#clojure – 2012-10-18

2012-10-18

00:04	ForSpareParts	Have any of you guys done real-timey stuff using Clojure? I'm trying to hack something together, and I'm a little hung up on creating behaviors that regularly act on a state.
00:04	ForSpareParts	Thought about attaching "update behaviors" to elements of the state, also considering an event queue-like...thing.
00:14	doomlord	you mean realtimey like games or something else
00:43	ForSpareParts	doomlord, Sorry I missed your message before, but yeah.
00:44	ForSpareParts	I'm interested in games specifically, but I'm guessing the problems I'm working on here have broader applications.
00:46	doomlord	i've only dabled with FP a bit. i've setup a gameloop in haskell.. my useage of haskell is very noob-ish but i managed it
00:47	doomlord	i had a record per entity, a list of entities, and each just had a function for re-generating itself "next frame" ... more like a C method (callback function) of doing polymorphic entities than a C++ method (classes etc)
00:47	doomlord	when I say "my haskell code is noobish" - i dont fully get monads yet, instead of the "state monad" i've just been manually piping a state around;
00:48	doomlord	(i made some helper function that made it easier)
00:49	ForSpareParts	Yeah, I wasn't really thinking in monads, myself.
00:49	ForSpareParts	Had a notion that I'd build a data structure for my state and map transforming functions over it, somehow.
00:49	ForSpareParts	I've found a few methods for doing this, but they seem... inelegant.
00:50	doomlord	my 'stone age' monad state replacement is a helper function like this:- pipeState3 f0 f1 f2 s .. this did the result0, s'= f1 s..result1 s''= s''' chaining :) but i need versions for N arguments pipeState2 pipeState4 etc
00:51	duck1123	does a send-off inside a dosync still only fire when the transaction completes?
00:51	doomlord	i was happy with the haskel representation , but haskel records suck - i find that a barrier
00:51	ForSpareParts	doomlord, Haskell's a neat language, but I'm not fond of its syntax -- personal preference.
00:51	ForSpareParts	Feels too math-y to me.
00:51	doomlord	i was happy with the core concept of data entity{ update::UpdateFunc, ....}
00:52	duck1123	also, what about logging inside a dosync? My send-off, nor my log statement seem to be firing
00:52	tomoj	sends are held until the transaction ends
00:52	tomoj	IO is not
00:53	doomlord	i like the type inference in haskell; but the record system puts me off.. also as you say, some of its syntax choices are also offputting;
00:54	duck1123	I wasn't expecting the io to be safe, it was only there because I can't figure out why the send isn't running
00:54	doomlord	its got infix.. but not . for element acess; that sucks. and i prefer F# \|> to haskell's $ ... its a bit like the threading macros in clojure i think
00:56	doomlord	is there a clojure interpreter that can be embedded in a c++ program :) if not perhaps it wouldn't be too hard to adapt one of the existing embeddable lisps to behave more like clojure
00:59	doomlord	sort of related to the gameloop discussion - are there any GUI frameworks that work in a functional way ..does that even make sense. will gui always be very impure by its nature. or do any of clojures features lend themselves to events etc
01:02	ForSpareParts	doomlord, I've seen some stuff about using functional-reactive programming for that -- somthing else I've been thinking about in relation to games.
01:02	ForSpareParts	Google "yampa" -- it's a Haskell library that implements it, though I can't say I exactly grok it yet.
01:03	duck1123	okay, I just took the agent out of the equation. It's working now.
01:05	doomlord	trying to find java opengl bindings.. they're not in the javasdk's by default are they? is it 'jogl' that i'm after
01:06	ForSpareParts	doomlord, LWJGL, perhaps?
01:18	abp	Hell-o
01:18	technomancy	salut
01:18	doomlord	has anyone here installed jogl?
01:19	doomlord	where is the best place to get it, how do you install it
01:22	abp	doomlord: google said that for example: http://stackoverflow.com/questions/1962718/maven-and-the-jogl-library
01:23	abp	doomlord: searching jogl maven, because you probably meant a dependency in leinigen when saying "how to install it".
01:26	aperiodic	doomlord: have you looked at penumbra? https://github.com/ztellman/penumbra
01:26	doomlord	i am a complete noob with java environment
01:28	doomlord	interesting so it turns begin/end , push/pop into single forms
01:29	doomlord	i think i'll probably stick with the interface as close to C as possible
01:30	brainproxy	hmm, I need to include the html entity ™ (and things like that) in my defelem, but it's getting espaced or something when hiccup generates the html from the template
02:45	doomlord	eek, where is sqrt
02:52	fredyr	(Math/sqrt 2)
02:52	fredyr	?
02:52	antares_	,(Math/sqrt 2)
02:53	clojurebot	1.4142135623730951
02:53	doomlord	ah. i was trying (use 'clojure.contrib.math) ... and gettnig errors
02:53	doomlord	yes that works here
02:53	fredyr	oh gotta try the bot
02:53	antares_	doomlord: clojure.contrib.* is obsolete, see http://dev.clojure.org/display/design/Where+Did+Clojure.Contrib+Go
02:54	doomlord	ok. the perils of distributed info
02:55	doomlord	is it possible to pull that into the main namespace.. i'm trying (use Math) & various combinations with clojure\; ' etc
02:55	fredyr	,(/ (+ 1 (Math/sqrt 5)) 2)
02:55	clojurebot	1.618033988749895
02:55	antares_	doomlord: Math is a java class (java.lang.Math)
02:56	antares_	you cannot refer to Java methods
02:56	doomlord	yikes,ok.
02:56	ChongLi	you can use import-static to turn them into macros
02:56	goracio	hello can anyone help with this https://gist.github.com/3910191 mongo query ...
02:56	doomlord	i was about to say, should i just make my own wrappers, but "import-static" will do that for me?
02:57	ChongLi	that's what it does
02:57	ChongLi	though it's not ideal
02:57	ChongLi	macros are not first class functions
02:57	clj_newb_234	mvn install:install-file ... <-- this installs *.jar fles. How do I instlal jnilibs ? So I can use them via lein.
02:58	doomlord	would it not do the same job as #defines in C :)
02:58	doomlord	oh.. you can't pass to map :(
02:58	ChongLi	yeah
02:58	doomlord	i think i'll be ok with that- it would have thrown me if i didn't know, but i'm much more likely to map with a lambda or something else i've written
02:59	ChongLi	to use something like sqrt with map
02:59	ChongLi	you need to wrap with a lambda
03:00	ChongLi	,(map #(Math/sqrt %) [4 9 16 25])
03:00	clojurebot	(2.0 3.0 4.0 5.0)
03:01	doomlord	,(map Math/sqrt [1 2 3 4 5])
03:01	clojurebot	#<CompilerException java.lang.RuntimeException: Unable to find static field: sqrt in class java.lang.Math, compiling:(NO_SOURCE_PATH:0)>
03:01	doomlord	,(map (fn[x](Math/sqrt x)) [1 2 3 4 5])
03:01	clojurebot	(1.0 1.4142135623730951 1.7320508075688772 2.0 2.23606797749979)
03:01	arrdem	Clojure doesn't have an "import *" now does it.
03:01	aperiodic	clj_newb_234: you just need to have some jars with the native extensions in the proper arrangement; lein will extract the native deps and put them on the classpath to be picked up by the loadLibrary statements
03:02	clj_newb_234	what is the "proper arrangement" ?
03:03	doomlord	clojure in ubuntu: i just discovered you can apt-get install clojure1.3 (up till now i just had 1.1 without installing something else manually).. is 1.3 still trash?
03:03	ChongLi	it's not trash, it's just old
03:04	doomlord	i was trying to get clojure working with emacs aswell... is there more to it than just (set-inferior-lisp "clojure")
03:04	ChongLi	yeah
03:04	ChongLi	nrepl
03:04	doomlord	slime sorry not emacs
03:04	aperiodic	clj_newb_234: this is the best documentation about it i've been able to find: https://github.com/swannodette/native-deps
03:05	ChongLi	you can use slime or nrepl
03:05	doomlord	having second thoughts about exposing myself to the JVM and letting it occupy any neurons in my skull , but clojure itself appears very interesting
03:06	ChongLi	what's wrong with the JVM?
03:06	clj_newb_234	aperiodic: nice thanks
03:07	doomlord	i've used c++ mostly. lisps weren't so shocking or 'far away' because they can be embeded
03:08	doomlord	i'd never touch java.
03:08	ChongLi	well if you do you're missing out
03:09	ChongLi	there are a lot of great java libraries you can use with clojure
03:09	doomlord	clojure is interesting, as is scala. java itself... its the combination of a certain style of OO and non-native that has a very negative reaction in my head
03:09	ChongLi	there are so many nice "small" languages out there that languish due to lack of library support
03:09	ChongLi	the fact that clojure can leverage java libraries is a huge advantage
03:10	doomlord	ok fair enough.
03:10	ChongLi	and clojure really takes the pain out of writing java anyway
03:11	aperiodic	also, there's a huge difference between "Java" and "the JVM"
03:11	ChongLi	thanks to macros and the lack of checked exceptions
03:12	doomlord	yeah, but they come up in word association. Like i say.. i've started with an interest in FP which made me look into this (i also want to try scala). i like lisp macros. i suspect scala will be easier for my c++ head geerally
03:13	ChongLi	I haven't looked at scala very much
03:13	ChongLi	it seems like a really weird language
03:13	ChongLi	due to being multi-paradigm
03:14	doomlord	heh. on the contrary i would say "it looks interesting due to being multi-paradigm" :) mult-paradigm seems the most useful to me
03:14	doomlord	shift style to whatever suits
03:14	ChongLi	I wouldn't be so sure
03:15	doomlord	haskell being very pure FP makes some basic tasks hard
03:15	ChongLi	there's always a price to be paid in terms of cognitive load
03:15	ChongLi	yeah, haskell can have a very high price
03:16	ChongLi	some of the type signatures get way out of hand
03:16	doomlord	scala compared to C++ , i think it will be a lightened load.
03:17	ChongLi	I don't know; scala's got quite a complex type system itself
03:17	doomlord	many things which require hacking with #defines should be easier when built-in
03:17	ChongLi	also I find that nothing beats S-exps for reducing cognitive load
03:18	ChongLi	having a lot of special syntax and weird operators can make you go cross-eyed after a while
03:18	doomlord	one of the main things i'm after in my 'other language quest' is reflection for serialization (is it reflection or introspection) .. the abliity to walk structure members.. in C++ its a mess
03:18	doomlord	haskell is really bad for special operators .
03:18	doomlord	<$> .. wtf
03:18	ChongLi	or it's really good
03:18	ChongLi	depending on your perspective
03:19	ChongLi	guys who like to play perl golf will feel right at home
03:19	doomlord	in lisp i can handle prefix maths just fine, its reminiscent of low level code plus doing vectormath in C.. i'm used to prefix notation helpers. What i find myself missing though is dot notation for component access.
03:20	ChongLi	what's wrong with using keywords?
