	bbloom	cappichino or GWT owns everything
01:20	bbloom	in the own everything view, you can assume the dom only changes in controlled ways
01:20	bbloom	but it makes it hard for you to safely employ css, etc
01:20	bbloom	or custom event handlers
01:20	bbloom	or dom inspection
01:20	bbloom	if somebody changes some shit in the web inspector, you're fragile little view of the world may break
01:21	bbloom	if you wanted to have an immutible snapshot of the DOM state, you simply can't get it. because even simple things like event handlers aren't queryable from any dom API
01:21	bbloom	there is no way to say getAllEventHandlers
01:22	bbloom	the advantage of owning everything is that you can do more powerful stuff, like dynamic layouts. you just absolutely position everything via javascript
01:23	tomoj	jquery's deep clone only copies event handlers which were added by jquery, I assume
01:23	bbloom	but then at that point, you're just using the dom as a finely optimized rectangle renderer with some native widgets for text input. and you'll want to skin everything for a professional app anyway, so you might as well skip the rectangle renderer and render arbitrary lines and shit :-P
01:23	bbloom	correct
01:24	bbloom	also, i used to do game dev, sooo i part of me wants to say fuck it and go off and implement a compositor
01:24	bbloom	that would be easier than dealing with the fucking dom :-P
01:25	bbloom	Raynes: are you seriously going to go implement this?
01:26	ro_st	implement what?
01:26	bbloom	canvas gui toolkit w/ clojurescript :-)
01:26	ro_st	from scratch?
01:27	bbloom	shrug
01:27	ro_st	ambitious.
01:28	ro_st	when it comes to rendering, you want short stacks for performance. clojurescript doesn't make for short stacks
01:28	bbloom	"short stacks" ?
01:28	ro_st	call stacks
01:28	ro_st	stack as in stack overflow
01:29	bbloom	yeah, but i'm not sure what you're getting at.... call overhead? since when does the height of the stack matter for performance?
01:29	ro_st	although, i've never done anything with canvas, so i could be totally wrong.
01:30	Raynes	bbloom: Am I going to what what?
01:30	ro_st	the less stuff that happens in hot loops, the faster they go. if your logic has to work through the cljs abstractions to get at the 'raw' js in that loop, that means you're losing some performance
01:30	bbloom	Raynes: I said canvas gui, you said "Obviously stealing your idea."
01:30	ro_st	for the same reason games use c/c++ for their 3d engines instead of lua
01:31	Raynes	bbloom: No sir, I was implying Mozilla stole your idea.
01:31	bbloom	ah, heh
01:31	Raynes	I wouldn't touch that with a 10 foot career.
01:31	bbloom	haha
01:31	ro_st	-grin-
01:47	mehwork	is there a diff between (defn- main) and (defn -main) ?
01:48	tomoj	the former defines a private function called 'main, the latter a public function called '-main
01:49	mehwork	i take it it's a bad idea to do (defn- -main)
01:50	ro_st	only if you want to be able to call main from outside the ns
01:55	tomoj	event handlers seem problematic (in the don't-own-everything approach), but I don't feel like I want an immutable snapshot of the dom which includes event handlers
01:57	bbloom	tomoj: i'm not actually going to build a canvas gui :-P just venting a bit. i'm still exploring the design space, so i'm not exactly sure how to deal with it
01:58	bbloom	i think the preferable approach is closer to angular.js
01:58	bbloom	but i want to embrace clojure's data view of the world
01:59	bbloom	currently, i have a prototype working that has a data structure, recursively expands templates and "styles", much like macro expansion, and the expanded form can be rendered to a dom
01:59	bbloom	clearly rebuilding the entire dom on every update is inefficient
01:59	bbloom	so i'm searching for a way to intelligently apply the delta to the dom
02:01	bbloom	it seems the only valid strategy is to assume no one is going to open the web inspector and muck with your elements. and you just pray that no extension is corrupting the dom in any serious fashion
02:02	ro_st	not a valid assumption to make at all. lastpass does stuff to the dom, for example
02:02	ro_st	both when rendering its noticebar and when populating fields with data
02:02	bbloom	hence the "serious fashion" qualifier
02:02	bbloom	it's one thing to add a top level div
02:02	bbloom	it's another thing to insert random elements and modify random properties scattered throughout the page
02:03	ro_st	true
02:03	bbloom	and when populating fields with data, a well behaved plugin would most likely trigger a value-change event
02:03	bbloom	which would be more or less indistinguishable from the user typing in that field
02:05	mehwork	is it wrong to put -main in core.clj? core.clj has (ns foo.core) and my project.clj has :main foo.core but when i lein run -m src/foo/core.clj i get an error
02:05	mehwork	it seems to work fie if i put the -main function in a diff file and set :main to that though
02:07	ro_st	and if you just lein run?
02:07	ro_st	no -m path
02:07	tomoj	(btw it's -m ns, not -m path: `lein run -m foo.core`)
02:08	tomoj	bbloom: the thing you're describing is one-way only, right? data->dom?
02:08	bbloom	tomoj: i have one-way working with complete-re-render
02:09	bbloom	tomoj: my goal is full two-way
02:09	bbloom	tomoj: with efficient incremental updates
02:09	bbloom	my little template/styles macro expanding jazz is actually quite simple and pleasant to work with
02:09	bbloom	it's roughly analogous to microsoft's WPF logical and visual trees
02:09	bbloom	basically, i have some "logical tree" which is "template expanded" into a visual tree
02:10	bbloom	and the visual tree is roughly a static snapshot of the target dom
02:10	bbloom	so i just hiccup-ize it right now for proof of concept
02:10	bbloom	works nicely
02:11	mehwork	tomoj: hmm thanks. lein run by itself worked but lein run -m foo.core still errors
02:12	tomoj	so ideally, in the data->dom case, web inspector tweaks would just be immediately undone? or a bit less ideally, next time the data changes?
02:14	bbloom	tomoj: i'm not super worried about web inspector tweaks, i just mention them as an annoyance
02:15	bbloom	tomoj: in theory, they'll immediately be overwritten by a re-render for that subtree
02:15	tomoj	guess it's irrelevant since we want two-way. I don't understand what two-way would look like
02:16	bbloom	so angular works by having "directives" for each bindable thing
02:16	bbloom	for example, ng-checked http://docs.angularjs.org/api/ng.directive:ngChecked
02:16	bbloom	i'm thinking that i'll do something similar
02:17	bbloom	something like {:tag "color" :attr/title "you see this on hover" :css/color "red" :bind/click xyz}
02:18	bbloom	er tag shoulda been "div"
02:18	bbloom	brain fart
02:20	tomoj	so the model is an identity and dom changes and bound to mutate it
02:20	bbloom	something like that
02:21	bbloom	my goal is an order of magnitude productivity win over everything else out there :-)
02:21	bbloom	it's a hobby project though... if i had to do real work starting tomorrow, i'd just suck it up and use angular + coffeescript :-P
02:34	tomoj	so you need one function to update a model given input from the dom, and one to update the dom given the model. and a binding is both functions?@
02:39	alex_baranosky	what's the current status of the Korma project? A quick look on clojars shows more than ten non-official Korma jars there
03:02	mehwork	is it bad to def a var, then on the next line def it again by apply'ing something to it? Effectively creating and manipulating the same state?
03:02	mehwork	or what is the convention
03:03	ro_st	not bad. you might want to (def foo (let [foo (something)] <alter foo>))
03:04	ro_st	the let binding replaces your first def
03:04	mehwork	i see, thanks
03:07	mehwork	where can i read about <alter>?
03:07	ro_st	that's just me saying "do something with foo"
03:07	ro_st	<alter> isn't actual clojure syntax
03:07	mehwork	oh :p
03:07	ro_st	sorry for the red herring :-)
03:07	mehwork	it's okay. I'm relieved
03:08	ro_st	(def foo (let [a 1] (inc a))) ; foo is 2
03:10	alex_baranosky	mehwork: imo that is bad form. something you only do if you absolutely must
03:11	alex_baranosky	mehwork: if you have mutable state why not use an atom or ref for it?
03:11	alex_baranosky	mehwork: do you really need this mutable state? <-- that's the real question. Usually the answer is "no".
03:12	mehwork	idk, i'm still a complete noob
03:15	ro_st	baby steps.
03:15	ebaxt	I'm having a "there
03:15	ebaxt	I'm having a "there's a function for that" moment, anyone? https://www.refheap.com/paste/7744
03:17	ro_st	ebaxt: looks like group-by might help
03:18	ebaxt	ro_st: I don't want a vector for each key since I know they are unique
03:35	amalloy	(into {} (for [x coll] [(:foo x) x]))?
03:35	amalloy	(into {} (map (juxt :foo identity) coll))
03:36	amalloy	ebaxt: ^
03:37	tomoj	my port of clojure.test to clojurescript seems to be basically working
03:38	mehwork	how do you convert all integer key values to strings in a map? eg {1:"foo", 2:"bar"} becomes {"1":"foo", "2":"bar"}?
03:38	mehwork	do i need to use doseq
03:39	tomoj	(into {} (for [[k v] m] [(str k) v]))
03:39	ro_st	(into) is a great bit of kit
03:39	tomoj	I wish it were (into-kv {} (map-k str m)), but it's not
03:41	mehwork	or is there an option to do it during map-indexed hash-map seqname
03:41	mehwork	it would be nice if map-indexed took a callback to filter through first
03:43	tomoj	wait, what? you're doing something like ##(map-indexed {1 :foo 2 :bar} [:baz :bing :bang :bong]) ?
03:43	lazybot	⇒ (:baz :foo :bar :bong)
03:44	amalloy	hah. i never thought of using a map as the map-indexed function
03:44	mehwork	interesting
03:45	tomoj	or, or ##(map-indexed hash-map [:foo :bar :baz])
03:45	lazybot	⇒ ({0 :foo} {1 :bar} {2 :baz})
03:45	ebaxt	amalloy: Thx, shorter at least ;)
03:45	tomoj	for the latter, ##(map-indexed #(hash-map (str %1) %2) [:foo :bar :baz]) :/
03:45	lazybot	⇒ ({"0" :foo} {"1" :bar} {"2" :baz})
03:46	noidi	,(zipmap (range) [:foo :bar :baz])
03:46	clojurebot	{2 :baz, 1 :bar, 0 :foo}
03:46	noidi	ah, that's not what you were doing
03:47	mehwork	that told me nested #()'s are not allowed when done inside a let
03:48	mehwork	noidi: what's the , mean
03:50	ro_st	it tells clojurebot to execute the code
03:50	ro_st	,(inc 41)
03:50	clojurebot	42
03:50	ro_st	,(* 3 7 2)
03:50	clojurebot	42
03:51	mehwork	,(= 42)
03:51	clojurebot	true
03:51	mehwork	,(= 42 "meaning of life")
03:51	clojurebot	false
03:53	tomoj	the double # tells lazybot to eval: ##(+ 1 2)
03:53	lazybot	⇒ 3
03:53	tomoj	&(+ 1 2) ; (or initial &)
03:53	lazybot	⇒ 3
03:54	mehwork	why does zipmap return the map reversed
03:55	ro_st	it's conjing onto a list, which conj's on to the beginning
03:55	ro_st	vectors conj onto the end
03:56	ro_st	this is for performance reasons
03:56	ro_st	not because it's sensible -grin-
03:56	mehwork	is that to do with lazy sequencing
03:56	ro_st	,(conj (list 1 2) 3)
03:56	clojurebot	(3 1 2)
03:56	tomoj	&(-> {} (assoc 1 2) (assoc 3 4) (assoc 5 6))
03:56	lazybot	⇒ {5 6, 3 4, 1 2}
03:56	tomoj	no list in zipmap
03:56	ro_st	,(conj [1 2] 3)
03:56	clojurebot	[1 2 3]
03:57	ro_st	oh. then it's something else. but what i showed about conj and lists/vectors is still good to know
03:58	clgv	what's the question?
03:58	tomoj	&(zipmap (take 9 (range)) (range))
03:58	lazybot	⇒ {8 8, 7 7, 6 6, 5 5, 4 4, 3 3, 2 2, 1 1, 0 0}
03:58	tomoj	&(zipmap (take 10 (range)) (range))
03:58	lazybot	⇒ {0 0, 1 1, 2 2, 3 3, 4 4, 5 5, 6 6, 7 7, 8 8, 9 9}
03:58	mehwork	yeah the docs for conj say they may be in different places depending on the concrete type
03:58	clgv	mehwork: that's what the scala people did not like as well ;)
04:01	tomoj	is there any hope for denotational semantics in clojure - not for proofs, just for design aid
04:04	clgv	tomoj: I think `conj` is the worst example and to a certain degree the querying and delting operations associated with it on sequential data.
04:05	clgv	,(peek '(1 2 3))
04:05	clojurebot	1
04:05	clgv	,(peek [1 2 3])
04:05	clojurebot	3
04:07	tomoj	do the semantic types include lists and vectors?
04:07	tomoj	IReduce?
04:08	tomoj	I guess I should read about semantics
04:10	mehwork	how do you sort a map by it's numerical keys? i tried (into (sorted-map) mymap) but it's coming out disordered
04:11	ro_st	,(into (sorted-map) {3 2 1 6 4 5})
04:11	clojurebot	{1 6, 3 2, 4 5}
04:11	clgv	mehwork: `sorted-map-by` let's you provide a comparator
04:12	ro_st	keys look sorted to me
04:13	tomoj	&(keys (zipmap (range 20) (range)))
04:13	lazybot	⇒ (0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19)
04:13	tomoj	coincidence!
04:14	tomoj	I misread as "#'keys looks sorted to me"
04:14	clgv	,(let [kvs (repeatedly 20 #(rand-int 100))] (println "kvs =" kvs "\nsorted-map =" (apply sorted-map kvs)))
04:14	clojurebot	kvs = (57 76 0 89 53 ...)
04:14	clojurebot	sorted-map = {0 89, 38 16, 42 15, 48 27, 49 16, ...}
04:15	clgv	&(let [kvs (repeatedly 20 #(rand-int 100))] (println "kvs =" kvs "\nsorted-map =" (apply sorted-map kvs)))
04:15	lazybot	⇒ kvs = (48 42 74 4 15 98 80 75 69 97 96 87 27 75 23 35 43 20 59 98) sorted-map = {15 98, 23 35, 27 75, 43 20, 48 42, 59 98, 69 97, 74 4, 80 75, 96 87} nil
04:15	alex_baranosky	is Korma compatible with Clojure 1.2.1?
04:16	clgv	alex_baranosky: you should put in the effort to upgrade. I've been sitting on 1.2.1 until 2-3 months ago. it's a pain when all the libs move on and you cant use a decent number of them anymore
04:17	clgv	&(let [kvs (repeatedly 20 #(rand-int 100))] (println "kvs =" kvs "\nsorted-map =" (seq (apply sorted-map kvs))))
04:17	lazybot	⇒ kvs = (9 2 32 84 86 5 34 7 63 9 66 47 97 86 62 71 19 68 82 33) sorted-map = ([9 2] [19 68] [32 84] [34 7] [62 71] [63 9] [66 47] [82 33] [86 5] [97 86]) nil
04:17	mehwork	ah sorted-map-by is what i needed cause now it's the values i need
04:19	alex_baranosky	cigv: in due time, but for the time being I am trying to fix a issue at work, by moving some old code off clj-record
04:20	alex_baranosky	I can do that faster than I can do a move to 1.4
04:55	clgv	alex_baranosky: guess you have to try korma with 1.2.1 to get an answer. a profile for clojure 1.2.1 in the project.clj would be a good sign for you
05:12	Raynes	alex_baranosky: Clojure 1.2.1 lolnoob
05:16	ejackson	Raynes: Now I know why you live out in the sticks - to avoid all the slaps !