03:20	doomlord	but clojure seems to help quite a bit with destructuring and callable objects, i like that
03:20	ChongLi	destructuring is really nice
03:20	doomlord	(object :component) is still not as good as object.component
03:20	ChongLi	especially the :keys form
03:21	doomlord	what i really like away from c++ is tuples, not having to declare or wrap multiple values.
03:21	doomlord	i like opportunities to avoid naming things
03:22	doomlord	in c++ you're having to make or refer to temporary classes alot
03:23	ChongLi	also did you know you can use sets as predicates for filtering?
03:24	ChongLi	,(filter #{:a :b :c :d} [1 9 :a 19 :c 'x \u])
03:24	clojurebot	(:a :c)
03:24	doomlord	i didn't realise that no
03:24	ChongLi	stuff like that is so cool and useful
03:24	doomlord	yes it definitely is
03:25	doomlord	the ability to make literal maps (like anon structs i guess) seems really good.
03:25	scottj	doomlord: it's actually easy to do object.component in your own clojure build, though you can't use "." http://youtu.be/OiaYxaONChg?t=2m
03:26	ChongLi	I love how clojure's data structures lead to a different style of working
03:26	doomlord	interesting
03:26	ChongLi	since you can make most things using only the built-in data structures
03:26	ChongLi	no need to declare new ones the way you do in most languages
03:26	doomlord	yeah the combination of maps and vectors is really powerful
03:26	ChongLi	and the sequences library
03:27	ChongLi	lets you leverage them all in a very clean way
03:27	doomlord	object:value , interseting
03:28	doomlord	i'll persevere with the parens. there's enough else to like
03:28	scottj	I think I also did #.object.value once
03:29	scottj	at that point though (:value object) is better :)
03:29	doomlord	its (slot-value object value) that is too much :) (object :value) is tolerable
03:31	doomlord	just tried writing vector maths routines, these are very pleasing with destructuring
03:31	ChongLi	yeah
03:31	ChongLi	have you looked into multimethods?
03:32	ChongLi	they're an extremely powerful feature of clojure that aren't talked about that much
03:32	ChongLi	a la carte polymorphism with ad-hoc hierarchies is really nice
03:32	doomlord	i'm aware of how CLOS does it, i much prefer this sort of thing to the traditional OO style of defining methods inside classes
03:32	doomlord	multiple dispatch is rather interesting too
03:33	doomlord	i really despise class heirachies. i like the way Go does things.
03:33	ChongLi	ad-hoc hierarchies are just really nice and clean
03:33	ChongLi	you use derives to build them
03:34	ChongLi	(derive ::rect ::shape)
03:34	ChongLi	for example
03:34	ChongLi	the other really cool thing about multimethods is they can dispatch on any arbitrary function
03:35	doomlord	what do you mean by that
03:35	ChongLi	(defmulti cool even?)
03:35	doomlord	someone said clojure gives more control over the dispatch compared to CLOS
03:36	ChongLi	in this case, even? is the dispatch function
03:36	ChongLi	with this multimethod you dispatch on whether something is even or odd
03:37	doomlord	a bit like pattern matching on steroids perhaps
03:37	ChongLi	so you don't even need to care about types, you dispatch on values
03:37	ChongLi	,(defmulti cool even?)
03:37	clojurebot	#<Exception java.lang.Exception: SANBOX DENIED>
03:37	ChongLi	oh
03:37	doomlord	can that emulate c++ style overloading , eg make 'add' work on values or vectors
03:38	doomlord	,(vector? [1 2 3])
03:38	clojurebot	true
03:38	ChongLi	well for that you usually use protocols or java interfaces
03:39	doomlord	i really wish there was a fixed version of c++. so many irritations in it seperate to the reason for using it, i.e. manual memory management with no complex runtime
03:40	ChongLi	I think many people find manual memory management a big irritation
03:40	noidi	doomlord, isn't D pretty much a fixed C++?
03:41	doomlord	must be reasons its not so popular . its still got a garbage collector in there so its more overlapping with other languages than c++ perhaps
03:41	ChongLi	D's GC is optional
03:43	noidi	doomlord, didn't you just mention manual memory management as one of the irritations that should be fixed? :)
03:43	noidi	ah, sorry, I misread
03:43	ChongLi	I think it could be read either way
03:43	doomlord	because its a core-feature rather than an irritation. My perception is: you use C++ where you need manual memory management. its still got enough improvements over C to be worth using
03:44	ChongLi	and yet linus hates it
03:44	noidi	I prefer C+Lua over C++
03:44	ChongLi	http://article.gmane.org/gmane.comp.version-control.git/57918
03:44	ChongLi	this always cracks me up
03:44	doomlord	yeah that seems a good combination
03:44	clojurebot	cemerick: Yeah, MacKay's book is great too.
03:45	doomlord	i find C++ templates are useful for expressing solutions to low level problems like making offset pointers. but the high level parts suck
03:47	noidi	https://twitter.com/AlecMuffett/statuses/252512427567108097
03:47	noidi	"We can't teach C++ anymore. It's like trying to explain "Lost" now - you had to be there. No one's going to watch all 6 seasons"
03:47	doomlord	heh.
03:52	arrdem	hey. I managed to teach myself the C parts of C++ in one summer...
03:52	arrdem	gave up when I hit the actual C++
03:53	arrdem	so can't be that hard right? XP
04:08	doomlord	is there a function to modify an element of a vector (??? [10 20 30] 1 5) -> [10 5 30]
04:09	doomlord	(nondestructive)
04:09	doomlord	probably want to do that as a batch... indices to replace and new values
04:12	Chousuke	doomlord: replace
04:15	hoeck	doomlord: vectors are associative
04:15	ivan	,(assoc [10 20 30] 1 5) ; or the thing that isn't replace
04:15	clojurebot	[10 5 30]
04:15	hoeck	,(assoc [1 2 3] 0 0 2 0)
04:15	clojurebot	[0 2 0]
04:15	doomlord	interesting
04:17	ejackson	lpetit: bonjour
04:18	lpetit	ejackson: hello dear
04:18	ejackson	lpetit: happy hacking-day :)
04:18	lpetit	thank you :)
04:19	lpetit	currently fighting my limitations in writing macro-defining macros :)
04:20	lpetit	ejackson: how are you?
04:26	ejackson	macro-defining macros !
04:26	ejackson	that turtles all the way down dude... you can't win !
04:26	ejackson	i'm fine, myself.
04:29	lpetit	ejackson: currently wrapping Eclipse' tracing API. Classical wrapping of (when (trace-option-enabled? ..) …) with a macro. Then there are 4-5 flavors of tracing calls => removing duplication between the macros code. And then, I'd like to have a macro creating the relevant macros for a specific eclipse plugin :)
04:30	lpetit	I believe I can fly ! :)
04:30	ejackson	it might be easier...
04:30	ejackson	:)
04:31	ejackson	you are a brave, brave man
04:36	lpetit	ejackson: I am a fool, you mean :)
04:36	ejackson	hahaha - not at all.
04:37	ejackson	but you'd better put some wine in the fridge for this evening :)
04:37	lpetit	:-D
04:38	ejackson	failing to do that would be foolish
04:39	clgv	"macros generating macros - thats perverse!" (structural quote... ;) )
04:40	lpetit	clgv: and it's for you
04:41	lpetit	I want to be able to leave traces in ccw, while not polluting everybody … so configurable traces … so Eclipse Trace API :)
04:41	clgv	lpetit: not for us all - the tools for a better future in debugging CCW? ;)
04:41	lpetit	… or how spending 5 hours what could have been hacked in 5 seconds :)
04:42	lpetit	clgv: yes
04:42	lpetit	but you meant "no", not "not", right?
04:42	clgv	I meant "not for us all?" ... badly puncutated...
04:47	lpetit	clgv: ok :). You're right, that's because I intend it to be generally useful that I didn't, yet again, throw a quick println until I remember to remove it … 5 months later, say ? :)
04:49	clgv	lpetit: I've been pretty busy last weekend and will be next weekend. so my changes wont be made until the end of next week...
04:49	lpetit	clgv: noted, thx for the status update
05:03	lpetit	clgv, ejackson : so far, I think I've been falling into every imaginable trap of macro writing. Let's start level 2 by externalizing macro-defining macro into its proper namespace :)
05:04	clgv	lpetit: oh I gather some experience with that. are you struggeling with somethin particular right now?
05:04	clgv	*gathered
05:04	ejackson	lpetit: aah, yeah I did that (if I understand you !)
05:04	clgv	I have written a Meta-DSL to generate a DSL which is used for concrete documents then
05:05	ejackson	otfrom: mornin' Mr Wayne
05:05	lpetit	clgv: no, not started yet, still finishing debugging current situation. But I guess I'll have to look closely at each auto-namespaced symbol because it will not have to be expanded into the macro-defining macro's own namespace, but rather within the defined macro's namespace
05:07	lpetit	arch, reflection code added by clojure adding noise to my stack and preventing eclipse to properly report calling function name :)
05:07	lpetit	s/arch/argh
05:10	otfrom	morning ejackson
05:23	Kototama	hi, any ideas why I get compilation warning with the jaqy library even if compiling without optimization? https://gist.github.com/3910617
05:24	Kototama	I mean jayq
05:31	mpenet	Kototama: this disapears with alpha4
05:44	Kototama	ah! okay
07:16	abp	Hm, how do I set up a Clojurescript REPL attached to the code running in my browser?
07:18	abp	I already start an nrepl server for my server-side code and connect to that using ccw repl, and i could use the cljs middleware of piggieback with that server too, but where/how is the connection to the browser made?
07:47	goracio	anyone use monger ?
07:48	antares_	goracio: I am the author of monger
07:49	goracio	antares: great have couple questions :)
07:49	goracio	antares_: great have couple questions :)
07:51	goracio	antares_: suppose we have this (mg/find-maps "messages" {:nik a }) and a is a vector, so how to choose all messages with :nik that is in vector
07:51	goracio	antares_: a for example ["dave" "martin"]
07:51	antares_	goracio: using MongoDB query operators
07:52	antares_	$in, $all, etc
07:52	goracio	antares_: and :nik field is a string like :nik "dave"
07:52	antares_	goracio: see http://clojuremongodb.info/articles/querying.html#using_mongodb_query_operators and MongoDB query docs
07:53	goracio	antares_: suppose we have this (mg/find-maps "messages" {:nik {$in : a} }) something like this ?
07:56	nonrecursive	Hi all - I'm trying to run my leiningen project and get two errors which alternate each time I try. The errors are "Too many arguments to def" and "java.lang.NoSuchMethodError: clojure.lang.RT.keyword"
07:56	nonrecursive	would anyone know how to debug this?
07:56	antares_	goracio: yes. It's the same as in mongo shell. See the docs.
07:56	antares_	nonrecursive: are you sure you are not using Clojure 1.2-only code?
07:57	nonrecursive	antares_: yeah, i'm using clojure 1.4 for a web app
07:58	antares_	nonrecursive: yes but are you sure you are not using old contrib libraries or any other 1.2-only code?
07:58	antares_	that exception is from the 1.2-to-1.3 migration age as far as I remember
07:58	goracio	antares_: seems works :) thanks
07:58	antares_	although "too many arguments to def" is probably an issue in your own code
07:58	nonrecursive	antares_: ah ok - I'm not sure about that
07:59	nonrecursive	antares_: it's bizarre because the error changes every time, even when I don't make changes in code
08:01	antares_	nonrecursive: files that do not require each other can be compiled in different order
08:01	clgv	nonrecursive:yeah your code has probably some wrongly place parantheses
08:01	nonrecursive	antares_: ok it looks like requiring clojure 1.3 instead of 1.4 fixed the issue
08:02	clgv	nonrecursive: thats no good sign
08:02	antares_	nonrecursive: 1.4 is a drop in replacement for 1.3, has been the case for me in almost 30 codebases
08:02	nonrecursive	clgv: I think the "too many arguments to def" must have come about through something funky with a macro in one of the libraries I'm using
08:04	clgv	nonrecursive: if I were you I'd switch back to 1.4 and hunt down the error in the code.
08:04	nonrecursive	ok now this is weird - i've switched the clojure version back to 1.4 and now it's working again
08:04	nonrecursive	i mean, still working
08:04	antares_	nonrecursive: run lein clean and try again
08:04	antares_	you possibly have some AOT-compiled code around
08:04	clgv	nonrecursive: well, then a "lein clean" would have solved it as well, I guess ;)
08:05	nonrecursive	oh shoot, the error is back
08:05	nonrecursive	after doing lein clean
08:06	nonrecursive	ok I'm going to dig into this more to see if I can resolve that "too many arguments" error - thanks for the help
08:07	clgv	lpetit: still around?
08:08	lpetit	still, and having greatly improved my macro-foo so far
08:08	john2x	hi.
08:08	lpetit	nested syntax quotes have no secret for me anymore :)
08:08	abp	lpetit: How is CCW Clojurescript integration going? ;)
08:08	clgv	lpetit: I was wondering whether maybe the history search should be displayed similar to the autocompletion in an additional widget.
08:08	john2x	is there a way to reload all the files currently loaded in the repl?
08:09	clgv	lpetit: ok insider comment ~'~'my-symbol ;)
08:09	abp	lpetit: About to start piggieback browser-connected ccw repl battle..
08:09	lpetit	clgv: and sometimes ~'~(symbol (name (ns-name ns)) "foo") :)
08:10	lpetit	abp: ?
08:10	clgv	lpetit: been there ;)
08:10	clgv	lpetit: your first reaction to the above suggestion?
08:11	abp	lpetit: I want to connect the repl of counterclockwise to a clojurescript-app running in the browser. Seems like it's not very straight forward from what I've read so far?
08:12	lpetit	clgv: so that before "accepting" (hitting enter?) the input area stays the same, and the filtered history appears in a list ?
08:12	nonrecursive	antares_: clvv: it looks like I get the error whenever I put (:use korma.core) in a namespace declaration
08:12	lpetit	abp: what's not straightforward is to launch this browser connected repl from ccw
08:12	clgv	lpetit: yeah that was the rough idea - but I dont know yet how to incorporate different search modes there.
08:13	clgv	lpetit: what I wanted to know is whether that would be rejected by you at once for some reasons ;)
08:13	lpetit	abp: but if you launch it straight from lein2 on the command line, you just need to get the port number, and it's straightforward from ccw to connect to this port: Window > Connect to REPL
08:13	lpetit	clgv: seems interesting idea
08:14	lpetit	clgv: no great idea out of the box right now
08:14	lpetit	(out of my head, I mean)
08:14	clgv	lpetit: and a lot of work compared with the current state of it
08:15	lpetit	clgv: indeed, which will probably be the compelling reason for you to postpone this enhancement :)
08:15	clgv	lpetit: very likely ^^
08:15	Tichodroma	(source my-stuff.core/foo
08:15	lpetit	abp: will be around for the next couple hours, so if you're having problem, feel free to ping me
08:15	Tichodroma	woops, wrong buffer
08:15	abp	lpetit: Ah ok, then I should try that first. I'm a little confused about the port I pass to start a Clojurescript repl.. Can I have a ring webapp and the repl run on the same port?
08:16	antares_	nonrecursive: I am not familiar with korma but they keep 1.2 compatibility and korma.core is 500+ lines long. Naked use may or may not be a good idea.
08:17	lpetit	abp: there can be so many ways to start a nrepl server from within your test environment. I must know a little bit more about how you do (or intend to do) start it.
08:18	lpetit	abp: if you want the ring web app and the repl on the same port, then you'll need your nrepl server to be served through http, right ? That means you'll have to use a library provided by cemerick which allows this. tl;dr: yes
08:18	lpetit	abp: https://github.com/cemerick/drawbridge
08:20	abp	I use lein-ring to start up my Ring webapp on port 3000. lein-ring supports an :init function specified in project.cls, which I then use to start an nREPL-server. I window/connect to repl in ccw onto the nREPL-server to eval Clojure. Now I want something similiar so I can have an clj and cljs repl side by side ;)
08:20	abp	*project.clj
08:21	lpetit	abp: so currently you're having ring started serving http on port 3000, and an nrepl server inside the same JVM started on another port
08:21	abp	lpetit: yep, nrepl on 7888
08:22	lpetit	abp: you could have started a :repl on port 7888, and used the :injections (or whatever) from it to start your ring handler LOL too :)
08:23	lpetit	abp: with the last version of CCW, it's easy: in the View for your "Clojure" REPL, there's a button to quickly open a second View connected to the same port.
08:24	lpetit	abp: in this second view, you can just issue the usual cljs incantations to transform it into a cljs repl, thanks to the magic of nrepl middle wares.
08:25	lpetit	abp: so the "hard" part will be in correctly configuring the piggyback middle ware for nrepl. At this point, this is purely a lein2/nrepl configuration thing, not really related to ccw (piggyback will make the thing transparent to ccw, afaik)
08:26	lpetit	abp: really, it'll be a piggyback / lein2 configuration battle :)
08:26	lpetit	(and I cannot help here, I have no experience)
08:28	abp	lpetit: Ok, then I will get my hands at it. Since diving into the whole cljs-thing I lost a lot of time by reading up things and got it working pretty quickly.. So more a problem of hesitation than anything else.. Thanks so far :)
08:31	nonrecursive	OK I think I've fixed the problem. Where I had (defdb db …) I now have (if (not (resolve 'db)) (defdb db …)
08:31	lpetit	abp: a comprehensive tutorial on setting up a whole ring+cljs with lein2 and piggyback is still missing in the clojure ecosystem, AFAIK. As well as its complement for (pre)prod situations with drawbridge to server the repl through HTTP/BASIC protection :)
08:34	abp	lpetit: Yes, theres a whole lot missing in terms of content on cljs atm, but nothing is really hard. Things like using a js-library under advanced compiliation seem quite hard but are eventually dead easy.
08:36	lpetit	abp: I don't doubt it's not hard in the end, and once you know it. Getting to it right now, though, seems scary. I miss some easy to grasp big picture drawings
08:36	abp	lpetit: That's it.
08:39	djanatyn	I started playing around with threads in sbcl and it's defeinitely not as fun as it is in clojure
08:39	djanatyn	clojure was my first experience with multiple threads. now I appreciate clojure's concurrency model even more
08:44	lpetit	clgv: yeepee ! macro-defingin-macro working !
08:44	lpetit	do you want to see the mess i created ? :)
08:44	clgv	lpetit: :) sure, lets see
08:46	lpetit	clgv: not final state, but a working state ! : https://gist.github.com/3911545
08:48	clgv	lpetit: not too messy. but I can only guess how to use it ;)
08:48	lpetit	I'll show you in a dozen minutes, there are still simplifications to be made
08:48	clgv	lpetit: dont hurry. I might be leaving early today since I caught a cold...
08:48	nonrecursive	clgv: antares_: ok I found the actual culprit, it was a bug in Korma - https://github.com/korma/Korma/issues/51
08:49	lpetit	clgv: ok
09:01	alex_baranosky	hey guys, what are some common solutions for handling composure routes that can optionally end in a trailing slash?
09:01	alex_baranosky	it looks like Compojure used to accept regexes as routes, but doesn't any longer
09:03	alex_baranosky	has anyone extended the Route protocol to regexes for their projects? https://github.com/weavejester/clout/blob/master/src/clout/core.clj#L71
09:08	LordPinky	alex_baranosky, ZvZ is such a beautiful matchup
09:13	xeqi	alex_baranosky: I've seen middleware that rewrites the :uri
09:13	xeqi	another option is to redirect instead of serve the page on both uris
09:13	alex_baranosky	xeqi: yeah seen that too… Seems silly to have to work around the framework
09:13	alex_baranosky	LordPinky: I don't understand
09:14	LordPinky	alex_baranosky, ah, you don't play StarCraft?
09:14	alex_baranosky	xeqi: what do you think of the Compojure protocol extending approach?
09:14	LordPinky	You should, everyone here does.
09:14	LordPinky	If you don't, you're not part of it.
09:14	LordPinky	Can't sit at the cool table if you don't bro, those are the rules and we do enforce them.
09:14	alex_baranosky	LordPinky: correction: everyone here does CLojure
09:15	AdmiralBumbleBee	not everyone here does clojure
09:15	LordPinky	just a ruise honestly, this channel is where all the people who play StarCraft on freenode come together under the praemise of discussing a programming language while they actually brag about their blink stalker micro.
09:15	AdmiralBumbleBee	pretty sure sgeo only does haskell with lots of parens ;)
09:16	alex_baranosky	xeqi: looks like it might be complicated (bad) to extend that protocol, because the String implementation is doing a ton of parsing etc
09:18	xeqi	alex_baranosky: yeah, I can't follow it on quick glance so it might be messy
09:30	lpetit	clgv: here's the way to install the code for ccw.core: https://gist.github.com/3911733
09:31	clgv	lpetit: pretty compact. after that statement it will trace every eclipse event?
09:31	lpetit	clgv: hell no :)
09:32	clgv	lpetit: what does it do? I see no config in that gist ;)
09:32	lpetit	clgv: after that statement, it will install in ccw.core.trace a handful of macros wrapping the eclipse tracing api
09:32	Sgeo	AdmiralBumbleBee, heh
09:34	lpetit	I'm now able to write (ccw.core.trace/trace :autocompletion "some log message") and it will check if the "ccw.core/autocompletion" Eclipse trace flag is set to true, and if so, it will add the appropriate call that will add a trace in the user's workspace/.metadata/trace.log file (with file/line number, thread, and the message)
09:34	AdmiralBumbleBee	Sgeo: I learn a lot from your musings, so no harm intended :)
09:35	antares_	Sgeo: do you mind a small haskell question? I am scared of hardcore academics in #haskell :)
09:35	lpetit	clgv: starting from Juno, Eclipse will let the user toggle the trace flags from the preferences > trace page.