05:16	Raynes	ejackson: Yeah, but that's only until February. Then I'll be in slapping distance of everyone I insult on the west coast.
05:18	ro_st	will me nattering on about how we've used Clojure end-to-end be interesting enough?
05:18	ejackson	I'll get you a gumguard for Christmas
05:18	ejackson	ro_st: experience reports are always super interesting.
05:18	Raynes	ejackson: Sweet
05:18	ro_st	what sort of details would be of value?
05:19	ejackson	edges, corners, when you had to do something ugly
05:19	ro_st	we're only launching on the 14th of jan, but by the time of the talk, we should be several thousand users strong
05:20	ro_st	we are using values-based programming end to end.
05:20	ro_st	using event sourcing for the main app's persistent data model
05:20	ro_st	datomic, clj, cljs, pallet
05:21	ro_st	i think there's probably enough meat in there
05:21	ejackson	yeah, people always like to hear about somebody actually winning !
05:22	ro_st	that's true
05:55	piranha_	ro_st: do you do in-browser app, or what is it in the end?
05:55	ro_st	it's a webapp, yes
05:55	ro_st	html5 multi-platform
05:55	ro_st	deskop/tablet for now. dedicated mobile ui sometime next year
05:56	piranha	aha, how do you do user interface? FRP?
05:56	ro_st	i'd love to get there. we probably have a half-baked FRP hiding in amongst our code
05:56	piranha	I see...
05:56	Raynes	Translation to human speak is: "it's all broke"
05:56	ro_st	we use shoreleave's remotes and pubsub along with enfocus for templating and event handling
05:57	piranha	I'm trying to get in by using Flapjax, so far it's... well, works somehow for simple things, not sure if it will be good for big stuff
05:57	piranha	I see
05:57	ejackson	Raynes: maybe a helmet, actually...
05:57	piranha	ro_st: how do you handle data on the client? :) Atoms? Or inside of the DOM? :)
05:57	ro_st	atoms
05:58	ro_st	but all the atoms are in one namespace; all the actual logic is pure
05:58	ro_st	on the client, the cljs uses the atom namespace as a singleton; on the server, i use binding to scope those atoms when working with the data
05:58	ro_st	we use lein-cljsbuild's crossovers to share the model code in clj/cljs
05:58	piranha	aha... I wrote my own abstraction for managing data in atoms, and I would love to hear some feedback (if I'm crazy or not, haha), clojure mailing list is keeping silence :))
05:59	piranha	ah, I see... I plan to write only browser part in clojure
06:00	ro_st	the atoms only hold transient data only; the persisted data is event sourced. so loading a document merely replays events into a running session and that reconstitutes the state
06:00	ro_st	never persist and repopulate atoms directly
06:00	ro_st	makes undo/redo dead easy to do.
06:01	ro_st	mind sharing a link to your post?
06:01	ro_st	Raynes: we're using the some things that FRP uses, like pubsub
06:02	Raynes	I wasn't the one who asked about FRP, but thanks for letting me know. ;)
06:02	ro_st	but by no means are we actually doing FRP. would love to spike out an FRP solution to properly understand it, but there's only so much damn time available.
06:03	ro_st	ah, i was responding to your 'it's all broke' comment :-)
06:03	ro_st	piranha: atoms + shoreleave-pubsub works really well
06:04	piranha	ro_st: https://groups.google.com/forum/#!topic/clojure/shkr_3AD3UM
06:04	piranha	ro_st: what means 'event sourcing'? I feel I don't really understand that :)
06:04	ro_st	ah, i've seen this. it's queued in my list
06:05	piranha	ah :)
06:05	ro_st	http://martinfowler.com/eaaDev/EventSourcing.html
06:05	piranha	thanks
06:06	ro_st	essentially, instead of transfering the full state, instead just transfer the user's actions, and re-constitute the full state by 'replaying' those actions in order
06:07	ro_st	stock markets are event sourced
06:07	piranha	I see...
06:07	piranha	hmmm, interesting
06:08	ro_st	why i love it is because we get undo/redo almost for free, and when we add websockets as a transport and start to do collaborative document stuff, all i need to do is add a bit of control flow logic and rewire pubsubs
06:08	ro_st	to have eg a websocket play events into a cljs workspace rather than keyboard/mouse events
06:10	ro_st	and we also have instant save/resume for mobile - no big state to upload. users just use the app and the rest happens like magic
06:11	piranha	indeed, sounds nice :)
06:12	ro_st	piranha: check shoreleave-pubsub. working demo code here: https://github.com/robert-stuttaford/demo-enfocus-pubsub-remote
06:12	piranha	ro_st: thanks! I will
06:13	piranha	reading about event sourcing now :)
07:29	ro_st	anyone connecting to nrepl servers via ssh?
07:30	ambrosebs	can anyone think of a statically typed language that has keywords that are also functions?
07:30	ambrosebs	or similar
07:36	ro_st	i don't know. that guy who did Typed Clojure seems like someone to ask, though
07:36	ro_st	-grin-
07:36	llasram	ambrosebs: Keywords in the syntactic sense, or in the Clojure sense?
07:38	clgv	ro_st: roflmao!
07:39	ambrosebs	llasram: (:a coll) looks up :a in its argument.
07:39	ro_st	cemerick_: thank you, thank you, thank you for nrepl
07:39	ambrosebs	I'll scratching my head how to make that not hacky
07:39	ambrosebs	*I'm
07:40	llasram	Oh. Hmm. Yeah -- the name defines what the value does. That is kind of weird when you think about it
07:42	ambrosebs	One idea I had was to parameterise Keyword with the key it looks up. :a is of (Keyword :a). But that has other issues.
07:42	clojurebot	Alles klar
07:42	cemerick	ro_st: :-) Anything in particular about it you're enjoying?
07:43	ambrosebs	I'm trying to get rid of my intersection type, but it comes in quite handy here. :a is just (I (Value :a) (All [x] [(U Any '{:a x}) -> (U x nil)]))
07:44	ro_st	datomic licenses by jvm process. starting up a repl consumes one process. when your production stack is already using all 3 available processes makes starting a repl impossible. but embedding nrepl server makes that unhappiness melt away :-)
07:45	ro_st	...processes, that makes*
07:47	ambrosebs	The definition of Keyword in CLJS gives a hint.
07:47	ambrosebs	We definitely need to abstract over its "k" field.
07:48	renderful	nice repl tip ro_st, that will come in handy one day
07:49	ro_st	also a lot saner when it comes to server ram
07:49	ro_st	makes me wonder why we've been hobbling along without nrepl embedded for so long
07:51	ro_st	cemerick: are you attending clojurewest?
07:52	cemerick	well, other REPL servers were certainly embeddable in the past, though perhaps not as frequently as nREPL is
07:52	cemerick	ro_st: probably
07:52	ro_st	buy you a beer.
07:52	cemerick	heh, thanks
07:53	cemerick	hopefully Alex looks favorably on my talk submission.
07:54	clgv	cemerick: as clojure book writer you have no "auto accept status" on clojure conferencs? ;)
07:54	cemerick	heh, no
07:55	cemerick	At this rate, all confs would be full-up with speakers without any submissions.
07:55	cemerick	It'd get a little boring. :-P
08:05	clgv	:D
08:15	squidz	where else can I find clojure videos besides the new clojure youtube channel?
08:19	ro_st	search vimeo for clojure
08:19	ro_st	search infoq.com for clojure
08:20	squidz	ro_st: okay so vimeo and infoq have a lot of clojure videos, thanks, anything else
08:20	ro_st	http://skillsmatter.com/
08:21	squidz	cool thanks
08:21	ro_st	plenty on those 3 for you
08:21	ro_st	90% of em, i'd say
08:21	squidz	yeah i think so, thanks
08:22	ro_st	sure thing
08:24	squidz	i'm almost done with Joy of Clojure, and am confident I can hack some clojure, but am kind of stuck on what I should actually program. I guess the opportunity will present itself at one point or another, but I can't wait
08:26	ro_st	what's your background?
08:26	squidz	I am working as a java dev developing a facebook app for career networking
08:27	squidz	backend stuff mostly
08:27	ro_st	oh then you must be loving clojure
08:27	ro_st	java devs love it once they invest enough time checking it out
08:27	squidz	yes so far i can't get enough of it. I took a detour and learned a little haskell but got frustrated with it so decided to check out clojure. I havent looked back
08:29	squidz	here is our app https://apps.facebook.com/emplido/
08:33	squidz	so far I was only able to really do one thing with clojure, and that is, sneak in a small clojure script into our project for maven to execute
08:43	ro_st	this looks ideal for clojure
08:43	squidz	really? why?
08:43	ro_st	it's well suited to the sort of data munging you're likely doing
08:44	squidz	ah I see. I guess the problem is as usual, how to convince others who know nothing of clojure
08:45	ro_st	users create data > data is collated to create information > information is queried for ratings and fuzzy matching > outbound messaging & report displays
08:45	ro_st	watch the talk from Amit at Zolodeck
08:45	ro_st	it's on infoq
08:45	ro_st	he uses clj + storm + datomic to great effect
08:47	squidz	ill check it out. What is storm btw?
08:47	ro_st	storm-project.net
08:47	ro_st	distributed real-time computation
08:47	ro_st	shit-hot.
08:48	squidz	looks cool, but isn't datomic proprietary?
08:48	ro_st	it is
08:48	ro_st	expensive, too
08:48	squidz	hm that makes it hard, but Ill check out the video anyways. Looks like it's interesting to learn
08:48	ro_st	still, its net effect is a radically lower cost of ownership for us
08:49	ro_st	far less developer time involved, and far less code
08:49	squidz	what did you use before?
08:50	ro_st	v1 stack is php+mysql
08:50	ro_st	v2 clj+datomic
08:50	squidz	oh okay, yes that must be much better
08:51	ro_st	way, WAY better :-)
08:51	squidz	ive never done anything with php, and I don't really want to
08:51	ro_st	don't
08:52	ro_st	it's horrible
08:52	squidz	I guess i'll just keep checkout out clojure videos until I get an idea of a miniproject. Small steps
08:52	AdmiralBumbleBee	php isn't that bad
08:52	ro_st	yeah. do you have clojurebook.com?
08:52	AdmiralBumbleBee	people just do awful things with it
08:53	squidz	ro_st: ive seen clojurebook.com, but I havent read any of it
08:53	ro_st	great book. highly recommended.
08:54	squidz	okay, i'm almost done with Joy of clojure. Maybe checking that out afterwards will be good
08:55	squidz	maybe starting from the practicums section, since the first part of the book may be a little repetitive
08:56	ro_st	yeah
08:56	ro_st	i'm going over the book again this holida
08:56	ro_st	haven't used it since the solid two weeks i spent reading it before i started coding
08:57	squidz	have you made any projects on github yet?
08:57	ro_st	yup http://github.com/cognician
08:57	ro_st	https://github.com/Cognician/fusebox is the most recent
08:59	squidz	cool
08:59	cemerick	ro_st: fusebox looks promising :-)
08:59	ro_st	thanks. i had a look around for a flags system in clj but found zip.
08:59	ro_st	i've got a nice little addon for datomic as well
09:00	ro_st	which defines an entity for the whole system, and then allows you to enable for particular entities or enums - for example, on user roles
09:00	ro_st	still need to tidy it up and release it
09:01	ro_st	there's almost no code to it
09:02	ro_st	we had 4 different projects doing the master/develop git branch thing and it was getting seriously tedious to manage versioning and dependencies between them
09:02	ro_st	so now we're just on master, and new stuff is wrapped inside fuse checks
09:03	WokenFury	cemerick: which is the best release of tools.nrepl for embedding? 0.2.0-RC1 or 0.2.0-beta9?
09:03	WokenFury	getting this error in RC1: java.lang.IllegalArgumentException: Multiple methods in multimethod 'print-method' match dispatch value: class clojure.tools.nrepl.server.Server -> interface clojure.lang.IDeref and interface clojure.lang.IRecord, and neither is preferred
09:03	squidz	what is the need for fuses though? To have global flags?
09:03	ro_st	fuses happen to look precisely like datomic attribute names, which isn't a coindcidence :-)
09:03	ro_st	squidz: http://stackoverflow.com/questions/7707383/what-is-a-feature-flag
09:04	cemerick	WokenFury: later is better
09:04	cemerick	Interesting re: the print error. I suppose I've never tried println'ing a server reference.
09:04	squidz	ro_st: ah i see
09:05	WokenFury	cemerick: afaik neither am I, just using (nrepl/start-server :port config/NREPL-PORT)
09:05	ro_st	we have one set of flags powering our web server, our background worker, and our cljs app
09:05	WokenFury	possibly because it's the last call in my main fn
09:05	ro_st	squidz: http://blog.asana.com/2011/04/using-flags-to-ease-new-feature-development/ is a great intro
09:06	cemerick	WokenFury: yeah, it's implicit if that's the last thing evaluated in e.g. a loaded file
09:06	WokenFury	aha. must've been that. sorted now
09:07	WokenFury	with beta9 it was spitting this out in console: #<Agent$Closeable$b14108b6@1ae57e12: {:ss #<ServerSocket ServerSocket[addr=0.0.0.0/0.0.0.0,port=0,localport=9991]>, :transport #<transport$bencode clojure.tools.nrepl.transport$bencode@706cac3b>, :greeting nil, :handler #<middleware$wrap_conj_descriptor$fn__8737 clojure.tools.nrepl.middleware$wrap_conj_descriptor$fn__8737@2d6f4087>}>
09:07	cemerick	Yeah, that's to be expected.
09:07	cemerick	RC1 is certainly better
09:07	cemerick	WokenFury: would you mind filing a bug @ http://dev.clojure.org/jira/browse/NREPL for the printing issue?
09:07	WokenFury	will do
09:11	acron^	Please excuse my ignorance, as I come from C/C++ so my mental model of operations in Clojure is lagging behind somewhat
09:11	acron^	I want to take a map and perform an operation on the value of each element in that map, depending on what they key is
09:11	acron^	I also don't want to keep fetching the map from the db
09:12	acron^	where would I even begin this, in the context of FP?
09:12	acron^	a massive 'do' function?
09:12	acron^	case?
09:13	Ember-	acron^: list comprehension?
09:14	acron^	do you have an example Ember-?
09:14	Ember-	acron^: http://clojuredocs.org/clojure_core/clojure.core/for
09:15	Ember-	and yes, you can pass a map to for
09:15	Ember-	and no, for is NOT a for loop
09:15	Ember-	in case of a map you are effectively processing key/value pairs
09:16	acron^	hmmm
09:16	Ember-	but it might not be an optimal solution for you, clojure has some much power in map and list processing
09:16	Ember-	some = so
09:16	Ember-	(what I was thinking when typed _some_ ??)
09:17	llasram	acron^: By "perform operation" do you mean "perform action for side-effects" or "generate a new value, assembling the sequence of new values into a new collection"?