09:35	lpetit	clgv: prior to Juno, the user (you) has to start Eclipse with the -debug (don't ask me why it's not named -trace) flag and the path (url?) to the file defining which debug flags are on
09:38	Sgeo	antares_, in general I don't mind, but right now I'm running late for class
09:46	antares_	Sgeo: ok
10:25	abp	xeqi: Do you have a clue how to get piggieback with a browser-repl working when the page containing the cljs is served from a ring app?
10:25	dfd	Do you folks know if there's a way to run just one test in midje? While I'm at it, what is the appropriate way to stub out a function "globally" (for a whole set of tests hitting the same function), except, er, on the tests where you don't want to stub it out...?
10:39	cemerick	abp: you mean the cljs side of the browser repl, or the cljs for your app?
10:40	abp	cemerick: Hi there. I want to execute cljs in my browser from a ccw repl.
10:41	cemerick	abp: Sure; to do so, you'll just need to start your REPL from Leiningen (so the piggieback middleware can be rolled in to the stack), and connect to it from ccw.
10:42	cemerick	You'll still need to set up your app to look for a browser-repl connection (e.g. on port 9000).
10:42	abp	cemerick: Currently I have that in lein-rings :init, starting an nrepl server with piggieback middleware
10:42	abp	cemerick: Ah ok, so that port is separate from everything else?
10:42	cemerick	ah, ok, that'll work too
10:43	abp	cemerick: Then I think it's just about to work ;)
10:43	cemerick	Yes, the browser-repl controls its own set of resources.
10:43	abp	cemerick: Let me try.
10:43	cemerick	There was talk a bit ago about trying to make it possible to route them through an existing ring app, but that's apparently not in the cards.
10:45	abp	cemerick: Your clojurescript dependencies in piggieback conflict with the way cljsbuild works, so i need to exclude them :x Or probably cljsbuild conflicts with everything else..
10:46	cemerick	You certainly need to be using the same version of clojurescript in both places.
10:46	abp	cemerick: Sure, I think cljsbuild uses a newer version, but not via lein deps but bundled.
10:47	cemerick	abp: nothing is bundled; the cljsbuild jar is ~8kb
10:49	abp	cemerick: https://github.com/emezeske/lein-cljsbuild/issues/105
10:49	abp	cemerick: Not bundled, but pulled, sorry.
10:54	abp	cemerick: Ehm, so I still need to serve an index.html from port 9000 then, not integrating the repl into my ring-apps cljs?
10:55	cemerick	abp: No; at least, I don't
10:55	cemerick	But, I don't grok that cljsbuild issue, either.
10:57	cemerick	AFAICT, cljsbuild always uses whatever version of ClojureScript you have in your :deps vector
10:59	abp	cemerick: Ok, then this issue would be obsolete. I don't have a ClojureScript dep in my :deps and no .jar in the leiningen deps folder. So somehow it's got ClojureScript.
11:00	cemerick	abp: deps folder? What version of leiningen are you using?
11:00	casperc	hey
11:01	casperc	stupid question: Is there a function in core which returns true if it gets nil and nil otherwise?
11:01	abp	cemerick: lein2, I don't mean a physical folder, only the view in eclipse "Leiningen Dependencies"
11:01	cemerick	ah
11:01	andrewmcveigh\|wo	casperc: nil?
11:02	casperc	that returns false if not nil unfortunately
11:02	andrewmcveigh\|wo	casperc: ah, sorry. didn't read your question properly.
11:02	casperc	it should be like this (if val nil true)
11:02	cemerick	casperc: what are you trying to do?
11:02	casperc	np
11:03	casperc	bit hard to explain the whole thing, but i need a nil value when i get something which is not nil
11:03	casperc	i guess i can just do the if thing, but i just feel stupid for doing it :)
11:04	cemerick	sounds like a specious requirement :-)
11:05	abp	cemerick: So, I'm doing the things described in the Browser REPL section of piggiebacks readme. The only difference is, I've a ring app running on localhost:3000, delivering the page my cljs is included in. When I call (repl/connect "http://localhost:9000/repl") from there, should that work? Despite me having no index.html served from localhost:9000 etc.
11:07	cemerick	abp: You have no control over what's served from 9000 at all; that server comes up when you call (cemerick.piggieback/cljs-repl ………)
11:09	antares_	so, for every Raynes we send to the conj they release 3 new tickets? That sounds like a price increase some may just accept. Now, where can we get additional Rayneses…
11:09	abp	cemerick: Ok, but the cljs-repl does not work. I can't evaluate anything.
11:10	cemerick	abp: make sure you're loading your web page that connects to :9000 after starting the cljs REPL.
11:22	notsonerdysunny	Is it possible to import 2 classes of same name from two different packages into the same clojure namespace?
11:22	notsonerdysunny	I seem to be running into name clash related compilation errors..
11:22	abp	cemerick: Doesn't work either. But I should be able to see the browser polling on localhost:9000 in chrome developer tools?
11:23	cemerick	Yes, though it only attempts to connect once; I don't think it polls.
11:23	abp	cemerick: Ok, but still a visible network connection. I've got none.
11:28	abp	cemerick: It's getting better, sorted out one error of my own.
11:29	abp	cemerick: Do you use lein-cljsbuild?
11:29	cemerick	yes
11:29	cemerick	though I still don't see how that's related to browser-repl
11:30	chouser	4clojure is down. Did I kill it?
11:30	abp	cemerick: Not at all. Having no issues with it? In terms of stability, reproducible errors.
11:30	cemerick	nope, it's been solid
11:30	chouser	hm, back up
11:31	abp	cemerick: Strange. I'm using shoreleave and jayq, but sometimes I get errors in the js-console that are gone after a clean and rebuild.
11:32	ohpauleez	abp: Shoreleave related errors or just general errors?
11:33	abp	cemerick: Done. browser-repl works, thanks for another great job on an astonishing tool. ;)
11:33	cemerick	abp: what was the issue?
11:35	abp	ohpauleez: General errors. I get a few warnings for Shoreleave while building, nothing severe, I'm still fiddling around how to structure my app reactive around pubsub and minimizing state. Having a blast. :)
11:35	xeqi	abp: glad to here you got it working
11:35	abp	cemerick: A bug in my code and not having repl/connect "http://localhost:9000/repl") eval'd.
11:35	cemerick	ah :-)
11:35	ohpauleez	abp: Good to hear. Feel free to ping me or email me (via the shoreleave ml) if you have any questions
11:36	abp	lpetit: Would it be hard to highlight (comment ...) blocks in ccw?
11:36	ohpauleez	there is one outstanding pubsub bug I need to patch up
11:36	abp	cemerick laughs hard. :D
11:36	ohpauleez	but I'm also sitting on a few other changes I need to polish up before releasing
11:36	lpetit	abp: you mean show them as ;#_ comments?
11:36	cemerick	abp: FWIW, I have this as a top-level in one of my app's cljs files:
11:36	cemerick	(when (-> js/window .-location .-hostname (= "localhost"))
11:36	cemerick	(connect "http://localhost:9000/repl"))
11:37	ohpauleez	cemerick: That deserves to be in Shoreleave haha
11:37	ohpauleez	my trick is query args (so I can use it in production)
11:37	abp	ohpauleez: Didn't ran into that bug. ;) I have one function publish changes from a user typing in an input. Not much more.
11:37	cemerick	ohpauleez: nah, shouldn't be in libraries :-)
11:37	abp	lpetit: Yes, please.
11:38	xeqi	ohpauleez: where is the shoreleave ml?
11:38	cemerick	abp: once ccw uses leiningen to start its REPLs, then stuff around piggieback etc. will get a lot easier
11:38	lpetit	abp, cemerick : dunno. Highlighting (comment) as ; or #_ comments … what if the user uses something else than clojure.core/comment ?
11:39	abp	When get things in cljs files executed? Is there some auto-wrapping into a jquery-ready like handler?
11:39	lpetit	cemerick: how much easier? I mean, the hard stuff will then still be made in lein2 configuration,and only grabbing nrepl server's port and starting client repl automatically will be easier.
11:40	hldfr	Hmm why isn't some named as some? , just like the every? sequence predicate ?
11:40	S11001001	hldfr: because some isn't a predicate
11:41	S11001001	,(some seq '[() [] [1]])
11:41	clojurebot	(1)
11:41	hldfr	,(some odd? [])
11:41	clojurebot	nil
11:41	hldfr	ah
11:41	cemerick	lpetit: The difference between setting up your own nREPL stack vs. using the easy lein2 configuration available for doing the same is significant.
11:42	cemerick	Nevermind other plugins, middleware, hooks, etc.
11:42	S11001001	hldfr: it's worth comparing some to (comp first filter), and considering which is appropriate in each situation
11:42	gfredericks	abp: I believe it's all at the top level; you can check the compiled file though
11:42	lpetit	cemerick: you mean in case one does not want to start lein2 from the command line without waiting for better integration with ccw ?
11:43	S11001001	,((comp first filter) seq '[() [] [1]])
11:43	clojurebot	[1]
11:43	cemerick	lpetit: Right; you lose code completion and such doing that.
11:44	lpetit	ooh, I wasn't aware of that
11:45	cemerick	lpetit: Reason #43 I've been harping on about completion as middleware, instead of using eval for it.
11:49	abp	cemerick: Seems like the next stop for me would be lerning to use lein2 properly. ;)
11:50	abp	lpetit: You can't get at the ns of a symbol for highlighting currently?
11:51	wingy	what version of eclipse should i download for clojure?
11:52	abp	gfredericks: Sure
11:52	wingy	Eclipse IDE for Java Developers or Eclipse Classic
11:52	lpetit	wingy: both should do. Go for "for Java Developers", I guess
11:52	lpetit	bbl
11:52	abp	gfredericks: Just wondering why everybody is happily hacking top-level code while that wasn't good in plain js
11:57	gfredericks	abp: everybody?
11:57	abp	cemerick: "GET http://localhost:3000/deps.js 404 (Not Found) main.js:165" shows up in the js-console. That happens randomly while using lein cljsbuild auto. That is a bug of an older ClojureScript version or not?
11:58	abp	gfredericks: Ok, some folks in examples I've seen.
11:58	gfredericks	abp: when I need dom-related side-effects I always wrap the call to the side-effecting function
11:58	gfredericks	something like (js/jQuery init)
11:59	gfredericks	abp: if you see people doing otherwise you have my permission to frown
12:00	abp	gfredericks: Yes, that's what I'm doing too, until now in an anonymus function, because I'm only tampering.
12:03	doomlord	was just trying to compile/run a couple of modules, the tools are telling me errors about classpaths .... is it not possible to just combine a few sourcefiles from a single commandline like with ghc or gcc
12:04	gfredericks	doomlord: are you using lein?