09:17	Ember-	list comprehension is maybe one of the hardest things to graps coming from an imperative world
09:17	acron^	llasram, the latter
09:18	tpope	I still haven't grasped it
09:18	Ember-	in that case list comprehension should be your tool
09:18	llasram	Ok, then like Ember- is saying, the `for` comprehension will do it. Although IMHO you get more milage from the composable sequence operations
09:18	acron^	lets imagine i have {:a 1 :b 2 :c 3} - I want :a to be doubled, :b to be halved and :c should be trippled
09:19	llasram	Oh!
09:19	llasram	Well that's differest
09:19	noidi	(for [[k v] my-map] (make-new-value k v))
09:19	Ember-	and yeah, I'd propably use some higher level functions
09:19	acron^	and i want the results in a new map using the original keys
09:19	llasram	So you want a different function applied to the value for each key?
09:19	Ember-	for comprehension is kinda like the ultimate low level swiss tool for stuff like that :)
09:19	noidi	where make-new-value is a multimethod with implementations for :a :b and :c
09:19	acron^	llasram: yes
09:20	acron^	so {:a 1 :b 2 :c 3} -> {:a 2 :b 1 :c 9}
09:21	Ember-	what noidi said
09:21	Ember-	for comprehension + multimethod
09:22	Ember-	but how many different keys do you have?
09:22	Ember-	<10 or more?
09:22	acron^	potentially, yes
09:22	acron^	10-20
09:22	Ember-	ok, are they of what type?
09:23	Ember-	and what kind of operations?
09:23	Ember-	are they totally different from each other?
09:23	acron^	it's string in, map out
09:23	acron^	for all
09:24	llasram	acron^: ##(let [update-by-key (fn [fs m] (merge-with #(%1 %2) fs m)), fns {:a #(* 2 %), :b #(* 0.5 %), :c #(* 3 %)}] (update-by-key fns {:a 1, :b 2, :c 3}))
09:24	lazybot	⇒ {:a 2, :b 1.0, :c 9}
09:24	Ember-	yes, but can your keys be eg. "a", 1, true and so on?
09:24	llasram	But this isn't honestly a typical solt of thing
09:24	llasram	sort even
09:24	llasram	What's your higher-level goal with this operation?
09:24	acron^	Ember-: keys are all str
09:25	acron^	llasram: i'm turning url paramters into monger queries
09:25	Ember-	hmm, sounds a bit hazardous maybe
09:25	Ember-	but I gotta run
09:26	Ember-	significant other called, time to meet her :)
09:26	acron^	so color=Blue,Red --> {$or [{:color "Blue"} {:color "Red"}]}
09:27	acron^	so color=Blue+Red --> {$and [{:color "Blue"} {:color "Red"}]}
09:27	acron^	etc
09:29	noidi	acron^, https://www.refheap.com/paste/7745
09:29	llasram	Interesting. I agree with noidi the multimethod approach is probably the way to go
09:29	llasram	You can provide a :default for giving things a default treatment
09:30	acron^	noidi, this is really helpful, thank you
09:31	squidz	noidi: me likey
09:34	noidi	,(seq {:a 1, :b 2, :c 3})
09:34	clojurebot	([:a 1] [:c 3] [:b 2])
09:34	noidi	so a sequential view of a map is a seq of map entries
09:34	noidi	,(class (first {:a 1, :b 2, :c 3}))
09:34	clojurebot	clojure.lang.MapEntry
09:35	noidi	you can add map entries to a map with conj
09:35	noidi	,(conj {:a 1} [b 2])
09:35	clojurebot	#<CompilerException java.lang.RuntimeException: Unable to resolve symbol: b in this context, compiling:(NO_SOURCE_PATH:0)>
09:35	noidi	,(conj {:a 1} [:b 2])
09:35	clojurebot	{:b 2, :a 1}
09:36	noidi	and "into" just conj's everything in its second argument (a sequence) into its first argument
09:36	noidi	so then there are only multimethods, which are too complex to be explained on IRC :)
09:37	acron^	:E
09:39	squidz	acron^: have you done anything with multimethods yet?
09:39	acron^	Nothing at all
09:39	acron^	reading up now
09:40	squidz	nows a better time than ever I guess
09:42	acron^	noidi does "map" internall cally seq on the passed in map then?
09:43	noidi	acron^, yes
09:43	acron^	gotcha
09:45	tbaldridge	almost every sequence function calls seq on its argument. Vectors are not seqs, so map wouldn't work with vectors otherwise.
09:47	mmitchell	weavejester: Quick question about Compojure... is it possible to dispatch on a query param and its value?
09:49	mmitchell	weavejester: i currently have 1 route to a controller. The logic in the controller is very complicated due to some wacky business logic around a single query param. I'd love to be able to completely separate the logic into 2 controllers using routes.
09:55	clgv	mmitchell: you just can dispatch the clojure function in the route depending on the query param
09:57	clgv	(GET "/myrout" request (dispatch request)) and (defn dispatch [{{:keys [myparam]} :query-params :as request] (if (= myparam "awesome) (awesome-fn request) (default-fn request))
09:59	mmitchell	clgv: very cool thanks!
10:08	yedi	is there like a protocol or spec or format or something that can be used for storing/sending boolean rules (combinations of logical statements)
10:09	weavejester	mmitchell: You could also write it a little more concisely - there's usually no need to pass the whole request out
10:09	weavejester	mmitchell: (GET "/foo" [p] (case p "foo" foo-route bar-route))
10:09	mmitchell	weavejester: you mean by passing only the :query-params?
10:09	mmitchell	ahh ok
10:10	weavejester	mmitchell: You can return another handler from a route
10:10	mmitchell	nice
10:23	ppppaul	anyone have suggestions for testing my ring?
10:24	ppppaul	i see there is a lib ring-mock... it good?
10:24	clojurebot	libraries is http://clojure.org/libraries
10:24	weavejester	ppppaul: There's some other systems built on top of ring-mock
10:25	weavejester	ppppaul: Let me see if I can find the name of the one I'm thinking of
10:25	ppppaul	i'm also interested in mocking out http calls (my server is talking to other servers, and right now i'm in a hairy place where i could really use some testing fast)
10:25	squidz	does anybody do a blog in clojure?
10:26	nathanic	squidz: this blog http://yogthos.net/ is written in clojure. src: https://github.com/yogthos/yuggoth
10:27	nathanic	it even supports .war deployment in an application server, which is nice for people with JEE day jobs like me
10:27	weavejester	ppppaul: Try looking at https://github.com/xeqi/kerodon
10:28	squidz	nathanic: thanks
10:28	weavejester	ppppaul: if you're testing a web app
10:28	weavejester	ppppaul: Or https://github.com/xeqi/peridot
10:28	weavejester	ppppaul: if you're testing a web service
10:29	ppppaul	i'm building an api
10:29	ppppaul	peridot looks sexy
10:30	ppppaul	kerodon looks like magic _
10:30	ppppaul	thanks weavejester
10:30	ppppaul	:D
10:33	ppppaul	i have another issue. i'm using ring via 'lein ring server-headless' and i would like to see my logs going to the console. i am having trouble doing this. it seems like my logs are going into some void
10:37	jweiss	how would one generate a keyword in the current namespace (without explicitly mentioning the namespace). i tried ns but that didn't work at runtime, ns value ended up being clojure.core.
10:38	jweiss	i mean, create it from a string
10:38	jweiss	,(keyword "::foo")
10:38	clojurebot	:::foo
10:38	jweiss	doesnt work
10:38	jweiss	,(keyword "foo" (-> ns .name name))
10:38	clojurebot	:foo/sandbox
10:41	dhm	jweiss: (intern ns (symbol (str (keyword "foo"))))
10:42	llasram	jweiss: Extracting from `ns` should in fact work... Are you sure you didn't have scoping issues?
10:43	llasram	&(keyword (-> ns ns-name name) "foo")
10:43	lazybot	⇒ :sandbox7657/foo
10:44	jweiss	llasram: eg, (defn mkkey [] (keyword (-> ns ns-name name))) did not work
10:44	jweiss	was returning clojure.core, not the namespace in which mkkey was defined.
10:44	jweiss	sorry left out the string part of the keyword but you get the idea
10:45	llasram	Oh! ns is dynamic, so you'll get the value of ns where mkkey is called
10:46	jweiss	llasram: yeah, so i was trying to figure out how to refer to the current ns, meaning at compile time
10:46	llasram	(let [kw-ns (-> ns ns-name name)] (defn mkkey [s] (keyword kw-ns s))
10:46	llasram	)
10:46	llasram	:-)
10:47	jweiss	just trying to avoid a gotcha where i refactor the name of the namespace and inadvertently introduce a mismatch if i forget to update the hardcoded name
10:48	llasram	Right. So like I have above, you capture the ns name string in the lexical scope of the function. When you re-compile the namespace, the forms get re-evaluated, so renames follow automatically
10:48	jweiss	llasram: ok thanks
10:49	ppppaul	i'm trying to do interactive ring dev via emacs. i have :dev-dependencies [[ring-serve "0.1.1"]] in my project however i'm not getting 'ring.util.serve on my classpath. any ideas?
10:49	ppppaul	reading docs form this url https://github.com/mmcgrana/ring/wiki/Interactive-Development
10:49	llasram	ppppaul: Are you using Leiningen 1?
10:49	ppppaul	2
10:50	llasram	Then :dev-dependencies -> :profiles {:dev {:dependencies [...]}}
10:50	ppppaul	interesting
10:50	ppppaul	i'll do that no
10:50	ppppaul	w
10:51	llasram	I think you probably want ring-server instead of ring-serve though. It's more recent, and doesn't depend on an ancient version of hiccup
10:51	llasram	(I don't do much webdev so I can't say definitively, but that's what I found when I was doing a tiny compojure app last week)
10:52	weavejester	ring-serve is out of date, but I haven't got around to adding support for emacs debugging into ring-server
10:52	llasram	Ah
10:52	weavejester	But the emacs debugging landscape for Clojure has changed a lot since then
10:52	weavejester	nrepl seems to be the future
10:53	llasram	Yeah -- what does "support for emacs debugging" mean exactly?
10:54	llasram	Oh, swank.core/break
11:02	ppppaul	llasram, the :profiles goes in my project.clj or it goes in .lein/profiles ?
11:04	llasram	ppppaul: project.clj
11:20	ppppaul	i started clojure with swank a few years ago. now i'm back into clojure and i see that i'm being told to use nrepl. can someone explain why?
11:21	drewr	ppppaul: no one in the clojure community supporting swank-clojure coupled with nrepl being actively developed
11:21	ppppaul	is there a big difference right now?
11:22	drewr	little things mainly; I still use swank-clojure
11:23	ppppaul	so i have the same commands in emacs when i use it?
11:24	nDuff	They're not exactly equivalent, no.
11:25	ppppaul	i have ring.util.serve on my classpath now, but when i try (use 'ring.util.serve) i get an error
11:25	ppppaul	classnotfoundexception :(
11:29	drewr	ppppaul: no, the bindings are slightly different and the functionality is somewhat divergent; eventually I expect the nrepl interaction to be vastly superior though
11:29	ppppaul	drewr, the nrepl looks interesting
11:30	ppppaul	right now my problem is getting ring working in emacs
11:30	ppppaul	:(
11:34	ppppaul	any help with this would be appreciated... the ring docs for this are painfully incorrect :(
11:34	weavejester	Which docs?
11:34	ppppaul	https://github.com/mmcgrana/ring/wiki/Interactive-Development
11:35	ppppaul	i do this -> user> (use 'ring.util.serve)
11:35	ppppaul	i get back classnotfoundexception
11:35	weavejester	ppppaul: That's an out-of-date repository you've linked to
11:35	weavejester	ppppaul: https://github.com/ring-clojure/ring/wiki/Interactive-Development
11:35	ppppaul	when i do lein classpath \| grep serve i find the class
11:35	ppppaul	uuuuuuuuuuugggggggggggg!
11:35	ppppaul	:(
11:36	weavejester	ppppaul: But the ring-serve instructions are still the same
11:36	ppppaul	lol
11:36	weavejester	What happens when you (use 'ring.util.serve)
11:36	weavejester	?
11:36	ppppaul	oh cool, you made it the lib
11:37	ppppaul	weavejester, pm
11:39	weavejester	ppppaul: It might be that ring-serve doesn't work with Jetty 7, which is used by the latest versions of Ring.
11:39	weavejester	ring-serve is an out-of-date library
11:40	weavejester	You might want to try using the Ring adapter instead
11:40	weavejester	Or just using lein-ring
11:41	ppppaul	ok, i didn't know about the alternatives
11:41	ppppaul	so long as swank/break works i'm happy
11:41	ppppaul	or if logging works
11:41	ppppaul	or anything to help me debug
11:42	weavejester	Logging will work with lein-ring or the standard adapter
11:42	ppppaul	logging via (print/pprint)/?
11:42	weavejester	I don't think anything works with swank/break. I'll need to find some time to look into the newer Clojure debugging tools
11:43	weavejester	swank seems to be being deprecated for Clojure in favor of nREPL
11:43	weavejester	ppppaul: Yep
11:45	ppppaul	hmm, it doesn't seem to be working for me
11:45	ppppaul	maybe i'm looking in the wrong place
11:46	weavejester	ppppaul: What's not working?
11:46	ppppaul	my print statements aren't making it to the console
11:46	weavejester	ppppaul: Which console? A normal terminal? Emacs swank? Emacs nREPL?
11:48	ppppaul	terminal
11:49	ppppaul	i couldn't get ring working in emacs
11:49	ppppaul	i'm using lein ring server-headless
11:49	ppppaul	however, my code is currently throwing stack traces
11:49	ppppaul	it would be awesome to be able to have breakpoints working on my server
11:50	weavejester	ppppaul: You might want to look into Ritz
11:51	ppppaul	looking
11:52	weavejester	But to be honest I haven't found breakpoints to be useful enough to bother with
11:52	ppppaul	:)
11:52	weavejester	At some point I'll probably try Ritz
11:52	weavejester	But
11:52	weavejester	If your application consists of many small, pure functions, you can try out all parts in a REPL anyway
11:53	ppppaul	my app is a bit hairy now. just started putting in tests
11:56	tomoj	hmm https://www.refheap.com/paste/1f9f6acaced723eff69ce1e4f
11:56	tomoj	clojurescript
11:56	llasram	I really want to just sit and watch someone who uses breakpoints in their workflow, so I can understand when they may be useful
11:57	tomoj	seems you can't expand to a try/catch with a js/* for the class?
11:58	tomoj	weird thing is that analyzing the output of cljs.analyzer/macroexpand-1 works fine
12:01	tomoj	dnolen: https://www.refheap.com/paste/1f9f6acaced723eff69ce1e4f
12:02	tomoj	(ana/analyze env (ana/macroexpand-1 env '(defcatch foo [] bar))) works fine
12:02	tomoj	but (ana/analyze env '(defcatch foo [] bar)) throws that error
12:03	tomoj	bug?
12:03	clojurebot	Paste the contents of project.clj and ~/.lein/init.clj along with the output of ls ~/.lein/plugins and lein version.
12:03	S11001001	llasram: I used them occasionally in Common Lisp, but never bothered in Clojure.
12:03	S11001001	Of course, SLIME's CL support for breakpoints and inspecting stacktraces is pretty spectacularly good
12:05	S11001001	llasram: I'd say FP makes them much less useful.
12:06	llasram	S11001001: I definitely used them in C, so maybe that's it
12:09	tomoj	hmm, now (ana/analyze env (ana/macroexpand-1 env '(defcatch foo [] bar))) has the same error
12:10	tomoj	but if I take the output of (ana/macroexpand-1 env '(defcatch foo [] bar)) and copy it to (ana/analyze env HERE), it works
12:10	tomoj	O_o
12:11	dnolen	tomoj: hmm no idea.
12:11	dnolen	tomoj: it looks like it thinks the name that holds the exception is js/Error
12:11	ppppaul	can anyone help me get lein to run my tests on file changes?
12:11	ppppaul	i would like to have a setup similar to testacular (JS) but for my clojure project
12:12	tomoj	when the error happens, the try* form is (try* bar (catch js/Error e__8915__auto__ 42))
12:12	dnolen	tomoj: yes that try* is wrong
12:13	tomoj	yeah, but if I macroexpand-1, it's (clojure.core/defn foo [] (try bar (catch js/Error e__8915__auto__ 42))). hmmm
12:14	dnolen	tomoj: and you're calling ana/macroexpand-1 ?
12:14	tomoj	yeah
12:15	dnolen	tomoj: it looks like somehow you're macroexpanding Clojure try and not CLJS try
12:16	tomoj	try is a special form in clojure, isn't it?
12:16	dnolen	tomoj: it is, but it's a macro in CLJS that expands to a try*
12:19	antoineB	how do you not enforce immutability in cljs? (dealing with gui stuff)
12:19	dnolen	antoineB: ?
12:21	antoineB	i use set! to do for exemple cross reference with a dom element and my type
12:22	antoineB	maybe you have some project with good design?
12:22	tomoj	aha
12:22	dnolen	antoineB: still not sure what you are asking.
12:23	antoineB	it's a bit unclear for me
12:24	tomoj	https://github.com/clojure/clojurescript/blob/ebe40c8a/src/clj/cljs/core.clj#L866
12:24	tomoj	the try macro checks for list?
12:24	tomoj	should it be seq? ?
12:24	antoineB	something like that http://en.wikipedia.org/wiki/Functional_reactive_programming
12:27	antoineB	tomoj: you will write "(catch " not "[catch " so i think i will be a list
12:28	tomoj	&(list? `(catch foo))
12:28	lazybot	java.lang.SecurityException: You tripped the alarm! catch is bad!
12:28	tomoj	well, it's false
12:28	tomoj	but that's probably not the exact reason why
12:30	dnolen	tomoj: yeah that definitely looks like the problem.
12:31	tomoj	I expected to find seqs in there, but looks like it is just the conses
12:33	dnolen	tomoj: if you can confirm that fixes it for you, I can commit that change here.
12:33	tomoj	yep, that fixes it
12:33	tomoj	at least, on my stupid test case
12:34	tomoj	by the way, I'm porting clojure.test to cljs.test by keeping the ns/var->test mapping in atoms cljs-side. will an official port have to wait until vars are reified?
12:35	dnolen	tomoj: fixed in master
12:35	tomoj	confirmed that that fixed my real case, (is (thrown? js/Error ...))
12:37	dnolen	tomoj: I see no reason to wait for that. run-tests can be a macro that uses the atom to generate all the test calls.
12:37	dnolen	tomoj: a port of clojure.test would be most welcome.
12:38	tomoj	I had run-tests as a macro
12:38	tomoj	and all the rest
12:38	tomoj	with the mapping clj side. but that seemed to break when using the repl
12:39	tomoj	now run-tests is a function that accepts symbols and looks them up in the atoms (cljs side)
12:40	tomoj	which incidentally means you can deftest in the repl, although I didn't really want that much
12:41	tomoj	I had to change the namespace to cljs.test, which breaks tradition, because there are macros
12:42	tomoj	could stay with clojure.test and have a clojure.test-macros or something
12:45	technomancy	llasram: the one time I found stepping through breakpoints to be really useful was when I was investigating a bug in Clojure itself
12:47	dnolen	tomoj: I think cljs.test makes the most sense. I think the only reason there are namespaces like clojure.string is because of the old namespace clobbered by locals bug.
12:48	dnolen	tomoj: well I guess the argument for clojure.test is that feature expression might arrive one day, and that's one less hassle. that's a big argument for clojure.test actually.
12:49	technomancy	aha; it was a bug where you could read a form containing an invalid map in clojure 1.2 and it wouldn't throw an exception until you tried to access the map
12:49	technomancy	that was a fun one
12:50	konr_trab	Any idea why lein [ring] would not generate the uberwar, keeping the generation process going on indefinitely?
12:51	weavejester	konr_trab: When a compile process goes on indefinitely that's usually an indication there's some loop outside of a function
12:52	weavejester	konr_trab: e.g. if I tried to compile (def foo (doall (repeat 1)))
12:52	hugod	le
12:55	konr_trab	weavejester: thank you! That was exactly the case!
12:56	mmitchell	anyone here using sublime text w/nrepl?
12:56	tomoj	dnolen: but that would mean putting the clojurescript parts of clojure.test in clojure.git?
12:57	technomancy	mmitchell: pretty sure it can't do that yet
12:57	mmitchell	technomancy: yeah seems so
12:58	dnolen	tomoj: really? I think I'm missing something here.
12:59	llasram	technomancy: Oh, wow. Yeah, I could see any tool you could weild being useful against that one
12:59	tomoj	would feature expressions let you define macros in clojurescript/.../clojure/test.cljs ?
13:00	llasram	I really need to fix erc-spelling-mode not working until I turn it off then back on...
13:00	technomancy	llasram: typically it's most valuable in worst-case-scenarios though
13:00	tomoj	otherwise you have to put them in clojurescript/.../clojure/test.clj
13:00	tomoj	which is.. taken
13:00	tomoj	the alternative I see is to have clojure/.../clojure/test.clj use feature expressions
13:01	tomoj	not sure how that would even work
13:01	dnolen	tomoj: perhaps talking past each other. I don't see why you can't have clojure.test in the CLJS repo as a macro include.
13:02	tomoj	won't it conflict with the existing clojure.test?
13:03	dnolen	tomoj: ah right. ok, then I guess it has to be cljs.test which has the macros. and people will need to resolve the ns differences w/ feature expressions
13:05	tomoj	I wonder how we can do jvm-side fixtures
13:05	tomoj	or if that requires something more like one.test
13:07	tomoj	I also wonder how to compile with load-tests false..
13:09	dnolen	tomoj: I think getting the minimal useful set is a good start :)
13:21	dakrone	jkkramer: hey, when will clojuresphere be updated?
13:21	dakrone	jkkramer: last updated was in October
13:23	jkkramer	dakrone: I'll run the script today. I keep running into issues getting a proper dump of projects from github and end up disablng the auto-update
13:23	dakrone	jkkramer: ahh okay
13:24	jkkramer	the project needs some love and I haven't had time
13:46	augustl	how do you typically manage the creation of the base schema when using datomic? I'm thinking some kind of leiningen task would do.
13:52	crease	technomancy: slept on it. think you're right about multimethods
13:56	aaelony	I have a map that may contain a variable number of (differing) keys, :k1, :k2, ..., :kn. Depending on the keys present, various functions specific to that particular key need to be called to process the values of that key. There may be a lot of keys to look for. Considering a macro or a function that tests if the map contains a paricular key and then finds the function to be called. Or possibily the Maybe monad? Curious to hear
13:56	aaelony	people's thoughts on how best approach. https://www.refheap.com/paste/7755
13:58	S11001001	aaelony: write a variadic version of update-in that takes pairs of keypaths and functions and ignores absent keypaths
13:59	aaelony	S11001001: nice idea, thanks. I wonder if I should make a collection of the keywords (and names of the functions for each) and then pass that around.
14:00	hyPiRion	aaelony: when :k1 exists, should you assoc :k1-post or :k1 with the value computed?
14:00	aaelony	for debugging purposes, I'm keeping the both for now. Ultimately, only the post will matter.
14:00	hyPiRion	okay
14:01	hyPiRion	I'll see what I can come up with on the fly.
14:01	aaelony	wow, cool thanks.
14:03	hyPiRion	I've been working with merge functions for some time now, so it's kinda relevant to what I'm currently working on :p
14:03	aaelony	i think it may be a kind of (cringing…) design pattern.
14:04	aaelony	was looking at http://clojuredocs.org/clojure_contrib/clojure.contrib.monads/maybe-m but not sure if it is a fit, or performant, clean, etc...
14:06	hyPiRion	aaelony: https://www.refheap.com/paste/7756
14:06	aaelony	hyPiRion: that looks really nice
14:07	hyPiRion	I don't know how the functions you have look like though, but it should be pretty fast too
14:07	hyPiRion	If there are many, you should probably pop a transient!, assoc! and persistent! in there.
14:07	aaelony	some functions call java libraries, others call regexes, or compute stuff...
14:08	aaelony	very cool
14:10	aaelony	i like the use of reduce-kv
14:10	hyPiRion	aaelony: If there are many keys - like, I think more than 16 hits per map, then this should be faster: https://www.refheap.com/paste/7757
14:11	hyPiRion	Though it is slower if you have many functions
14:11	aaelony	there are something like between 10 and 50 keys
14:11	hyPiRion	few*
14:11	hyPiRion	Then the first one should do just fine
14:11	hyPiRion	I think...
14:11	aaelony	super to know both, will test
14:12	hyPiRion	Oh yeah, I haven't actually tested them, so I may have some typo in there. But the idea is idea, so it shouldn't be hard to fix it.
14:12	hyPiRion	/s/is idea/is there/
14:13	aaelony	understood
14:13	TimMc	hyPiRion is making up for the swearjure session yesterday. :-P
14:13	aaelony	for the more intense functions I might throw a future or something in there too
14:13	ivaraasen	TimMc: swearjure?
14:13	aaelony	big thanks
14:14	hyPiRion	aaelony: Ah, sounds smart. You should do two passes then - one where you future everything, and another where you deref everything.
14:14	hyPiRion	And good luck
14:14	aaelony	thanks - sounds good
14:14	hyPiRion	ivaraasen: swearjure is clojure without alphanumerics
14:15	hyPiRion	like, uh
14:15	hyPiRion	,(`[#'~(`[~@'`[~@[]]](+))](+)) ; But I don't want to go there today
14:15	clojurebot	(var clojure.core/apply)
14:15	TimMc	ivaraasen: https://gist.github.com/3916042
14:15	TimMc	ivaraasen: and https://gist.github.com/3036120
14:16	ivaraasen	beautiful
14:16	TimMc	This is what will make everyone want to use Clojure.
14:17	hyPiRion	"Screw brainfuck and javascript, Clojure is the new obscuria"
14:17	hyPiRion	TimMc: I think the first function we should get is juxt, by the way
14:18	TimMc	hyPiRion: JS is way better at this: http://utf-8.jp/public/jjencode.html
14:18	hyPiRion	It'd be easy to do conditionals then
14:18	TimMc	Ugh, except that site has encodign issues.
14:18	hyPiRion	TimMc: Currently yeah, but Clojure has potential
14:19	hyPiRion	and it runs on the jvm!
14:26	gtrak	TimMc: that's amazing
14:26	llasram	(inc TimMc)
14:26	lazybot	⇒ 27
14:29	squidz	if I have a function (myfun [p1 p2 p3] ...) how can I use a sequence (v1 v2 v3) for the parameters of myfun?
14:29	amalloy	&(doc apply)
14:29	lazybot	⇒ ------------------------- clojure.core/apply ([f args] [f x args] [f x y args] [f x y z args] [f a b c d & args]) Applies fn f to the argument list formed by prepending intervening arguments to args. nil
14:30	squidz	amalloy: ah thanks, so that's what apply is for
14:37	Anderkent	Can anyone recommend a graph library for clojure 1.4?
14:38	ohpauleez	Anderkent: What sort of graph problem/features are you looking for? Like a Graph DB, or just nice graph DSs?
14:39	brehaut	or graph rendering?
14:39	Anderkent	transitive clojure, grouping, topological sort
14:39	Anderkent	basic stuff
14:41	hiredman	~ping
14:41	clojurebot	PONG!
14:43	ohpauleez	Anderkent: I usually use a mix of the standard data structures, with functions to treat them like graphs, and zippers.
14:43	ohpauleez	But you might also look at this: https://github.com/jkk/loom
14:44	ohpauleez	There was an old contrib you could resurrect as well.
14:44	Anderkent	right, that's more or less what I found, though I hoped that there would be something already stable for something as basic as what I need. Thanks
14:44	ohpauleez	Or find a solid Java library, and just use it via interop (wrapping things into functions where it helps composition)
14:45	brainproxy	Anderkent: if you're at all interested in a graph database, OrientDB is pure java and pretty nice, and there is a clojure wrapper, clj-orient
14:45	ohpauleez	Anderkent: No problem - for most of my graph problems, I just end up using the basic clj ds's
14:46	ohpauleez	especially with clojure.walk and zippers
14:46	brainproxy	the apis for the db support various graph operations, as you might expect
14:46	brainproxy	eh, looks like clj-orient is lagging behind orientdb by a couple of point releases
14:58	bbloom	dnolen: saw you starred namin / TAPL-in-miniKanren-cKanren-core.logic, and saw in the readme: " It turns out that implementing general set operations in a pure logic system, when the domain is not already known, is an extremely difficult open problem."
14:58	dnolen	bbloom: that's mostly about the difficulty of unifying sets, which isn't interesting to me at all.
14:59	bbloom	dnolen: are there any decent articles, papers, things to look at regarding infinite domain constraint problems? i'm curious how that would even work
14:59	dnolen	bbloom: now that we have a constraint system, I think we can bring most Clojure operations as constraints
14:59	SegFaultAX	dnolen: What's the best online resource to learn about unification?
15:00	dnolen	we do (assocc x k v) the way we do (+fd x y z)
15:00	dnolen	we can do
15:00	bbloom	dnolen: yeah, that's clearly the next step for core.logic and i'm excited to see it unfold
15:01	amalloy	SegFaultAX: http://starwars.wikia.com/wiki/Unification_War
15:01	SegFaultAX	amalloy: Hah :)
15:01	bbloom	dnolen: I was just curious to understand "stubornly see for myself their limitations in expressing type systems."
15:02	dnolen	SegFaultAX: chapter 3 of William Byrd's thesis is a good introduction to the idea via miniKanren's implementation
15:02	tbaldridge	amalloy: illogical in this context that article is
15:02	SegFaultAX	dnolen: You wouldn't happen to have a link handy, would you?
15:02	bbloom	SegFaultAX: http://gradworks.umi.com/3380156.pdf
15:02	bbloom	i've been meaning to read it myself
15:02	SegFaultAX	bbloom: Danke.
15:03	dnolen	bbloom: well Kiselyov & Byrd have show you can do some interesting things - HM type inference / checking / inhabitation very succinctly
15:03	dnolen	bbloom: nominal logic programming is also very interesting - Prolog style programming to do inferences about names, binding and scope
15:04	dnolen	bbloom: and I have a hunch that with CLP(Set) you could probably get through most of Types & Programming Languages
15:04	bbloom	dnolen: k, got Cheney & Urban's paper about that open
15:05	mehwork	is there a more clojuric/idiomatic way of doing System/exit
15:06	ivaraasen	dnolen: know of any any semi-technical (applied) primers on logic programming?