12:04	doomlord	just started trying that aswell
12:05	doomlord	is this extra complexity imposed by the jvm
12:06	doomlord	(heh i guess i could try a script that concatenates source files :) )
12:06	gfredericks	doomlord: lein is pretty easy once you're familiar with it
12:07	doomlord	do i need a whole lein generated directory structure for a library
12:07	gfredericks	one big win over manually compiling source files is that with lein it is trivial to tap into the ecosystem of clojure/java libraries
12:07	gfredericks	yeah that's generally what you want; `lein new ...` will set one up for you
12:08	gfredericks	putting the source files in directories that match the namespace is standard java/clj procedure
12:08	doomlord	ok i guess i was expecting this 'middle ground' of one step up over a single sourcefile to be as easy as c ... i suppose at least if this chews up more neurons its knowledge applicable to a range of tools not just clojure
12:09	doomlord	ok thats my first mistake
12:09	doomlord	i just setup a lein project and stuck all my sources in there
12:09	doomlord	(in the one directory it made)
12:10	gfredericks	I'm sure it's technically possible to compile a pile of files if you use things like clojure.core/load, but using require (which I believe assumes the aforementioned directory structure) is _much_ more idiomatic
12:10	doomlord	in this 'middle ground' my "library" is merely one file
12:11	noidi	creating a new project with lein only takes a couple of seconds, so there's not much need for a middle ground, imo
12:11	gfredericks	there's a lot of uniformity gain with a small amount of effort
12:11	noidi	lein new foo; cd foo; lein repl
12:11	doomlord	foo\src\foo .. :(
12:11	gfredericks	hundreds of libraries on github with standard directory structures
12:12	doomlord	ok i guess i can stil a bunch of my helpers together to avoid directory overhead
12:13	noidi	doomlord, there's also this https://github.com/mtyaka/lein-oneoff
12:13	noidi	oops, apparently that's for leiningen1
12:14	noidi	so disregard that :)
12:14	doomlord	if i've just got 4-5 sourcefiles ... but 4 i want to re-use - i'm going to need 2 lein projects then i guess
12:15	doomlord	deeply nested directories (foo/src/foo) makes commandline use much more awkward :(
12:16	doomlord	if you've got hundreds of sources then great
12:16	gfredericks	could make a symlink
12:17	doomlord	"if you use things like clojure.core/load"... how would that look
12:18	gfredericks	I think you just give it a filename and it evals the file
12:19	doomlord	1st attempt: in the repl, in the directory where the source is - "(load "sourcrname.clj") ... errors about classpath :(
12:20	gfredericks	oh there's a load-file as well
12:20	gfredericks	that must be what I was thinking of
12:20	doomlord	so far i've used "$ clojure <singlesourcename.clj>" to run a single file project. (very tempted to make a script to concat sources and run that..)
12:21	gfredericks	you can go ahead building your own thing if that's what you enjoy doing, but anytime you ask for help about it people will ask you why you're not using lein
12:21	doomlord	ok load-file works - thanks :)
12:21	gfredericks	np
12:22	doomlord	ok i guess "lein" is worth getting into if i want to, say, dable with clojure + scala
12:22	doomlord	or when i've written more than 5 sources
12:22	gfredericks	and it will ease your ability to do lots of other common build-related tasks
12:23	gfredericks	e.g., make executable jarfiles
12:23	doomlord	so by default java land has many .classes which can be loaded add-hoc? but jarfiles are a self-contained bundle?
12:24	konr_trab	Is there a way to deconstruct the keys and values of a hash in a defn? For example, {:foo {:bar "baz"}} would have :foo set to a, :bar set to b and "baz" to c?
12:24	doomlord	map gets them i think
12:25	doomlord	i think mapping a hash gets[key value] 's passed in?
12:25	doomlord	,(map #(print %) { :a 1 :b 2 :c 3})
12:25	clojurebot	([:a 1][:c 3]nil [:b 2]nil nil)
12:26	konr_trab	hmm, but can't I do something like (let [{a {b c}} {:foo {:bar "baz"}}] (list a b c))?
12:27	gfredericks	konr_trab: you can
12:27	gfredericks	well
12:27	doomlord	you could build a vector and destructure the vector perhaps? is there a better way
12:27	gfredericks	if you don't know the keys ahead of time then it's messier
12:27	gfredericks	that is a much less common thing to want to do
12:28	gfredericks	,(let [[[a [[b c]]]] {:foo {:bar "baz"}}] (list a b c))
12:28	clojurebot	#<UnsupportedOperationException java.lang.UnsupportedOperationException: nth not supported on this type: PersistentArrayMap>
12:28	dnolen	konr_trab: that doesn't really make sense given there can be many more key value pairs - hash are unordered so you don't know what you would get.
12:28	gfredericks	hrm :(
12:28	AdmiralBumbleBee	I thought you couldn't destructure nested maps?
12:28	doomlord	,(let [ [k0 v0][k1 v1] ] (map #([%1 %2]) { :foo 1 :bar 2} ) )
12:28	clojurebot	#<CompilerException java.lang.RuntimeException: Unable to resolve symbol: k1 in this context, compiling:(NO_SOURCE_PATH:0)>
12:29	dnolen	AdmiralBumbleBee: you can
12:29	AdmiralBumbleBee	has that always been the case?
12:29	doomlord	,(let [[ [k0 v0][k1 v1] ] (map #([%1 %2]) { :foo 1 :bar 2} )] (print k0 v0 k1 v1) )
12:29	clojurebot	#<ArityException clojure.lang.ArityException: Wrong number of args (1) passed to: sandbox$eval127$fn>
12:29	dnolen	AdmiralBumbleBee: yes
12:30	gfredericks	doomlord: I think the classfiles in the jars are loaded in the same way; the jvm has a 'classpath' which can be a combination of raw directories and jarfiles containing classfiles
12:30	doomlord	,( let [ [[k0 v0][k1 v1]] (map #([%1 %2]) { :foo 1 :bar 2}) ] (print k0 v0 k1 v1) )
12:30	clojurebot	#<ArityException clojure.lang.ArityException: Wrong number of args (1) passed to: sandbox$eval163$fn>
12:30	doomlord	,( let [ [[k0 v0][k1 v1]] (map #(list %1 %2) { :foo 1 :bar 2}) ] (print k0 v0 k1 v1) )
12:30	clojurebot	#<ArityException clojure.lang.ArityException: Wrong number of args (1) passed to: sandbox$eval199$fn>
12:31	AdmiralBumbleBee	I even have code in the project I'm working on that destructures a nested map… and I wrote the code
12:31	doomlord	,( let [ [[k0 v0][k1 v1]] [[:a 1][:b 2]] ] (print k0 v0 k1 v1) )
12:31	clojurebot	:a 1 :b 2
12:31	doomlord	(map #(list %1 %2) {:a 1 :b 2})
12:31	doomlord	,(map #(list %1 %2) {:a 1 :b 2})
12:31	clojurebot	#<ExecutionException java.util.concurrent.ExecutionException: clojure.lang.ArityException: Wrong number of args (1) passed to: sandbox$eval261$fn>
12:32	AdmiralBumbleBee	doomlord: could you use a local repl?
12:32	doomlord	heh sorry
12:34	doomlord	,( let [ [[k0 v0][k1 v1]] (map (fn[x]x) { :foo 1 :bar 2}) ] (print k0 v0 k1 v1) )
12:34	clojurebot	:foo 1 :bar 2
12:39	doomlord	multi-files: ok i appear to be able to use "load-file" from a single source ... thats enough for the minute but i dont get to make my own namespaces without figuring out lein i guess
13:10	TimMc	&(-> "\ud834\udd1e" (.substring 0 1) (.codePointCount 0 1))
13:10	lazybot	⇒ 1
13:10	muhoo_	doomlord: wat?
13:11	doomlord	i'oll figure out lein once i've written more code..
13:12	TimMc	doomlord: Don't use load-file.
13:13	muhoo	there's not much to figure out about using lein
13:13	TimMc	doomlord: https://github.com/baznex/crosscram/tree/master/src/crosscram -- just takea look at this project structure.
13:22	AtKaaZ	,(quote 1 2 3)
13:22	clojurebot	1
13:23	AtKaaZ	maybe it should err ? if the programmer expected something else other than ignoring 2 3
13:25	AtKaaZ	added to the list:) https://gist.github.com/3895312
13:29	AtKaaZ	quote is a special form ? ##(apply quote '(1 2 3))
13:29	lazybot	java.lang.RuntimeException: Unable to resolve symbol: quote in this context
13:30	antares_	AtKaaZ: yes
13:30	AtKaaZ	cool, thanks
13:32	AtKaaZ	this is cool: ##(. 1 toString)
13:32	lazybot	⇒ "1"
13:32	AtKaaZ	,(. 1/2 toString)
13:32	clojurebot	"1/2"
13:32	AtKaaZ	,(str 1/2)
13:32	clojurebot	"1/2"
13:33	AtKaaZ	hmm I notice str uses this internally: ##(. 1 (toString))
13:33	lazybot	⇒ "1"
13:33	S11001001	AtKaaZ: try on nil
13:33	AtKaaZ	,(. nil toString)
13:33	clojurebot	#<NullPointerException java.lang.NullPointerException>
13:34	AtKaaZ	that's why I should use str, good point:)
13:34	AtKaaZ	but why is str using (. 1 (toString)) ?
13:34	AtKaaZ	instead of (. 1 toString)
13:34	AtKaaZ	,(. 1 (str "toString"))
13:34	clojurebot	#<IllegalArgumentException java.lang.IllegalArgumentException: No matching method found: str for class java.lang.Long>
13:35	AtKaaZ	so I can't execute clojure code there
13:39	AtKaaZ	looks like the two forms are equivalent: http://clojure.org/java_interop#dot
13:39	AtKaaZ	(. Classname-symbol (method-symbol args)) or (. Classname-symbol method-symbol args)
13:41	AtKaaZ	except that the () is looking for method specifically, and without () is looking for field then method, as far as I can tell by the exceptions
13:41	AtKaaZ	,(. 1 (bitCount))
13:41	clojurebot	#<IllegalArgumentException java.lang.IllegalArgumentException: No matching method found: bitCount for class java.lang.Long>
13:41	AtKaaZ	(. 1 bitCount)
13:41	AtKaaZ	,(. 1 bitCount)
13:41	clojurebot	#<IllegalArgumentException java.lang.IllegalArgumentException: No matching field found: bitCount for class java.lang.Long>
13:43	AtKaaZ	that doc actually says so if I read :) i'll stfu now
14:11	dimovich	hello people...
14:12	dimovich	how hard is finding some freelance Clojure projects?
14:12	dimovich	does anyone have experience in this area?
14:15	dnolen	wow did know about :strs or :syms feature of map destructuring
14:15	dbushenko	never seen such projects
14:16	hiredman	dimovich: from what I have seen, clojure jobs tend to be around core infrastructure, not really around stuff you send out or contract
14:16	Frozenlo`	dnolen: learned it around 4 days ago. Blew my mind :P
14:16	dnolen	dimovich: I think it depends on what you do - data / analytics folks don't seem to have a hard time I think.