15:06	amalloy	usually the answer is to type out (System/exit 0) then hit backspace 19 times and do something more flexible
15:06	bbloom	mehwork: if you're using agents, remember to call (shutdown-agents) other than that exit 0 is fine
15:06	SegFaultAX	,(count "(System/exit 0)")
15:06	clojurebot	15
15:07	amalloy	oh, i made a mistake with wc
15:08	dnolen	ivaraasen: I always recommend starting w/ a good Prolog text, Sterling & Shapiro Art of Prolog or Bratko's Prolog for AI book
15:08	mehwork	15?
15:09	ivaraasen	dnolen: will look into it. thanks
15:09	mehwork	the number of chars of the string, ah
15:09	SegFaultAX	amalloy: Even 25 would have been fine :) ##(count (str `(System/exit 0)))
15:09	lazybot	⇒ 25
15:09	mehwork	what's ##
15:10	dnolen	ivaraasen: there's an early draft PDF of Concepts Techniques & Models of Computer Programming on CiteSeer, also a good read especially to understand the constraint aspect of core.logic
15:10	bbloom	mehwork: inline evaluation for lazybot ##(inc 5)
15:10	lazybot	⇒ 6
15:10	mehwork	is lazybot on github
15:11	amalloy	$whatis source
15:11	lazybot	source is http://github.com/flatland/lazybot
15:12	ivaraasen	dnolen: I believe that book is part of my coursework next year actually
15:13	dnolen	ivaraasen: that's cool! :)
15:13	ivaraasen	we have a course on programming paradigms using Mozart/Oz
15:14	ivaraasen	looking forward to it
15:14	hyPiRion	ivaraasen: I have the book if you need. I'm not in Trondheim now though
15:15	hyPiRion	And I suppose it's going to be next year, not the upcoming semester.
15:15	mehwork	whenever i read through others' clojure code i never see any if statements. Are they frowned upon beause there's usually a better alternative in clojure?
15:15	ivaraasen	correct
15:17	mehwork	basically i want to know what the best way to write: run this function if this var is equal to 5
15:17	hyPiRion	mehwork: no else?
15:18	mehwork	nope, nothing approaching a switch/case
15:18	SegFaultAX	mehwork: With a lack of an else branch, when is usually good.
15:18	mehwork	nor ternary
15:18	hyPiRion	(when (= 5 var) (funcall))
15:18	mehwork	ah thanks
15:19	mehwork	can you when / else or do you have to if/else
15:19	hyPiRion	mehwork: if
15:19	dnolen	bbloom: Nada Amin is at EPFL doing PL things like this http://www.cs.uwm.edu/~boyland/fool2012/papers/fool2012_submission_3.pdf
15:20	hyPiRion	,(when (= 1 1) (prn 1) (prn 2)) ; <-- mehwork
15:20	clojurebot	1
15:20	clojurebot	2
15:20	SegFaultAX	mehwork: And before you ask, if you have more than 1 branch/alternate, cond is usually good.
15:20	mehwork	,(if (= 1 1) (prn "hmm"))
15:20	clojurebot	"hmm"
15:20	hyPiRion	^ or case, if that's working
15:20	mehwork	shouldn't i just do that then since it reads better
15:21	mehwork	even if there's no else
15:21	bbloom	dnolen: interesting. i'd be excited to see logic programming subsume typed programming
15:21	hyPiRion	mehwork: Noone's stopping you :p
15:21	mehwork	'when' feels like it's something that'd be related to a loop or event of some kind
15:21	hyPiRion	If you feel it reads better, then do it that way.
15:21	mehwork	hyPiRion: i know :p just wondering what the convention is
15:21	dnolen	bbloom: she's porting Byrd's nominal logic stuff to core.logic which pretty exciting. It's forcing a generalization of the constraint stuff I have in place and in pretty interesting. I don't know of many constraint solvers that focus on reasoning about programs - so perhaps we going down a interesting path not tread by other constraint solvers.
15:22	SegFaultAX	mehwork: It's arguably more idiomatic to use `when` if there is no alternate.
15:22	hyPiRion	mehwork: There's no "when"-over-"if" rule, if that's what you're asking
15:22	dnolen	bbloom: heh, I don't see that happening. But certainly I think having logic programming available makes it easier to understand model what type systems can do.
15:22	hyPiRion	But it explicitly tells the reader that there's no else-part here
15:22	mehwork	ok
15:22	amalloy	hyPiRion: well, more like there's two conflicting when/if rules
15:23	ivaraasen	I mostly use "when" when the else clause would just return nil
15:23	bbloom	dnolen: i could see type systems being expressed in terms of extensible logic programs
15:23	bbloom	dnolen: i'm nto saying types go away, i'm saying that logic programming becomes the underlying mechanism by which types are more useful and less restrictive
15:23	hyPiRion	amalloy: conflicting?
15:23	hyPiRion	wat
15:24	hyPiRion	teach me, amalloy.
15:24	SegFaultAX	mehwork: One word of caution (and this is mostly personal choice), don't use else with if-not. Negating a negation can be quite confusing.
15:24	amalloy	hyPiRion: most people's style is to use 'when wherever possible, such that 'if implies there's two clauses; if you see an 'if with only one clause from such a programmer, it indicates a mistake
15:24	dnolen	bbloom: perhaps, tho typed logic programming is a thing - see Mercury.
15:25	amalloy	a non-trivial minority use 'when only if there are side effects; thus the presence of a 'when helps notice the trickier spots with mutation
15:26	SegFaultAX	amalloy: That usage of when hadn't occured to me, but thinking back I've definitely seen code that used a convention similar to what you just described.
15:26	bbloom	dnolen: will look at it
15:27	amalloy	SegFaultAX: technomancy is the most vocal about preferring that style, if you want to ask him about it
15:27	hyPiRion	Well, I'm in some middle spot then. I excplicitly expand (if test res) to (if test res nil) and use when pretty seldom. Probably when I need mutation.
15:27	ivaraasen	dnolen: seen this paper? not sure if all of it would translate to core.logic, but it's a pretty neat example of applied Prolog. http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.52.5750
15:27	bbloom	dnolen: my interest in all this stuff is purely intellectual entertainment at the point in time
15:28	brehaut	amalloy: now i want 'when!
15:28	clojurebot	amalloy: therfor I return [previous] if rest is empty
15:28	SegFaultAX	brehaut: What would the bang signify?
15:28	amalloy	really, clojurebot? really? what do i have to unlearn you to avoid these echos from my past?
15:29	amalloy	clojurebot: forget amalloy: therfor I return [previous] if rest \|is\| empty
15:29	clojurebot	I forgot that amalloy: therfor I return [previous] if rest is empty
15:29	brehaut	SegFaultAX: that its for doing side effecting operations that are unsafe within a transaction, just like it always does
15:29	brehaut	it would probably be worth making 'when! wrap 'when in 'io!
15:30	dnolen	ivaraasen: WOW that looks awesome. Yes using core.logic to generate Clojure programs is definitely a goal, not sure how easy it would be translate that. Someone should try :)
15:30	dnolen	bbloom: it is very intellectually entertaining :) tho it's been exciting to see people put this logic stuff into production.
15:31	SegFaultAX	dnolen: I watched a talk about miniKanren where they procedurally generated all the scheme programs that evaluated to 6. That was pretty neat.
15:31	dnolen	SegFaultAX: yeah that's a great talk, the relational interpreters are neat.
15:31	ivaraasen	SegFaultAX: that sounds awesome. any recordings of the talk?
15:31	SegFaultAX	ivaraasen: Yes, let me find it.
15:32	dnolen	SegFaultAX: it's pretty easy to port that stuff to core.logic now Nada Amin has already done the quine generator.
15:32	brehaut	(defmacro when! [test & body] (list 'io! (list* when test body)))
15:32	SegFaultAX	ivaraasen: http://blip.tv/clojure/dan-friedman-and-william-byrd-minikanren-5936333
15:33	ivaraasen	SegFaultAX: thanks
15:34	SegFaultAX	brehaut: I didn't even know io! was a thing til just now. Thanks for that.
15:34	SegFaultAX	ivaraasen: No problem! :)
15:35	SegFaultAX	dnolen: I hadn't heard about that! Source?
15:35	TimMc	brehaut: Alternatively, (defmacro when! [test & body] `(when ~test (io! nil ~@body)))
15:35	lucky	hi
15:36	dnolen	SegFaultAX: http://github.com/namin/TAPL-in-miniKanren-cKanren-core.logic/blob/master/clojure-tapl/tapl/src/tapl/quines.clj
15:36	brehaut	SegFaultAX: incidentally, compare 'swap! (for atoms) to 'send (for actors) for the semantic differences of bang names: send is safe in transactions because its held till the transaction succeeds, whereas atoms always swap
15:36	TimMc	(move the io! inside)
15:36	brehaut	TimMc: i wondered about that too. im not sure about more precise warning vs consistently warning
15:36	TimMc	Also note the nil.
15:37	SegFaultAX	brehaut: Could you clairfy held till the transaction succeeds? Does that mean that it won't actually send until it's comitted by the STM?
15:37	TimMc	io! takes an optional first string arg...
15:37	brehaut	SegFaultAX: yes
15:37	brehaut	TimMc: oh really. huh
15:37	amalloy	TimMc: `(io (do ~@body)) seems more legible to me
15:39	SegFaultAX	amalloy: But that also prevents you from using a custom error message, no?
15:39	amalloy	so does his putting the nil in
15:39	SegFaultAX	Oh, true.
15:40	SegFaultAX	amalloy: Is there any cost to having a redundant do block?
15:40	amalloy	no
15:41	ivan	IDEA costs less than Sublime Text today
15:41	amalloy	and it wouldn't matter if there were, because i'll bet you a dollar (io! nil a b c) expands to (io! nil (do a b c)) anyway
15:41	SegFaultAX	dnolen: I don't think I'm quite far enough long in my core.logic understanding to absorb this yet. But it is quite short and elegant _looking_.
15:42	SegFaultAX	amalloy: That's why I asked if it incurred any cost since yours would expand to (io! nil (do (do a b c)))
15:42	dnolen	SegFaultAX: heh yeah, it'll make sense in good time
15:42	amalloy	well, i suppose that's true, it probably does. doesn't cost anything though; it's just as expensive as a {} in a C language
15:45	amalloy	i prefix all my macros with `(do (do (do (do ~@whatever)))) because i want the user's code to sound like it's humming
15:45	TimMc	:-D
15:46	brehaut	(do (do (do (do (repeat "yeah")))) ← nirvana song
15:46	zerokarmaleft	`(do (do (do (do (do (do (do (do ,@menamena))))))))
15:46	wink	lol
15:46	TimMc	damn you
15:46	dnolen	,'(foo ~x ~@y)
15:46	clojurebot	(foo (clojure.core/unquote x) (clojure.core/unquote-splicing y))
15:46	zerokarmaleft	oops, semi-schemified that
15:48	amalloy	zerokarmaleft: good thing you didn't totally schemify it. it doesn't work nearly as well with (begin (begin (begin ...)))
15:53	ivaraasen	SegFaultAX: my jaw dropped when I saw the generated code. so cool
15:54	bbloom	but did they ever get 120 to produce 5! ? :-)
15:55	SegFaultAX	ivaraasen: Tell me about it :)
15:59	hyPiRion	ivaraasen: yeah, it's pretty awesome.
15:59	dnolen	bbloom: I believe one of their students might have
15:59	bbloom	dnolen: bad ass.
16:00	bbloom	how long before rich can write down a slide deck and generate datomic? :-)
16:00	dnolen	bbloom: haha, it really only works for very simple languages - things bog down when you add too many features
16:00	bbloom	dnolen: of course, extreme combinatorial explosion
16:01	hyPiRion	dnolen: Are you telling us Clojure is not simple!?!?
16:01	hyPiRion	;)
16:01	dnolen	hyPiRion: :)
16:02	hyPiRion	Mr. Hickey has been lying to us the whole time, then.
16:02	dnolen	hyPiRion: heh "toy" or "model" language may be a better choice of words.
16:02	hyPiRion	dnolen: I know what you meant :p
16:02	bbloom	dnolen: in an entertainingly abstract philosophical sense, all programming is a search problem: you're looking in the infinite sequence of 0s and 1s for the program that does what you want. but that would take infinite time, so you need to employ some novel ideas to reducing that search space. i could imagine if you started encoding those ideas into your distribution strategy, you could solve more and more difficult logic problems
16:03	egghead	~search and constraint~
16:03	clojurebot	I forgot that ~paste is forget ~paste
16:03	TimMc	~botsmack
16:03	clojurebot	clojurebot evades successfully!
16:03	dnolen	bbloom: yeah, Will used some called condp where you could weight different language forms to eliminate unlikely programs
16:03	dnolen	at the Conj
16:03	bbloom	dnolen: yeah, something like that
16:04	bbloom	dnolen: would be interesting to also describe things like structural recursion, such that logic programs could try to solve "what's the base case? what's the reductive step?" to generate a recursive algorithm
16:04	bbloom	dnolen: expressing those novel ideas is challenging, for sure
16:05	TimMc	Novel ideas? They were probably codified in the 70s. :-P
16:05	bbloom	TimMc: i mean novel in the sense that they are worth remembering
16:09	ivaraasen	bbloom: my dream would be to have a parser do amortized analysis of an algorithm, rewrite it and generate a complementary data structure as well, i. e. a monoid for a finger tree. but that would probably be insanely difficult to do correctly
16:09	hyPiRion	Oh, the sufficiently smart compiler.
16:10	Raynes	weavejester: Are you going to put out a new version of codox?
16:11	mehwork	in a .clj file how do you go about a global/reusable var?
16:14	TimMc	mehwork: What do you mean by a global var?
16:15	ivaraasen	hyPiRion: do you know any NTNU faculty members that are enthusiastic about FP/logic?
16:15	mehwork	ah def alone does it within the namespae it seems. Does order matter? eg do you have to declare the var at the top before you can use it
16:15	TimMc	mehwork: Vars are declare-before-use, yes.
16:15	mehwork	TimMc: just something i can def at the top of the file and then reuse it in diff methods in defn's
16:16	TimMc	Sure, you could (def foo (atom 5)) and then later refer to foo.
16:16	mehwork	when you declare a ns at the top of a file clojure automatically applies it to just that entire file
16:18	mehwork	TimMc: why an atom instead of just (def foo 5)?
16:19	TimMc	mehwork: Just an example.
16:20	mehwork	think i need a good book on clojure. piecing it together over google and irc is kinda hard
16:21	S11001001	mehwork: check the clojure mailing lists; several book discussions have occurred
16:27	hyPiRion	ivaraasen: About logic, I'm not 100% sure. We had Tore Amble, but he died (:'( ) in January this year. Keith Downing is the only FP-guy I know of
16:28	hyPiRion	We're an endangered species up here, it seems.
16:28	ivaraasen	too bad
16:28	ivaraasen	PVV had a Clojure course though, right?
16:29	mehwork	do lispers believe in reincarnation or just recursion
16:30	ivaraasen	mehwork: they believe in a higher order at least.
16:30	mehwork	:)
16:30	mehwork	and parallel dimensions
16:31	mehwork	hopefully in at least one of them i'm not a noob right now
16:31	gtrak	do you guys ever feel like you want a null-object?
16:31	hyPiRion	ivaraasen: No, NVG had. I remember since I held it :p
16:32	gtrak	or some kind of way to specify and propagate contracts up a function stack? (checked exceptions.. shudder)
16:34	hyPiRion	gtrak: null-object?
16:34	gtrak	recursive docstrings...
16:34	hyPiRion	I'm pretty happy with nil.
16:34	hyPiRion	,(class nil)
16:34	clojurebot	nil
16:34	gtrak	hyPiRion: yea, it's not the nil that bothers me so much, it's the (if ..)
16:35	gtrak	we use it for convenience, if-let, but it's not always obvious what it signifies
16:36	gtrak	about 3 functions deep and it gets confusing... something needs to return a value
16:37	m0smith	rhc app start -a jbossas-7
16:37	m0smith	Application jbossas-7 does not exist
16:37	gtrak	nullity-overloading is obfuscating... I guess is what I mean.
16:40	hyPiRion	whut
16:41	hyPiRion	An element is truthy if it's not nil or false.
16:41	hyPiRion	(Except (Boolean. false))
16:42	gtrak	hyPiRion: I mean for example... say you have a database... (store/delete ... returns the entity or nil if it doesn't exist)... somewhere up the stack you can rely on this and use an if-let
16:42	gtrak	if there's too many steps in between the behavior and the branching... it's confusing
16:43	gtrak	I don't want to make all my functions polymorphic, because that's called java :-)...
16:44	mattmoss	something like -?>
16:44	gtrak	I wonder if there's a better way to think about it
16:44	hyPiRion	So it's the nil-punning you don't like?
16:44	gtrak	yea
16:45	gtrak	and the fact that it's so darn convenient :-)
16:45	hyPiRion	It's a design decision with pros and cons
16:46	gtrak	ah, yea, I'm not criticizing the feature, but rather looking for an alternative
16:47	hyPiRion	oh, right. check out the -?> as mattmoss mentioned
16:47	hyPiRion	it will short-circuit when nil is given as a result in the pipeline
16:48	gtrak	right :-).. I suppose that makes nil-punning more evident
16:48	hyPiRion	You could also return a keyword instead if that makes sense
16:48	hyPiRion	:not-found for your database-example
16:49	hyPiRion	Though that depends on whether the object can be :not-found, of course.
16:49	gtrak	right, or exceptions.. it's an interesting lib-design problem
16:49	weavejester	Raynes: Codox 0.6.4 released
16:49	Raynes	weavejester: You're a good guy.
16:49	Raynes	weavejester: Hope you feel better. :)
16:50	arrdem	,(inc weavejester)
16:50	clojurebot	#<CompilerException java.lang.RuntimeException: Unable to resolve symbol: weavejester in this context, compiling:(NO_SOURCE_PATH:0)>
16:50	arrdem	(inc weavejester)
16:50	lazybot	⇒ 6
16:50	arrdem	hur dur been a while
16:50	clojure-newb	cemerick: Hi, just trying to connect to your mock friend app with basic auth via curl…. but I'm getting stuck… getting a 404 atm, should be possible shouldn't it ?
16:50	weavejester	Raynes: I made the mistake of going to work a little too soon, so I felt absolutely terrible on Wednesday, but a little better now :)
16:53	janusmcintyre	Is there a more idiomatic way to do something like https://www.refheap.com/paste/7763 ?
16:54	janusmcintyre	I want to resolve the namespace + function from a string at runtime.
16:57	hyPiRion	,(-> "clojure.pprint/cl-format" read-string resolve meta :ns)
16:57	clojurebot	#<Namespace clojure.pprint>
16:57	hyPiRion	,(-> "clojure.core/read" read-string resolve meta :ns)
16:57	clojurebot	#<Namespace clojure.core>
16:58	hyPiRion	Don't know if that's better though
16:58	janusmcintyre	hyPiRion nice
16:58	janusmcintyre	,(-> "clojure.core/read" read-string resolve meta :ns)
16:58	clojurebot	#<Namespace clojure.core>
16:58	janusmcintyre	,(-> "foo.bar/read" read-string resolve meta :ns)
16:58	clojurebot	nil
16:59	hyPiRion	It'll resolve names available in the current namespace, so I'm not sure if it's good.
16:59	hyPiRion	,(-> "clojure.set/union" read-string resolve meta :ns)
16:59	clojurebot	nil
16:59	hyPiRion	,(require '[clojure.set :as set])
16:59	clojurebot	nil
16:59	hyPiRion	,(-> "clojure.set/union" read-string resolve meta :ns)
16:59	clojurebot	#<Namespace clojure.set>
17:02	janusmcintyre	hyPiRion it doesn't resolve namespaces that are not in the current namespace
17:03	hyPiRion	yeah, I figured as much
17:04	melipone	hiya!
17:04	melipone	how do I get all the values of an assoc list? I knew how to do that and I forgot
17:04	cemerick	clojure-newb: definitely; but, instead of looking at mock_app.clj (which I only wrote for testing, really), go check out https://friend-demo.herokuapp.com/
17:05	justin`	melipone: vals
17:05	cemerick	the sample apps there are very minimal, and should be easy to copy, paste, and run with
17:05	melipone	justin`: thnaks
17:05	cemerick	FWIW, I'm not spreading that URL around yet; I want to have ~10 example apps ready, and it needs to not look like trash.
17:06	cemerick	clojure-newb: ^^
17:06	janusmcintyre	hyPiRion I guess that code is not that bad then.
17:08	hyPiRion	janusmcintyre: I've not fiddled too much with ns and resolving vars in code, so I'm not the guy you should ask ;)
17:09	hyPiRion	It's not looking bad though, if that's what you're asking.
17:10	janusmcintyre	hyPiRion Thanks anyways. I searched around Github for something similar, but found nothing. If this is too hackish, I guess I'll find out soon enough
17:10	clojure-newb	cemerick: thanks… must have missed that
17:10	yedi	is there like a protocol or spec or format or something that can be used for storing/sending boolean rules (combinations of logical statements)
17:10	cemerick	didn't miss anything; I've not linked it anywhere yet
17:14	clojure-newb	cemerick: thanks for the pointer
17:24	melipone	quit
17:40	Thad_	"Victim's rights group" ,<-- want to see if this str contains the values "rights" OR "group"
17:42	callen	Thad_: http://clojure.org/other_functions#Other%20Useful%20Functions%20and%20Macros-Regex%20Support ?
17:42	callen	Thad_: just do an (or ...) of two presence-checking regexes?
17:43	pendlepa1ts	or use the regex #"(rights\|group)"
17:44	callen	pendlepants: I was trying to make it composable and able to treat as a list or something rather than a single block of regex.
17:44	callen	pendlepants: so that if it needs to be dynamically changed at runtime it doesn't lead to some kind of stygian horror.
17:49	hyPiRion	,(some #(.contains "victim's rights group" %) ["rights" "group"])
17:49	clojurebot	true
17:50	hyPiRion	,(some #(.contains "not in this one!" %) ["rights" "group"])
17:50	clojurebot	nil
17:52	Thad_	hyPiRion: perfect.. just use Java's .contain .. right ??
17:52	lazybot	Thad_: What are you, crazy? Of course not!
17:52	hyPiRion	Thad_: yes, it's not harder :)
17:53	hyPiRion	hey lazybot, you're talkactive today eh
17:53	hyPiRion	$findfn [1 2] 2
17:54	lazybot	[clojure.core/second clojure.core/last clojure.core/count clojure.core/peek clojure.core/fnext clojure.core/rand-nth]
17:56	borkdude	funny that rand-nth comes up
17:57	borkdude	$findfn [1 2] 2
17:57	lazybot	[clojure.core/second clojure.core/last clojure.core/count clojure.core/peek clojure.core/fnext clojure.core/rand-nth]
17:57	borkdude	$findfn [1 2] 2
17:57	lazybot	[clojure.core/second clojure.core/last clojure.core/count clojure.core/peek clojure.core/fnext clojure.core/rand-nth]
17:57	borkdude	(rand-nth [1 2])
17:57	borkdude	,(rand-nth [1 2])
17:57	clojurebot	2
17:57	callen	,(rand-nth [1 2])
17:57	clojurebot	1
17:58	borkdude	I would expect findfn sometimes not to list it, other times it would
17:58	aperiodic	$findfn [1 2 3 4 5 6 7 8 9] 9
17:58	lazybot	[clojure.core/last clojure.core/count clojure.core/peek]
17:58	borkdude	$findfn [1 2] 2
17:58	lazybot	[clojure.core/second clojure.core/last clojure.core/count clojure.core/peek clojure.core/fnext]
17:59	borkdude	ah there it is('nt) ;)
17:59	justin`	jackpot
18:01	Thad_	hyPiRion: This channel has a REPL ???
18:01	lazybot	Thad_: Yes, 100% for sure.
18:01	Thad_	cool
18:03	aperiodic	~botsnack
18:03	clojurebot	botsnack is scoobysnack
18:04	aperiodic	clojurebot: forget botsnack \|is\| scoobysnack
18:04	clojurebot	I forgot that botsnack is scoobysnack
18:04	aperiodic	~botsnack
18:04	clojurebot	Thanks, but I prefer chocolate
18:05	aperiodic	seems odd that inferences trump commands
18:06	technomancy	odd is what clojurebot does best
18:06	amalloy	aperiodic: botsnack isn't a command either
18:06	callen	$botsnack
18:06	lazybot	callen: Thanks! Om nom nom!!
18:08	borkdude	$inventory
18:08	amalloy	it's just a thing someone taught him a fixed reply to, a while ago
18:08	hyPiRion	$findfn [1 2] 2 2 2
18:08	callen	amalloy: don't demystify it.
18:08	lazybot	[clojure.core/if-not clojure.core/dosync clojure.core/sync clojure.core/condp clojure.core/nth clojure.core/get clojure.core/and clojure.core/locking clojure.core/io! clojure.core/when]
18:09	aperiodic	amalloy: oh, i see. that's less odd, then
18:09	hyPiRion	curious thing, this guy. I asked for fns, but got more or less just macros.
18:10	callen	$findfn [0 1] 1
18:11	lazybot	[clojure.core/second clojure.core/last clojure.core/peek clojure.core/fnext]
18:11	callen	k, so I'm not insane.
18:11	callen	well I am, but not about that.
18:11	amalloy	fnext. if there's any function we don't need in clojure.core, that's gotta be it
18:12	borkdude	,(doc fnext)
18:12	clojurebot	"([x]); Same as (first (next x))"
18:12	callen	LOL
18:13	callen	amalloy: at least it's not (caadddddr ...)
18:13	callen	amalloy: I actually figured it was people like you that enjoyed having fnext in core.
18:13	callen	amalloy: so you could golf harder on 4clojure.
18:13	amalloy	fnext instead of second? no thanks
18:14	callen	saves one character though!
18:14	amalloy	callen: have you looked at https://groups.google.com/group/clojure/msg/0dfda3a9360cbc4b ? it's a cute cadar-thing
18:14	hyPiRion	I could swear that was the docstring for second
18:15	hyPiRion	,(doc second)
18:15	clojurebot	"([x]); Same as (first (next x))"
18:15	hyPiRion	Oh, I see.
18:15	amalloy	somewhere i have a version with fewer bugs, but the broken version is just as entertaining
18:15	Thad_	Should I get an array / list when returned for this: (re-find #"rights\|group" "Victims' rights group") ???
18:15	lazybot	Thad_: How could that be wrong?
18:15	callen	amalloy: that tickles my black little ex-CL'er using soul.
18:16	callen	[cadr "1.0.2-SNAPSHOT"] <--- totally worth the bytes added to the jar.
18:17	amalloy	that implementation makes me so sad
18:18	callen	amalloy: did you use CL?
18:18	amalloy	for about a month
18:18	callen	because I have to say, that pretty well numbed me off to "bad libraries"
18:18	callen	oh, it was my first real language
18:18	callen	after GW-BASIC and C.
18:18	callen	I just kinda assumed after CL that most communities had god-awful libraries like that.
18:18	callen	then I learned C# and Python. That was a bit of a shock.
18:26	callen	are async/await in C# analogous to futures and promises in Clojure?
18:28	callen	I guess they aren't.
18:28	callen	solves a similar problem though.
18:29	brehaut	callen: promise / future concepts are analagous to the primatives await/async in .net are built on i believe
18:29	brehaut	callan: the API differs though
18:30	brehaut	callen: i imagine that with (protocol) monads you could produce something very similar though
18:32	callen	brehaut: well await is really just "dereference this promise and wait please.", right?
18:32	callen	or future, I guess.
18:32	callen	not promise. Sigh, sorry for bad clarity.
18:32	callen	brehaut: I decided to go with Hiccup btw, for templating.
18:32	callen	so raspberries
18:33	callen	main reason is that frontend guy wants to learn clojurescript anyway. Thinks hiccup would be a good way to force him to sink-or-swim with Clojure. Wouldn't normally be viable and definitely couldn't get away with it at the day job.
18:33	brehaut	callen: right. i think thats pretty much it. implementing and return bind on futures (or promises) lets you have nice syntax though
18:34	callen	brehaut: to be fair, with C#, if you don't make it very obvious, the community won't know how to use it.
18:34	callen	brehaut: I think that's partly why they made F#, to differentiate the community a bit and give them a good test bed for features for C#.
18:35	brehaut	callen: for reference, F# implements a comptuation expression over its primative values to do just that (computation expression ~= monads/do notation)
18:43	callen	brehaut: I've always like F#'s balance of strict/lazy semantics
18:43	callen	it's a pity the Haskell people used their language as a test-bed for lazy semantics.
18:50	mehwork	is there a way using lein to make it so m clojure app is self executable, e.g., my if my project's name is 'foo' to be able to just invoke it with ./foo rather than java foo.jar
18:51	AimHere	That introduces platform dependencies straight off, and loses you one of the reasons you're using the JVM in the first place, I'd have thought
18:54	technomancy	mehwork: there's a lein-bin plugin that sorta does that; not sure if it's been maintained
18:54	gtrak	mehwork: there's existing solutions to turn jars into executables, by bundling a jvm
18:55	mehwork	ok so it's bsaically acceptible to distrubute my app and have my users type 'java bla.jar'?
18:55	znDuff	mehwork, java -jar bla.jar is the usual thing, yes.
18:55	znDuff	mehwork, ...many operating systems will also let users directly run jars as executables, btw
18:55	znDuff	mehwork, ...including Linux, using the binfmt_misc kernel extension.
19:06	Bergle_1	ooh nice personal licenses 75% off jetbrains (intellij + others) http://www.jetbrains.com/specials/index.jsp
19:07	mehwork	end of the world sale lol
19:08	mehwork	(thanks technomancy gtrak znDuff)
19:08	hyPiRion	I've heard GNU is doing the same. They reduced their prices on emacs by 100% today
19:08	gtrak	100% less freedom?
19:08	gtrak	where do I sign up?
19:08	mehwork	i heard they were doing the opposite: for the end of the world they were making proprietary code
19:10	mehwork	free as in north korea
19:10	znDuff	Is the JetBrains site actually usable now?
19:10	znDuff	I tried buying an upgrade earlier today, and it was completely swamped.
19:11	ivaraasen	who needs JetBrains when you have ed?
19:11	ivan	you can buy but you don't get a license immediately
19:11	mehwork	idk, the idea of a sale on software feels like an abusive spouse who promises to only hit you once because it's valentines day. Still not excited
19:12	mehwork	i've been using vim without any clojure plugins. It's been painfulish
19:12	znDuff	...yes, that would.
19:13	mehwork	but i'm just learning the basis of the lang still, not working on a large project
19:13	znDuff	I can't do Clojure without paredit.