14:17	dimovich	thanks for the tips
14:18	pepijndevos	Is there am ebook of Lisp In Small Pieces?
14:18	hiredman	https://twitter.com/wunki/statuses/228408001655087104
14:19	nickmbailey	dnolen: woah, and now i know about :strs and :syms, thanks :)
14:22	dnolen	pepijndevos: not that I know of, though that seems like a book worth having in print.
14:22	pepijndevos	dnolen: yea, they seem rather expensive though.
14:25	dnolen	pepijndevos: in this case I think it's worth it - filled with goodies you can keep coming back to.
14:26	pepijndevos	dnolen: heh, once I have some money to spend, i'll consider it.
14:32	DaReaper5	Hi, what is the server address to this irc?
14:37	DaReaper5	I have a noobie clojure issue/question: what is the best approach for generating a report (like Jasper Reports) using clojure as the web server
14:40	hiredman	DaReaper5: I would just use jaspher reports
14:40	hiredman	DaReaper5: it's an existing java library, clojure interops very well with java, and your done
14:40	hiredman	:/
14:40	DaReaper5	I am thinking about using DynamicJasper to simplifiy things.
14:40	hiredman	you're
14:40	hiredman	whatever, I don't know anything about jasper, just that it is a java library
14:40	DaReaper5	as well*
14:50	TimMc	DaReaper5: That's not a question that can actually be answered meaningfully.
14:50	DaReaper5	Ok, to simplify my current issue: How do i enable a clojure web server to allow a client to download a file?
14:50	DaReaper5	TimMc i agree
14:51	DaReaper5	... that may not be the best wording either
14:51	TimMc	A file that is statically sitting on the server, or one that is generated dynamically by the server?
14:53	DaReaper5	an example of both would be nice. However, I need the client to be able to download dynamically generated files more so.
14:56	amalloy	DaReaper5: look up what http headers/content you need to send, and then do that
14:57	pjstadig	jasper is a disaster
14:57	pjstadig	at least it was when i used it from java or rails or something
14:57	pjstadig	maybe clojure could make it nicer
15:00	gfredericks	pjstadig: I find your claim plausible because it rhymes
15:00	DaReaper5	pjstadig i am hoping to avoid uglyness and issues by using DynamicJasper
15:00	DaReaper5	haha
15:01	TimMc	DaReaper5: It's really all about the Content-Disposition header or some such.
15:02	TimMc	Everything else is glue.
15:03	DaReaper5	ok, just fishing for possibly a link that has an example. I think i will dig into the current clojure web service code and see what we do currntly...
15:12	lnostdal	perhaps a strange question, but the clojure printer sometimes outputs what i guess must be memory addresses for some objects it prints .. will this output change based on the GC doing its work?
15:12	hyPiRion	what do you mean?
15:12	lnostdal	i suppose the GC might move objects around in memory .. and the addresses printed for the objects will then change
15:13	lnostdal	(move things around to avoid fragmentation etc.)
15:13	amalloy	&(str (lazy-seq nil))
15:13	lazybot	⇒ "clojure.lang.LazySeq@1"
15:13	hiredman	lnostdal: it is the object's hashcode
15:13	amalloy	hm. lazy-seqs don't print as their hashcode anymore?
15:13	lnostdal	hiredman: oh, ok
15:13	gfredericks	&(str (range))
15:13	hyPiRion	heh.
15:13	lazybot	Execution Timed Out!
15:13	hiredman	the default impl for Object is often the pointer, but it gets cached, since yes, the gc can move thrings around
15:13	raek	&(pr-str (lazy-seq nil))
15:13	lazybot	⇒ "()"
15:14	lnostdal	hiredman: i see .. interesting .. thank you :)
15:14	hyPiRion	,(pr (Object.))
15:14	clojurebot	#<Object java.lang.Object@427c6d2e>
15:14	lnostdal	hmm, this opens up another question; since it is cached .. is there a chance another object might end up with the same address then?
15:15	lnostdal	(since the real underlying address might be freed up while the GC moves things around)
15:15	amalloy	lnostdal: it wouldn't matter if that happened
15:15	hyPiRion	lnostdal: I'm quite sure that's a negative, the JVM would be strange if that would occur.
15:15	hyPiRion	at least not visibly.
15:15	lnostdal	ok
15:16	gfredericks	hyPiRion: based on what hiredman said I would assume the opposite
15:16	hiredman	lnostdal: there is, but it doesn't matter
15:16	hiredman	it's just a hash collision for Objects
15:17	gfredericks	I suppose it would be unreasonable to expect a runtime to assign all objects a permamently unique ID
15:19	hiredman	not a lot of utility there
15:19	gfredericks	and lots of wasted space
15:19	hiredman	that is basically what a pointer is, but the pointer can change
15:19	TimMc	Object.equals() must have some guarantee that objects can't be moved in memory in between address dereferences.
15:20	hiredman	and you don't want the hashcode changing
15:24	hiredman	TimMc: I think are confusing Object.equals() with the object identity check it uses
15:24	hiredman	you
15:24	amalloy	TimMc: i can't imagine why it would make that guarantee
15:24	TimMc	Mmm, right.
15:24	TimMc	.equals would just use ==
15:25	gfredericks	but if you have a pair of object references, and the GC decides to move the object somewhere else, it will have to change those pointers, right?
15:25	TimMc	amalloy: I don't mean the contract of .equals() in general, just Object's impl.
15:25	hiredman	or use shudder handles
15:26	gfredericks	hiredman: are those pointer pointers?
15:26	hiredman	yes
15:26	amalloy	TimMc: right. i don't see why it would make that guarantee. there's no point
15:27	gfredericks	I must be missing a detail about this
15:27	oskarth	I have a function that is defined by a macro which returns "{:foo 0}" when evaluated. How do I test this using clojure.test? "(is (= {:foo 0} BAR-FN-CALL))" fails.
15:27	amalloy	when the gc moves an object around, it also necessarily must update all references to that object in active stack frames
15:28	TimMc	Oh, hmm... are you saying that when I have a reference to an object, I actually have a pointer to the real memory location?
15:28	gfredericks	amalloy: so what we're saying is that the jvm just won't let == be executed while it's in the middle of updating the stack frame?
15:28	amalloy	so even if (a == b) goes like push a, <gc....>, push b, compare, if a has been moved after pushing, the compare still sees the current value
15:28	hiredman	uh, what do you think the word reference implies?
15:28	TimMc	That does seem to ring more true.
15:28	TimMc	hiredman: I was in fact thinking of pointers to pointers. :-/
15:29	gfredericks	amalloy: implicit in your statement is that <gc....> is atomic
15:29	hiredman	TimMc: that is sort of an unfortunate overload
15:29	TimMc	I haven't had to think about memory management details for a while.
15:29	amalloy	gfredericks: indeed, that is implied. and not necessarily true, but the jvm cooperates with your bytecode to make sure it appears atomic
15:30	hiredman	http://www.artima.com/insidejvm/ed2/jvmP.html might be a good reference (I have not read it)
15:30	gfredericks	amalloy: all of the things make sense now
15:30	hyPiRion	Huh
15:30	hyPiRion	,(vector)
15:30	clojurebot	#<ExecutionException java.util.concurrent.ExecutionException: java.lang.NoClassDefFoundError: Could not initialize class clojure.lang.RT>
15:31	Bronsa_	wtf
15:31	hiredman	sandbox is broken, it should right itself in 10 minutes
15:32	madsy	hiredman: It's your bot?
15:32	amalloy	&(vector) in the mean time
15:32	lazybot	⇒ []
15:33	gfredericks	&((reduce comp (repeat 15 vector)))
15:33	lazybot	⇒ [[[[[[[[[[[[[[[]]]]]]]]]]]]]]]
15:33	madsy	&(map {} [1 2 3] [:a :b :c])
15:33	lazybot	⇒ (:a :b :c)
15:33	madsy	&(map into [1 2 3] [:a :b :c])
15:33	lazybot	java.lang.IllegalArgumentException: Don't know how to create ISeq from: clojure.lang.Keyword
15:33	madsy	Eh, I suck. Been a while since I wrote any clojure code
15:33	gfredericks	(def identity (partial {} :foo))
15:35	hyPiRion	(def identity #(do %))
15:36	gfredericks	now accepting submissions for most surprising implementation of identity
15:37	gfredericks	&((#'and 7 8) 19)
15:37	lazybot	java.lang.ClassCastException: java.lang.Boolean cannot be cast to clojure.lang.IFn
15:37	gfredericks	&((@#'and 7 8) 19)
15:37	lazybot	java.lang.ClassCastException: java.lang.Boolean cannot be cast to clojure.lang.IFn
15:38	Bronsa	&(#'and nil nil 23)
15:38	lazybot	⇒ 23
15:39	gfredericks	oh I shoulda used partial
15:39	Bronsa	yeah
15:40	gfredericks	&(#'and nil nil 1 2 3)
15:40	lazybot	⇒ (clojure.core/let [and__3822__auto__ 1] (if and__3822__auto__ (clojure.core/and 2 3) and__3822__auto__))
15:40	llasram	(def identity (partial 'identity nil))
15:42	mattmoss	,(range 6 3 -1)
15:42	clojurebot	(6 5 4)
15:43	mattmoss	&(or 25 (range 6 3 -1))
15:43	lazybot	⇒ 25
15:43	Bronsa	,(or 1)
15:43	clojurebot	1
15:44	mattmoss	&(range (or 25 6) 3 -1)
15:44	lazybot	⇒ (25 24 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4)
15:45	AdmiralBumbleBee	mattmoss: what are you trying to do?
15:45	mattmoss	Chicago song lyric: "Twenty-five or -six to four..."
15:45	aperiodic	gfredericks: (def id (fn [x] @(reify clojure.lang.IDeref (deref [_] x))))
15:46	mjg123	What is the current status of Source Maps in ClojureScript? Do they work? Can I use them?
15:46	aperiodic	gfredericks: more "convoluted" than "surprising", i guess
15:46	dnolen	mjg123: they do not work. I'm the only person that has worked on them. very half-finished.
15:46	hyPiRion	(def identity (comp (partial apply reduce nil) reverse (partial list nil)))
15:46	dnolen	mjg123: feel free to lend a helping hand tho ;)
15:46	mattmoss	&(range (+ 20 (or 5 6)) 3 -1)
15:46	lazybot	⇒ (25 24 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4)
15:47	mattmoss	There we go.
15:47	aperiodic	international obfuscated clojure identity contest?
15:48	gfredericks	yes.
15:48	mjg123	dnolen: not sure how much help I could be - but there must be a good ref. somewhere? What's browser support like?
15:48	gfredericks	((apply comp (take 10000 (cycle [:foo (partial hash-map :foo)]))) 41)
15:48	gfredericks	&((apply comp (take 10000 (cycle [:foo (partial hash-map :foo)]))) 41)
15:48	lazybot	⇒ 41
15:49	mattmoss	o_O
15:50	hyPiRion	I assume we're going to bring in the big guns now?