19:13	ivaraasen	IDEA does look very nice
19:14	n_b	IDEA is great for Java, could not make me dump vim for anything else
19:14	hyPiRion	So it's worth buying even though you're not programming Java frequently?
19:14	gtrak	the problem with me buying IDEA is it assumes I will have to use enough java some day to need it :-(
19:14	gtrak	it's depressing
19:14	Bergle_1	i can drive emacs a bit, im good with vim, but i am spending more time in emacs for paredit :) even though i dont like emacs much - maybe ill get use to it
19:15	Bergle_1	ive done a fair bit of javascript and webstorm or any of the ide's is great for javascript with refactoring.
19:19	n_b	Bergle_1: tried paredit.vim?
19:21	Bergle_1	no havnt.
19:22	Raynes	Bergle_1: I use Emacs with evil-mode.
19:22	Raynes	I enjoy it very much
19:23	Bergle_1	im plodding away with emacs-live, which i realise is lots more than just paredit, but some bits seem amazing
19:23	Raynes	Note that I was an Emacs user that tried Vim and then went back to Emacs and started using evil-mode. I don't use it because I have an attachment to vim keybindings, I use it because I like modal editing and thing it is a good fit in Emacs.
19:23	ivan	non-modal editing makes me corrupt something in my .org files every few hours
19:27	hyPiRion	Uh, is there any specific reason why numerator and denominator throws exception when given integers?
19:28	hyPiRion	It's okay that they call a method on a Ratio, but couldn't one make a protocol on top of it instead?
19:29	bbloom	hyPiRion: seems reasonable for those methods to be polymorphic. it might just be a case of evolution, an accident of history
19:29	seangrove	Is there a way to list keys for watchers given a ref?
19:30	bbloom	hyPiRion: did you search the mailing list? you may want to start a thread and or submit a patch
19:30	seangrove	ref/atom/whatever was passed into (add-watch ...)
19:30	hyPiRion	bbloom: No, I'll do that later on
19:30	amalloy	the code in Numbers.java can't use protocols
19:30	bbloom	seangrove: not that i know of
19:30	mehwork	there's no clojure convention for global variables?
19:31	bbloom	hyPiRion: that's the accident of history that amalloy is referring to: clojurescript is able to use protocols throughout, but jvm clojure can't for many things
19:31	mehwork	like all caps or an underscore prefix, etc?
19:31	bbloom	mehwork: nope, because there are no global variables, all top level vars are always namespaced
19:31	seangrove	earmuffs?
19:31	bbloom	seangrove: earmuffs are for dynamic variables
19:31	hyPiRion	bbloom amalloy alright, thanks for input
19:32	hyPiRion	,(doseq [p '[bbloom amalloy]] (prn (list 'inc p)))
19:32	clojurebot	(inc bbloom)
19:32	lazybot	⇒ 3
19:32	clojurebot	(inc amalloy)
19:32	lazybot	⇒ 36
19:33	amalloy	seangrove: ##(.getWatches (atom nil))
19:33	lazybot	⇒ {}
19:33	bbloom	ah amalloy didn't know that was there
19:33	amalloy	nor did i
19:34	mehwork	bbloom: i'm having a hard time knowing when to create a top-level var and a local one that i pass around
19:34	seangrove	Is ## a lazybot prefix?
19:34	hyPiRion	seangrove: yup
19:34	mehwork	seangrove: yeah
19:35	seangrove	Ok, phew
19:35	mehwork	as is ,
19:35	hyPiRion	mehwork: that's clojurebot, slight differences.
19:35	bbloom	mehwork: use a local whenever the value is only relevant in one spot
19:35	hyPiRion	But basically, yeah
19:35	seangrove	hrm, maybe .getWatches isn't in cljs?
19:35	bbloom	mehwork: use a top level whenever you want to use something throughout a namespace or between namespaces
19:35	bbloom	seangrove: in cljs you can just #(.-watches %) probably, but it's an implementation detail
19:36	seangrove	Ok, I'll take a look at it then, thanks
19:36	mehwork	bbloom: well my problem is diff than that. It's more to do with 'config' variables. I guess they should go at the top so they're easier to spot
19:36	bbloom	mehwork: ah yeah, so config variables are labeled ^:dynamic and often (but not always) annotated with "earmuffs"
19:37	bbloom	(def ^:dynamic config {:foo "default"})
19:37	mehwork	:o
19:37	bbloom	then you can use (binding [config {:foo "omg"}] ...) to temporarily override those values
19:37	mehwork	good to know thanks
19:37	bbloom	(doc binding)
19:37	clojurebot	"([bindings & body]); binding => var-symbol init-expr Creates new bindings for the (already-existing) vars, with the supplied initial values, executes the exprs in an implicit do, then re-establishes the bindings that existed before. The new bindings are made in parallel (unlike let); all init-exprs are evaluated before the vars are bound to their new values."
19:37	bbloom	but generally, top levels are good for config, functions, constants, etc
19:42	julienddd	Hi all, new to Clojure here. I've played with TryClojure and it's been great. What's the next thing to get started ?
19:42	julienddd	I'm mostly interested in web app development, but I need more practice with the syntax, etc before jumping in
19:42	SegFaultAX	julienddd: Perhaps a book? I suggest Joy of Clojure or Clojure Programming.
19:43	julienddd	anything free and online ?
19:43	mehwork	bbloom: well when would i want to use that config binding over a normal top level for a config value?
19:44	dhm	julienddd: i think clojure programming is a better book when you're starting out
19:44	hyPiRion	julienddd: Is the syntax, the concepts or both new to you?
19:44	brehaut	julienddd: if you know about functions and maps you are about ready to start mucking about with web stuff (ring)
19:44	bbloom	mehwork: so each var has what's known as a "root binding" which is the top level default value
19:44	julienddd	brehaut: Both, I'm a java/c guy
19:44	dhm	julienddd: yeah i recommend ring too. the examples are pretty small and self-contained
19:45	bbloom	mehwork: you can override it on a per-thread basis using binding, so you'd want that if you're going to have multiple clients of your API that need their own config. for example current-user or something like that
19:45	mehwork	ok
19:45	SegFaultAX	julienddd: Then I doubly suggest picking up a book. There are loads of new concepts that you'll be learning along the way and it would be good to have a guided tour.
19:46	hyPiRion	julienddd: For syntax and learning about functions, the webpage http://www.4clojure.com/ is great
19:46	bbloom	mehwork: you can also use a top level var like a mutable storage, but it's better to stick an atom into that var, and then mutate the atom... this way you can pass around a pointer to the atom and use it's compare & swap semantics
19:46	bbloom	vars are primarily only mutable for interactive development
19:46	cemerick	dhm: You can find a ToC and such for the O'Reilly book @ http://clojurebook.com
19:47	julienddd	Thanks everyone
19:47	SegFaultAX	julienddd: See cemerick's note.
19:47	hyPiRion	julienddd: But as other say, I'd recommend finding some book. If not, then http://java.ociweb.com/mark/clojure/article.html may or may not be of help
19:47	dhm	julienddd: yeah, see chapter 9 of cemerick's book. they have a bunch of java interop stuff there that might be fun to play with as a java person
19:47	dhm	cemerick: thanks
19:48	hyPiRion	I've seen it pop up time to time, but I haven't read it myself.
19:53	nightfly_	Practical Clojure was a good introduction for me after I had a few months experience with Common Lisp
19:54	seangrove	For those familiar with the google closure/cljs compiler output, I'm trying to track down the source of an implementation of a protocol in the chrome web console
19:55	seangrove	When a function/closure is lifted out of another function and turned into a delegate, where can I find the source for that function?
19:56	mrowe	sorry if this is a faq, but I've had no luck finding anything... I am getting ClassCastExceptions in clojure.lang.Reflector/invokeInstanceMethod when passing an int (e.g. 1) to a method that expects an Integer. I kind of hoped autoboxing would happen somewhere... do I have to do it manually?
19:56	seangrove	For example, any function that takes variable arguments will call __some_other_function at the end, but I'm not sure where __some_other_function is stored
19:57	amalloy	mrowe: boxing is not the problem; somethine else is but we can't tell from your description
19:57	mrowe	amalloy: ok. what more info do you need?
19:58	amalloy	well, some actual code. so far i have "i pass an integer to a method expecting an integer, but it doesn't work"
19:58	bbloom	seangrove: a good trick is to stick a # in front of a form and paste it into the rhino repl
19:58	bbloom	seangrove: that echos a pretty printed bit of java code
19:58	mrowe	amalloy: :) what is the preferred pastebin again?
19:58	hyPiRion	refheap
19:58	amalloy	anything is fine. refheap or gist
19:58	amalloy	znDuff will complain if you use pastebin
19:59	mrowe	https://www.refheap.com/paste/7768
20:00	mrowe	if I call it with an object that has a method setMaxCount(Integer count) and a params map { :max-count 1 } it fails, but { :max-count (Integer. 1) } works
20:00	amalloy	&(class 1)
20:00	lazybot	⇒ java.lang.Long
20:00	bbloom	Raynes will complain if you don't use refheap
20:00	mrowe	amalloy: ah
20:01	amalloy	it would work if the method took int rather than Integer, i think
20:01	mrowe	yeah
20:01	amalloy	the compiler does that downcast for you
20:01	mrowe	I think that's why it's worked for me before, but is tripping up now
20:01	mrowe	amalloy: cheers
20:09	seangrove	bbloom: Do you primarily use the rhino for your cljs repl env?
20:09	bbloom	seangrove: no, rhino is slow as hell. it's just convenient to run ./script/repljs to get a rhino repl
20:09	seangrove	Heh, thought so
20:12	technomancy	oh dear
20:12	technomancy	oh dear oh dear
20:12	technomancy	" hasattr(object, name) The arguments are an object and a string. The result is True if the string is the name of one of the object’s attributes, False if not. (This is implemented by calling getattr(object, name) and seeing whether it raises an exception or not.)"
20:12	seangrove	Damn, it seems like Shoreleave just stopped adding atom watchers...
20:12	technomancy	http://docs.python.org/2/library/functions.html#hasattr
20:12	brehaut	technomancy: im sorry to hear
20:12	callen	technomancy: what's wrong?
20:12	technomancy	callen: the last sentence?
20:13	callen	technomancy: Python?
20:13	technomancy	brehaut: makes clojure.core/contains? look positively well-adjusted
20:13	technomancy	callen: yeah
20:13	callen	technomancy: Python is my day job, so I can sympathize.
20:13	brehaut	technomancy: amazingly thats the most sane way to do it due ti the amount of 'magic happens' in object member lookups
20:13	callen	technomancy: Python is resents what I subject it to. Like 6 call-deep pipes of map/lambda/filter/reduce
20:13	callen	Python resents*
20:14	brehaut	technomancy: absolutely
20:14	callen	brehaut: oddly, if getattr(obj, "field", False): hasn't burnt me yet.
20:14	brehaut	callen: because it does all that magic for you.
20:14	callen	brehaut: most Pythonistas have enough taste not to abuse method_missing-esque DSLs like Ruby.
20:14	callen	I had an epiphany earlier today about Ruby and Python, actually.
20:15	callen	I realized Python had more function-oriented than class-oriented APIs than Ruby because Ruby doesn't have real namespacing.
20:15	callen	for some reason, this hadn't occurred to me before.
20:15	bbloom	Python seems to not worry about using exceptions for control flow
20:15	bbloom	it's part of the iterator protocol!
20:15	tpope	callen this recently dawned on me as well
20:15	bbloom	StopIterationExcepetion
20:15	brehaut	bbloom: death to iter
20:15	callen	bbloom: slingshot?
20:15	callen	tpope: it makes me happy to know this wasn't 100% obvious to everyone else.
20:16	callen	tpope: it seriously explains a lot though.
20:16	TimMc	bbloom: Maybe that's where prototype.js picked that up from.
20:16	tpope	I guess I've never seriously worked in a language with real namespaces before
20:16	callen	I hate class-based APIs.
20:16	callen	they always end up abusing some stupid stateful series of instantiated class-scope data to make it work.
20:16	bbloom	tpope: it's actually a big deal. i've been considering writing a blog post about the importance of actual namespaces
20:16	callen	bbloom: you should tell the C people about it, see what they think. :P
20:16	bbloom	i was considering an inflamatory hacker news bait title like: "The Node.js REPL is Broken"
20:16	bbloom	:-P
20:17	tpope	it's growing on me
20:17	callen	bbloom: I knew namespaces were a big deal, what I hadn't done was connect that to why Python users use more functions and Ruby people use more classes for exposing functionality.
20:17	bbloom	callen: yeah, it makes sense
20:17	callen	FQNS 4life.
20:17	seangrove	callen: Maybe
20:17	bbloom	callen: but even python suffers from a dynamism fail with namespaces
20:18	technomancy	callen: modules work great in Ruby for namespacing, but for some reason people prefer classes. I don't understand it at all.
20:18	callen	bbloom: erm, yeah, I'm guilty of a few things there.
20:18	seangrove	Mix-ins are kind of an interesting patch to get class-based things to look more like function-based things
20:18	bbloom	if you do `from foo import bar` and then re-def bar in foo, the previously imported name is not modified
20:18	callen	technomancy: you can't be trusted, you work for sie enemy.
20:18	bbloom	hence you see advice to `import foo.bar as bar`
20:18	callen	seangrove: the mix-ins vs. MI thing is a total canard.
20:18	seangrove	MI?
20:18	clojurebot	Nakatomi space is http://bldgblog.blogspot.com/2010/01/nakatomi-space.html
20:18	callen	seangrove: most Python multiple-inheritance gets used like Ruby mixins.
20:18	callen	there's very little difference.
20:18	tpope	yeah you can't do the selective import thing in ruby
20:19	callen	beyond the fact that Python gives you a little more power.
20:19	seangrove	Ah, I wasn't thinking of inheritance
20:19	bbloom	there's also a few decent traits based libraries for python
20:19	tpope	and there's no real replacement for refer
20:19	bbloom	which are slightly less shitty than mixins and multiple inheritence
20:19	callen	most Ruby code I've seen is require "blah" -> BOOM GOES THE REFERENCES
20:19	callen	(enjoy yer monkeypatchin')
20:19	bbloom	def Module.const_missing
20:19	bbloom	cringe
20:19	callen	I've never seen anything approaching sensible namespacing in Ruby.
20:20	tpope	imagine if you had to call Clojure::String.join each time
20:20	callen	bbloom: ick.
20:20	callen	another thing I noticed about classes/objects is that the composability and dimensionality is annoying
20:20	bbloom	UserPreferences -> const_missing -> user_preferences -> 'models/user_preferences.rb' -> require -> "oh that was cool" -> THE HORROR
20:20	callen	when you use very method-centric APIs you're basically forced to treat every single piece of code like a special snowflake
20:21	callen	instead of as being a generic sequence, IFn, etc.
20:21	bbloom	callen: or abuse send :-)
20:21	callen	I originally came to Clojure thinking it would fix everything I didn't like in CL. Realistically, it fixed what I didn't like about Python and Ruby.
20:21	bbloom	i've been guilty of writing modules with functions like frob_foo and frob_bar and then doing send("frob_#{thing.class.name.underscore}")
20:22	bbloom	it's sorta like multimethods :-P
20:22	callen	bbloom: that is unholy
20:22	seangrove	Dynamic!