15:50	gfredericks	"Ah, that's what's slowing it down. That call to identity creates five thousand array maps"
15:51	dnolen	mjg123: already supported in Chrome
15:51	dnolen	mjg123: http://www.html5rocks.com/en/tutorials/developertools/sourcemaps/
15:52	dnolen	mjg123: is the primary resource, they link to the Google SourceMap v3 doc
15:52	mjg123	dnolen: thanks.
15:52	aperiodic	dnolen: what stage are they at, exactly? last i heard you could read and write the newest kind of source map, and the main work left to do was have the cljs compiler emit the necessary info
15:52	dnolen	mjg123: http://github.com/clojure/clojurescript/tree/source-map, is the branch
15:53	dnolen	mjg123: https://github.com/clojure/clojurescript/tree/source-map/src/clj/cljs you can see source_map.clj and the source_map directory there.
15:53	dnolen	aperiodic: yes we can pass a flag so that we generate the usual sourcemap - but this needs to read back in a merged with whatever information was produced during emission.
15:53	dnolen	aperiodic: Clojure 1.5.0 now has column data, so now's a good time to wrap this up.
15:55	dnolen	http://github.com/clojure/clojurescript/blob/source-map/src/clj/cljs/compiler.clj#L207 is some basic emission of source map info when we encounter a var/binding
15:58	dnolen	so ClojureScript is pretty fast on Raspberry Pi via V8
15:58	muhoo	really? wow.
15:58	scriptor	nice
15:59	dnolen	http://twitter.com/swannodette/status/258994509755609088
15:59	muhoo	i was running jvm clj on a beagleboard for a while
15:59	Frozenlo`	dnolen: What have you done! Nooo! No I'll have to buy one!
15:59	dnolen	well according to Martin Trojer
15:59	Frozenlo`	*now
15:59	muhoo	jit compillation and repl was painful, but with the embedded-optimized jvm, an aot compiled clj project worked well
16:00	muhoo	beagleboard ships with node.js
16:00	Frozenlo`	While on the subject, anyone tested jvm clj on the raspberry?
16:00	scriptor	is anyone still selling them, actually?
16:02	Mr_Bond	there is something called cubieboard as well which looks cool
16:02	Mr_Bond	prob not long they get gigE as well
16:02	muhoo	1gb a8? nice
16:06	Mr_Bond	It's crazy how cheap small computers are getting these days
16:09	Frozenlo`	Eh. Code will eat the world :P
16:11	hyPiRion	hm
16:11	hyPiRion	(def identity #(`[~@%&](+)))
16:11	gfredericks	you win
16:12	Bronsa	wtf
16:13	gfredericks	&(#(`[~@%&](+)) "swearing in clojure")
16:13	lazybot	⇒ "swearing in clojure"
16:13	Cheiron	Hi, what is the idiomatic way to perform an action on each item in a sequence but I don't care to collect the output . doseq is one option, any other ones?
16:13	Bronsa	is github down?
16:13	gfredericks	Cheiron: I like doseq; not sure why you need anything else
16:14	thorbjornDX	Bronsa: it is for me
16:14	gfredericks	you can do it with dorun and map
16:14	aperiodic	yup
16:14	Cheiron	just expanding my knowledge :)
16:15	mattmoss	Github was fine for me just a few minutes ago... now very very slow.
16:15	hyPiRion	Bronsa: Looks down from here.
16:15	Cheiron	about the new reducers. it is a built-in hadoop feature for Clojure?
16:15	scriptor	https://status.github.com/ minor interruption apparently
16:15	Bronsa	just as i was git pushing :(
16:20	thorbjornDX	gfredericks: awesome
16:20	Bronsa	gfredericks: lol
16:20	dnolen	Cheiron: nothing to do w/ Hadoop http://clojure.com/blog/2012/05/08/reducers-a-library-and-model-for-collection-processing.html
16:21	scriptor	git bash now means something very different
16:22	TimMc	hyPiRion: Now do factorial. :-P
16:23	hyPiRion	~Factorial
16:23	clojurebot	factorial is not something you need alphanumeric characters for: https://gist.github.com/3036120
16:24	hyPiRion	Except it's down.
16:24	gfredericks	this is that scene where the local champion walks into the arcade with his groupies
16:24	hyPiRion	heh
16:24	Bronsa	youtube is down too.
16:25	gfredericks	well that's like the whole internet then
16:25	scriptor	cached: http://webcache.googleusercontent.com/search?q=cache:RmuaRRSWE90J:https://gist.github.com/3036120+&cd=1&hl=en&ct=clnk&gl=us
16:27	DaReaper5	wow that is rare
16:27	TimMc	hyPiRion: Seriously though, that's a pretty good identify fn.
16:27	gfredericks	"But watch out." he warns and leaves the arcade.
16:27	Frozenlo`	What is happening?! Is this the internet kill switch?
16:29	xeqi	how are we still talking?
16:29	TimMc	Your local IRC server is running simulations of the other participants.
16:29	TimMc	beep boop
16:30	gfredericks	does not compute
16:30	emezeske	TimMc: For a second I thought you were the real TimMc. But you gave away your robotness with that beep boop.
16:30	DaReaper5	this explains everything
16:31	hyPiRion	TimMc: If you by good mean obscure, then thanks.
16:31	TimMc	xeqi: When the netsplit is over, your client will use eventual consistency protocols to correct your logs.
16:32	DaReaper5	ok serious question: i still cannot figure out how to return dynamically created files with a clojure web service
16:32	DaReaper5	i can do static files
16:32	DaReaper5	and do not want to create temp files
16:32	TimMc	It's just output streams and headers.
16:33	DaReaper5	ok ill look more into output streams in clojure
16:33	xeqi	DaReaper5: what do you have so far?
16:33	TimMc	Pick a web lib. Write a thing that does some things when a request comes in. Write the response to the provided output writer/stream.
16:34	Mr_Bond	muhoo: I came across this as well: http://lispm.dyndns.org/ccl
16:35	DaReaper5	xeqi: I have a clojure web service created by someone else that acts as an API to a database. I am giving myself a crash course on how to generate and return reports so i can add this
16:35	DaReaper5	functionality to the web service
16:35	duck11231	DaReaper5: This is a pretty good getting started guide for compojure: http://www.booleanknot.com/blog/2012/09/18/routing-in-compojure.html
16:35	xeqi	(inc duck11231)
16:35	lazybot	⇒ 1
16:36	DaReaper5	just fyi, i am still very new to clojure but am fine with everything else involved (java and client front end stuff)
16:37	DaReaper5	thanks duck11231
16:39	TimMc	gfredericks: I can conj onto a list or vector in "swearjure", but I can't figure out how to destructure.
16:40	TimMc	&(let [existing [1 2 3]] `(~(+) ~@existing))
16:40	lazybot	⇒ (0 1 2 3)
16:40	gfredericks	TimMc: you don't have any general way to get locals do you?
16:40	gfredericks	just #() args?
16:41	TimMc	Right.
16:41	TimMc	I can't bind.
16:41	gfredericks	is flat lambda calculus turing-complete?
16:41	TimMc	Flat?
16:41	gfredericks	#() can't be nested
16:41	TimMc	Oh, heh.
16:42	Bronsa	oh, github is being DDoSed
16:42	TimMc	I guess if I can write an interpreter in swearjure then it is.
16:42	AimHere	Who would DDoS github?
16:42	TimMc	Bitbucket? :-P
16:42	xeqi	ah, beat by TimMc
16:43	TimMc	An extortionist, perhaps.
16:43	gfredericks	leave $1000 in the whole in the south tree in the park and we will let your servers go
16:43	xeqi	A company trying to lower th price on an aquihire?
16:43	gfredericks	s/whole/hole/
16:45	xeqi	oooh, someone who really really wants to make sure their project is backed up by constantly creating and pushing to new repos
16:46	Mr_Bond	is that on beanstalk?
16:47	DaReaper5	TimMc thanks again for your help
16:47	hyPiRion	,(#([`[~@(%(+))~@(%(*))]](+))(#([`[~@%&]](+)) [1 2 3] [4 5 6]))
16:47	clojurebot	[1 2 3 4 5 ...]
16:48	hyPiRion	Ahh, concatenation.
16:48	Bronsa	tfw
16:48	scriptor	...
16:48	hyPiRion	You can destructure with that TimMc
16:48	thorbjornDX	hyPiRion: what have you done
16:48	TimMc	I'm so proud.
16:49	duck11231	I'm a little scared
16:49	technomancy	great time to remind folks how easy it is to mirror to gitorious
16:50	technomancy	and use http://p.hagelb.org instead of gist =)
16:50	emezeske	Phagelborg
16:50	emezeske	That's fun to say
16:50	gfredericks	I almost had to type that out myself
16:51	gfredericks	emezeske: it NEEDs to be a flavor of bagel
16:51	hyPiRion	"You will be assimilated" - hagelborg
16:51	xeqi	I'm not suppose to be using refheap.el for my pastes?
16:52	technomancy	oh I guess
16:52	technomancy	just saying: fewer moving parts -> less downtime
16:52	technomancy	though refheap is on heroku, which is awesome
16:52	TimMc	hyPiRion: I'm all ears.
16:54	Raynes	Conj flights are booked.
16:54	Raynes	o/
16:54	emezeske	\o
16:54	gfredericks	\\\\\o
16:54	dnolen	o/
16:54	TimMc	\o/
16:54	xeqi	/o\
16:54	TimMc	\o/ /o\ \o/ /o\
16:55	TimMc	~\o/~ help I'm drowning
16:55	clojurebot	forget ~paste is gist.github.com
16:55	TimMc	clojurebot: Fickle, are you?
16:55	clojurebot	It's greek to me.
16:55	emezeske	\o/ /o/ /o\ \o\
16:55	gfredericks	clojurebot: I forgot that ~paste is gist.github.com
16:55	clojurebot	Ack. Ack.
16:55	technomancy	gfredericks: head asplode
16:56	hyPiRion	So if you want to produce e.g. (fn [[a b] c] (+ a b c)) you can just do ##(#(+(%(+))(%())(%(+()())))(#([`[~@(%(+))~(%())]](+))(#([`[~@%&]](+)) [4 7] 9)))
16:56	lazybot	⇒ 20
16:56	gfredericks	technomancy: you gotta recognize when the bots want to play; to not do so is negligent
16:57	TimMc	hyPiRion: If I give you [0 1 2 3 4] can you give me [1 2 3 4]?
16:57	bhenry	is noir still the preferred way to do web with clojure?
16:58	TimMc	bhenry: It's never been the preferred way -- it's just one way.
16:58	emezeske	bhenry: My suggestion is to use compojure, FWIW. Opinions vary.
16:58	Frozenlock	bhenry: That's what I use, but I think most use compojure directly.