20:22	bbloom	when you don't want to make a class for every enum in your database, it's quite reasonable heh
20:23	callen	bbloom: no I understand it, I've had to do some evil stuff in Python as well. I just resent being forced to do so.
20:23	bbloom	agreed
20:23	callen	bbloom: most Python meta-programming for me involves getattr, setattr.
20:23	callen	because it works against classes, modules, etc.
20:23	callen	so I can dynamically snag database models by stringly name, etc.
20:23	callen	getattr(models, "User
20:23	callen	", None)
20:23	callen	then I'll usually wrap up the None/Option/MaybeMonad stuff with a bind wrapper
20:23	callen	because I hate if-checking None.
20:23	brehaut	class Foo(call_to_class_generater("blah")):…
20:24	brehaut	my least favorite python dynamism trick ever
20:24	callen	brehaut: I try to avoid piercing the veil between runtime and compile-time in Python and Ruby.
20:24	callen	Python tends to bite you very hard if you do things that trip-over into compile-time.
20:24	callen	good example of this: default args
20:24	callen	obviously if you understand the semantics it's not a big deal, but it's really not a great idea.
20:25	callen	it'll confuse most people who haven't run into it before.
20:25	callen	it's like a "run once" macro.
20:31	lewis1711	I just try to avoid python fullstop
20:32	callen	lewis1711: you could do a lot worse.
20:32	callen	lewis1711: I bet I could name 20 programming languages that are worse.
20:33	technomancy	lewis1711: if you know of another SSH implementation that wouldn't require writing a C extension to OpenSSH lemme know =)
20:33	lewis1711	Cobol, Brainfuck, raw machine code... :)
20:33	lewis1711	technomancy: fair play. it does have nice libraries
20:34	technomancy	oh man there are pastebin links flying fast in this channel
20:35	callen	lewis1711: Erlang, most C libraries (yes, they're their own languages. Don't believe me? Look at ffmpeg. F!@#ing crazy frog.), Ruby, Common Lisp, C++, JavaScript, Go, Pascal, Java, Scala, Vala, C#, Haskell, Perl, PHP, Smalltalk, OCaml, APL
20:35	callen	oh and Ada.
20:35	technomancy	that's cheating
20:35	technomancy	also, you forgot VB
20:35	brehaut	and VBScript
20:36	brehaut	seperate class of evil
20:36	technomancy	also, OCaml is awesome
20:36	tpope	cobol if you really want to dig into the junk drawer
20:36	brehaut	also haskell
20:36	lewis1711	sorry were you listing all the languages people should use instead of python?
20:36	technomancy	and smalltalk, wtf dude?
20:36	lewis1711	PHP... I dunno that doesn't belong there. I'll give you that
20:36	callen	technomancy: there is a hypothetical parallel universe where OCaml and I are like this: http://www.youtube.com/watch?v=5utC5fvY-Zs
20:36	callen	technomancy: we are not in that universe.
20:37	technomancy	that is some wonderful CG
20:37	lewis1711	meh, suspend your disbelief :)
20:37	callen	we are in the universe where obscure academic francophones implemented OCaml and batteries included came out way later than it should've been. Also, I don't want to work for Jane Street.
20:38	technomancy	I guess it depends on whether you're talking about quality of the language or suitability for use at $DAYJOB
20:39	callen	technomancy: you've got a point there, but I try to incorporate my nighttime funsies into my day job
20:39	callen	technomancy: privilege of working in startups I guess.
20:39	brehaut	huh 'ring.util.request/body-string
20:39	brehaut	handy
20:41	seangrove	How can I use a local copy of a cljs library in lein-cljsbuild?
20:42	seangrove	Nevermind, I think I'm doing it with clojurescript already, can probably figure it out via :dependencies
20:42	gfredericks	lein checkouts!
20:42	callen	so am I the only one trying to find ring.util.request right now and failing at life?
20:43	weavejester	callen: It's in Ring 1.2.0-SNAPSHOT.
20:43	callen	oh god dammit
20:43	callen	weavejester: thank you.
20:43	callen	weavejester: do you know which github branch that is?
20:43	weavejester	I'm going to try and release Ring 1.2.0 sometime within the next month or so
20:44	weavejester	callen: master
20:44	callen	that's what I was in.
20:44	callen	wait wait, I'm a moron. h/o
20:44	callen	no no, false alarm.
20:44	callen	weavejester: ring.core?
20:44	callen	ring-core, that is.
20:44	weavejester	Right
20:45	callen	oh. wrong fork.
20:45	weavejester	https://github.com/ring-clojure/ring/blob/master/ring-core/src/ring/util/request.clj
20:45	callen	weavejester: mmcgrana has too much google juice man.
20:45	callen	sigh, sorry.
20:46	brehaut	shit. i bet im not helping with that
20:46	callen	brehaut: how so?
20:46	brehaut	by linking to the mmcgrana repos
20:46	callen	brehaut: s'not cool man, you confuse idiots like me :(
20:47	brehaut	huh. either i have completely forgotten about fixing the links, or someone has benevolently hacked my site
20:49	brehaut	turns out you can thank utvc
20:49	callen	I started listening to the Neverending Story OST because of the earlier reference. I need to do this everytime I code.
20:52	TimMc	callen: ┬─┬ノ( º _ ºノ)
20:53	brehaut	（╯°□°）╯︵ ┻━┻
20:56	callen	TimMc: I'm happy to announce that the war is over, we were never at war with Oceania.
20:56	callen	TimMc: we've always been at war with Eurasia all along.
21:30	TimMc	&(let [& [1 2 3] [& &] & [& &] & [& &] &] &)
21:30	lazybot	⇒ (1 2 3)
21:44	gfredericks	&(let [& [1 2 3]] &)
21:44	lazybot	⇒ [1 2 3]
21:46	Guest77224	#lighttable
21:57	TimMc	gfredericks: Well, that too.
21:58	TimMc	Did you see the Swearjure session yesterday evening?
22:03	gfredericks	wat?
22:03	clojurebot	For Jswat: start clojure with -Xdebug -Xrunjdwp:transport=dt_socket,server=y,suspend=n,address=8888
22:03	TimMc	gfredericks: Y'know, Curje.
22:03	gfredericks	currj?
22:03	TimMc	Clojre without alphanumerics.
22:04	TimMc	and the letter \u, apparently...
22:04	gfredericks	oooh
22:04	gfredericks	man I thought you were talking about a meetup or something
22:04	gfredericks	like some hip city whose clojure group is called swearjure
22:04	TimMc	Mmm, swearjure meetup...
22:04	gfredericks	no I was at a meetup yesterday evening actually
22:05	TimMc	gfredericks: There is a meetup with a name reminiscent of that, I think.
22:05	gfredericks	seajure?
22:05	clojurebot	seajure is the seattle clojure group: http://seajure.github.com
22:05	TimMc	Hah, no.
22:05	gfredericks	clojurebot: swearjure?
22:05	clojurebot	No entiendo
22:06	gfredericks	data science on the obama campaign
22:09	TimMc	Oh hey, when did the clojure.org docs link move to Github and get versioned? (... v1.3, v1.4, v1.5)
22:10	TimMc	,(doc reset-meta!) ;; what's this useful for?
22:10	clojurebot	"([iref metadata-map]); Atomically resets the metadata for a namespace/var/ref/agent/atom"
22:12	TimMc	,(when-first [x (range)] x)
22:12	clojurebot	0
22:15	TimMc	Hmm, how is drop any different from nthrest?
22:16	amalloy	TimMc: nthrest eagerly consumes the first n elements
22:17	TimMc	Mmm, right.
22:20	gfredericks	while drop waits till you call first on it? so it's like rest/next?
22:21	gfredericks	(defn drop [n coll] (lazy-seq (nth-rest n coll)))?
22:21	amalloy	yeah, i think that's correct
22:21	TimMc	It looks like drop doesn't allow GC on the head until the first call to first or seq.
22:28	brehaut	i think i must be completely ignorant on how to use java.io classes
22:28	brehaut	but how do you create a stream that can be read from and written to‽
22:28	amalloy	brehaut: start from the assumption that they're rubbish
22:28	brehaut	wonderful :/
22:28	amalloy	well, there's no such thing
22:28	brehaut	amalloy: that seems to be the case
22:28	amalloy	but it sounds like you might be happy with a Piped[Input\|Output]Stream pair
22:29	amalloy	depending on what you want that stream for. file writing, or what?
22:29	TimMc	Concurrent communication?
22:29	brehaut	amalloy: interstitial stream for catting multiple other stream things to
22:30	brehaut	eg, i have a bunch of files from clojure.java.io/resource and i want to make one stream that reads from them in order so i can feed it to std in of a conch program
22:31	brehaut	so far i can only assume ive grabbed the whole problem complete bass ackwards
22:31	amalloy	$google SequenceInputStream javadoc
22:31	lazybot	[SequenceInputStream (Java 2 Platform SE v1.4.2)] http://docs.oracle.com/javase/1.4.2/docs/api/java/io/SequenceInputStream.html
22:31	brehaut	phew
22:31	brehaut	thanks
22:31	amalloy	yes, copying to a "buffer" is pretty backwards
22:31	amalloy	you want to construct a stream which delegates. in this case, that is already written for you (though with a terrible API)
22:32	amalloy	you'll also want ##(doc sequence-enumeration), i think it's called
22:32	lazybot	⇒ nil
22:32	amalloy	&(doc seq-enumeration)?
22:32	lazybot	⇒ nil
22:32	brehaut	enumeration-seq
22:32	amalloy	no, that's the wrong direction
22:32	brehaut	oh
22:33	amalloy	brehaut: you could also just steal from me: https://github.com/flatland/io/blob/develop/src/flatland/io/core.clj#L30
22:33	brehaut	ah that will do
22:33	brehaut	thanks
22:33	amalloy	i wonder what the "opts" arg is for
22:34	brehaut	consistency with clojure.java.io/… ?
22:35	jkkramer	forward-compatibility with a version yet to be written
22:39	amalloy	ah yes, consistency with c.j.io/make-input-stream
22:39	amalloy	i think in theory you can write (c.j.io/input-stream [stream1 stream2])
22:40	amalloy	and it calls (make-input-stream [stream1 stream2] nil), so we register an implementation of that
22:41	brehaut	clojure.java.io is one of those libs that if you arent writing clojure regularly is quite confounding
22:42	brehaut	likewise
22:42	gfredericks	you just try things and it always works for the 4th thing you try
22:42	brehaut	it doesnt help that the java library is obtuse at best
22:42	gfredericks	then you forget the details of what made it work and repeat that 2 months later
22:42	amalloy	it's certainly not as transparent as (most of) clojure.core
22:43	gfredericks	,(doc monitor-enter)
22:43	clojurebot	Gabh mo leithscéal?
22:43	amalloy	thinking about it now i'm quite impressed by how readable clojure.core is, on the whole
22:43	brehaut	yeah both statements are true
22:44	gfredericks	I'm surprised that the define-the-first-four-arities pattern was never generalized
22:44	gfredericks	or automated somehow
22:44	brehaut	my impression of c.j.io is that it looks like it does a lot of clever for me, but i just cant work out what exactly
22:44	amalloy	gfredericks: it's pretty hard to do. i think at some point rich welcomed someone to solve it
22:44	brehaut	gfredericks: i think heroic efforts were undertaken around 1.2/1.3 but it stumped allcomers
22:45	gfredericks	amalloy: yeah I had the sense that might be the case
22:45	amalloy	i spent an hour or two on it and gave up
22:45	brehaut	i think fogus might have got pretty close but not close enough
22:45	gfredericks	it's a bit like loop unrolling isn't it
22:46	amalloy	loop unrolling seems more structured to me
22:57	TimMc	What's this about arities?
22:57	gfredericks	I guess I need to find an example
22:58	alandipert	gfredericks: you mean like [] [a] [a b] [a b c & more], the stuff that's all over math?
22:59	TimMc	Oh, all those monoids!
22:59	gfredericks	alandipert: I suppose...math?
22:59	gfredericks	https://github.com/clojure/clojure/blob/master/src/clj/clojure/core.clj#L600
22:59	gfredericks	^ there's an example
22:59	gfredericks	oh no not monoids
22:59	gfredericks	it's specifically when there's an unnecessary number of small arities defined before the vararg case
22:59	brehaut	amalloy: am i missing something? i should be able to just require flatland.io.core and it should registers a make-input-stream implementation for seqable things (such as vectors) right?
23:00	alandipert	gfredericks: oh, i was thinking math because i've hung out there… another neighborhood in core.clj, lol. but yeah it's all over
23:00	gfredericks	I mean probably they are monoids but they're defined that way for perf, not just for being monoids
23:00	amalloy	brehaut: that's the plan, yep
23:00	brehaut	i must be doing something stupid then
23:03	technomancy	flatland.io.core? =(
23:04	brehaut	whats the opposite of turtles?
23:04	technomancy	shredder
23:05	brehaut	cause its those, all the way up
23:05	brehaut	its shredders all the way up
23:05	amalloy	(dec shredder)
23:05	lazybot	⇒ -1
23:06	TimMc	[core.core/core "1.0.0-SNAPSHOT"] (:require [core.core.core.core.core :as core])
23:06	brehaut	http://www.rottentomatoes.com/m/core/
23:07	brehaut	audiences agree: core is not good
23:12	TimMc	brehaut: Oh lord, you prompted me to read the plot of that on Wikipedia... christ that's terrible.
23:13	brehaut	TimMc: its a better movie than its plot
23:13	brehaut	but only just
23:13	brehaut	and because its firm mocking itself
23:13	TimMc	Is it?
23:13	brehaut	only just
23:13	brehaut	the plot is rubbish, the movie is slightly entertaining rubbish
23:14	brehaut	i still wouldnt watch if given a choice
23:22	alandipert	it does have that sweet shuttle LA river landing scene, which is how i justify naming my namespaces core
23:24	brehaut	lol
23:41	frozenlock	woohoo! Cljs in cljs!
23:45	clj-newb-234	is (= (m-bind (m-result value) function)) (function value)) correct? or should i be (= (m-bind (m-result value) function) (m-result (functoin value))) ?
23:45	clj-newb-234	http://onclojure.com/2009/03/06/a-monad-tutorial-for-clojure-programmers-part-2/ (in referene to)
23:46	amalloy	clj-newb-234: the latter looks right to me
23:47	TimMc	frozenlock: Really?
23:49	amalloy	clj-newb-234: i think actually it's the former. let's take the specific case of the List monad: f is then a function from x->[x]. so (m-result 5) is [5], and (m-bind [5] f) calls (f 5), which is expected to return a list
23:49	frozenlock	TimMc: taken from HN https://github.com/kanaka/clojurescript
23:49	clj-newb-234	amalloy: looking at http://www.haskell.org/haskellwiki/Monad_Laws , it appears to be the former
23:51	TimMc	frozenlock: Hmm, what makes this not self-hosted?
23:53	frozenlock	?
23:54	clj-newb-234	can "(m-bind (m-result value) function)" be wrritten in domonad notation? I don't believe it's equiv to: "(domonad [x (m-result value)] (function value))"
23:56	TimMc	frozenlock: Well, it claims to not be a self-hosted compiler, but it also claims to be a (mostly working) CLJS compiler that can compile itself.
23:57	TimMc	Hmm, or is it that the standard Clojurescript compiler can compile it, but it has some parts it can't compile for itself yet?
23:58	frozenlock	That would make sense