16:58	TimMc	I found it great for getting something simple running very quickly.
16:58	hyPiRion	TimMc: Yeah, probably. Let me massage it.
16:59	TimMc	hyPiRion: The best I can figure is a recusive procedure that takes element 1 and recurs to get element 2..n, then prepends.
17:00	TimMc	How do you tell when it is done recurring?
17:00	ToBeReplaced	bhenry: I had never done web development before, and I found that Noir hid too many things from me and insulated me from decisions I would rather arrive at myself >> instead went with compojure/ring/enlive >> noir is compojure/ring/hiccup with higher-level functions
17:01	TimMc	hyPiRion: Vectors don't support the 2-arity get: ##([1 2] 5 :nope)
17:01	lazybot	clojure.lang.ArityException: Wrong number of args (2) passed to: PersistentVector
17:01	bhenry	maybe i'll just go back to mustache? did anyone find major problems with that library by cgrande?
17:02	ToBeReplaced	i thought it wasn't supported anymore
17:02	cemerick	bhenry: mustache isn't very widely used at all
17:02	bhenry	dag. i liked it better than compojure
17:02	TimMc	hyPiRion: Oh, but ##({[] :empty} [] :not-empty)
17:02	lazybot	⇒ :empty
17:03	TimMc	Clearly the next step is to write a swearjure assembler.
17:03	ToBeReplaced	i think compojure/ring are pretty solidly established for at least a little while... whether you use noir on top of that is preference
17:03	hyPiRion	TimMc: Surely.
17:03	cemerick	bhenry: FWIW, if you've not looked at compojure in a while, you may find it more to your liking. I remember there was someone else in here that loved mustache, but only vs. ye old-time compojure.
17:03	TimMc	curje
17:04	hyPiRion	TimMc: As a vector or as varargs? ##(#({%`[~@%&]}%) 0 1 2 3 4)
17:04	lazybot	⇒ [1 2 3 4]
17:04	bhenry	cemerick: that could be. i haven't started a new web project in clojure for some time. i'll look at it again.
17:06	hyPiRion	Hum, it shouldn't be that far for a vector version
17:14	TimMc	hyPiRion: Oh hey, that's right!
17:14	TimMc	hyPiRion: For some reason I was thinking that %& was all-args, not rest-args. >_<
17:16	TimMc	Now you just need apply. :-P
17:20	hyPiRion	yeah dangit
17:20	TimMc	which can't be had
17:30	TimMc	OK, it can be done, but it would be longer than factorial.
17:30	TimMc	It needs a helper fn and two if clauses.
17:31	TimMc	And my code's done compiling, so I'll have to do it another time. :-P
17:31	hyPiRion	oh dog.
17:32	hyPiRion	Alright, new identity function time
17:32	TimMc	hyPiRion: Basically, take the factorial code and add an initial fn that takes the input and concats [1 []] to the arg list, then passes that on to the real recursive fn.
17:32	hyPiRion	(def identity #(`%%%))
17:33	TimMc	wat
17:33	ToBeReplaced	it seems like there are some contradictions between the Library Coding Standards and the examples in The Joy Of Clojure. Generally, which should I assume is more up to date?
17:34	hyPiRion	TimMc: Bleh, I was kind of hoping one could use some tricks with backquoting.
17:34	TimMc	&('foo 5 5)
17:34	lazybot	⇒ 5
17:34	TimMc	_nice_
17:34	hyPiRion	(#(identity `%) :foo)
17:34	hyPiRion	,(#(identity `%) :foo)
17:34	clojurebot	p1__27__28__auto__
17:35	ToBeReplaced	examples: TJoC uses (:key collection) to extract an element instead of (collection :key), and the very first example DSL in the book has (:use [clojure.string :as str :only []]) which I would have thought is now meant to be (:require [clojure.string :as str])
17:36	TimMc	ToBeReplaced: The issue of key vs coll as fn is mostly one of intended semantics.
17:37	TimMc	E.g. if your associative coll could be nil, you'd use the key as the fn for sure.
17:37	gfredericks	ToBeReplaced: and the second point sounds correct
17:37	TimMc	&(:foo nil)
17:37	lazybot	⇒ nil
17:37	TimMc	If the key is unknown, the associative coll would be used as the fn.
17:37	gfredericks	or if the key isn't a keyword
17:38	ToBeReplaced	i thought in that case you would use (get collection :key)
17:38	gfredericks	also that
17:38	gfredericks	at the very least maps as functions are useful instead of #(get collection %)
17:39	ToBeReplaced	ah, right
17:41	ToBeReplaced	maybe i'll just write down all of the things that come up that are confusing idiomatically as a newcomer and forward it along to the authors
17:42	TimMc	hyPiRion: I know how to stop the recursion!
17:42	ToBeReplaced	or maybe i just gotta get used to people using :use anyway :)
17:43	TimMc	If your input is [0 1 2 3 4] and you have collected [1 2 3 4] as rest, then you can check if (= [0 1 2 3 4] `[0 ~@[1 2 3 4]]).
17:43	aperiodic	wait, symbols are functions?
17:43	aperiodic	this is news to me
17:43	gfredericks	aperiodic: just like keywords
17:44	technomancy	aperiodic: a feature never before invoked on purpose
17:44	TimMc	&(:key 5 6)
17:44	lazybot	⇒ 6
17:44	hyPiRion	TimMc: Sweet
17:44	gfredericks	aperiodic: it's a secret of clojure which, once you know it, causes you to never again write bad code
17:44	aperiodic	oh sweet
17:45	aperiodic	when would you want to actually use that, if you're not trying to turn clojure into brainfuck?
17:45	technomancy	afaict it's only used for obfuscation purposes
17:45	gfredericks	symbols-as-keys might come up in macros
17:46	gfredericks	Due to a copyright claim by Microsoft this code sample has been obfuscated.
17:47	aperiodic	but you'd need to quote or unquote-quote the symbol in the macro either way
17:47	hyPiRion	It may be used to lookup stuff in a symbol table. (cue bad joke)
17:48	aperiodic	although i guess you might have fully qualified symbols in the collection
17:48	emezeske	~badum
17:48	clojurebot	Huh?
17:48	gfredericks	hyPiRion: a symbol, a keyword, and a string walk into a bar
17:48	emezeske	~rimshot
17:48	clojurebot	Badum, tish
17:49	gfredericks	I think it's totally conceivable for a complicated macro to have a map from plain symbols to some kind of values, and to want to look up symbols in that map
17:50	hyPiRion	gfredericks: The string isn't IFn though.
17:50	gfredericks	hyPiRion: that's the punchline!
17:50	gfredericks	now to figure out the middle of the joke...
17:51	hyPiRion	heh, well, the keyword is the only one to guaranteed have a colon.
17:51	gfredericks	,(keyword ":://")
17:51	clojurebot	::://
17:52	hyPiRion	or multiple.
17:53	gfredericks	Three stringoids walked into a bar. Two fell out. Who was left?
17:54	aperiodic	i don't know, gfredericks. who?
17:54	amalloy	gfredericks: even if a macro does have a map from symbols to something, it's likely to be more natural to use the map as a function than the symbol
17:55	amalloy	eg, `(do ~@(map sym->value symbols))
17:55	gfredericks	aperiodic: one of them!
17:56	hiredman	I like to tell jokes that ask the question "..., what's left?" and then tell people "no, that way" and point left after they try and guess
17:57	aperiodic	gfredericks: you like telling shaggy dog stories, don't you?
17:59	gfredericks	aperiodic: I couldn't orally perform such a thing; my straight faces have a halflife of ~5sec
17:59	aperiodic	that's probably a good thing
18:00	aperiodic	i once very nearly got thrown off of a moving bus for telling a half-hour-long version of the "you're not a monk" joke
18:00	TimMc	heh
18:03	Pip	Is Clojure cool shit?
18:03	TimMc	Very cool shit.
18:03	TimMc	But not the coolest.
18:04	TimMc	I don't know, what are you trying to ask?
18:04	thmzlt	Pip: clojure is just legacy bullshit (https://twitter.com/ryah/status/258634435161899009)
18:05	TimMc	aperiodic: I once heard a 20-30 minute shaggy dog story (something to do with a monestary, so maybe the same one?) and the "performer" kept noting digressions that he wouldn't make.
18:05	TimMc	It was pretty impressive.
18:05	gfredericks	(->> shit (sort-by coolness) (take 3))
18:05	Pip	lol
18:05	Pip	That's really impressive
18:05	aperiodic	TimMc: yeah, it sounds like the same one
18:05	technomancy	yeah you really need to get on a real non-legacy language like C++
18:06	technomancy	everything else just keeps confusing everyone
18:06	emezeske	Is this ryah guy a troll?
18:07	hiredman	yes
18:07	technomancy	obvs
18:07	aperiodic	TimMc: i'm impressed; 30 minutes or so is the longest i've been able to stretch that
18:07	TimMc	It might have only been 15 minutes, for all I know -- it felt like 30. :-P
18:07	technomancy	emezeske: so much so that it's actually kind of impressive
18:08	emezeske	technomancy: :)
18:09	thmzlt	technomancy: touche
18:10	dnolen	emezeske: normally wouldn't get any attention but he's the creator of Node.js
18:10	technomancy	dnolen: https://mobile.twitter.com/robey/status/258676358086029312 =)
18:10	emezeske	dnolen: Ahhh, I thought that name sounded familiar
18:14	thmzlt	too bad I blocked HN here, would love to see what kind of discussion is going on there today
18:15	emezeske	thmzlt: None, AFAICT.
18:15	technomancy	you can catch up at http://shitryandahlsays.tumblr.com/
18:15	technomancy	I think this is below even the standards of HN
18:16	emezeske	technomancy: Even below proggit's radar.
18:16	emezeske	Impressive
18:16	scriptor	oh cool, clojure needs to be euthanized
18:16	scriptor	at least that'll be quick and painless
18:18	amalloy	it's aught, isn't it?
18:19	emezeske	Can't say I've ever typed it out before :) aught looks better.
18:19	hiredman	~google aught
18:19	ToBeReplaced	wow, that guy's awesome... great tumblr
18:19	clojurebot	First, out of 199000 results is:
18:19	clojurebot	Aught - Definition and More from the Free Merriam-Webster Dictionary
18:19	clojurebot	http://www.merriam-webster.com/dictionary/aught
18:20	hiredman	~google ought
18:20	clojurebot	First, out of 8420000 results is:
18:20	clojurebot	Ought - Definition and More from the Free Merriam-Webster Dictionary
18:20	clojurebot	http://www.merriam-webster.com/dictionary/ought
18:21	hiredman	oh, I see, they are both words, that through be off for a second, I assumed aught being correct (in this case) would have more hits
18:21	emezeske	It sounds like "aught" is a bastardization of "naught"
18:21	emezeske	Which makes sense, as "naught" means nothing (zero)
18:21	aperiodic	aught prounoun 1) anything 2) all, everything; noun 1) zero 2) nothing
18:21	aperiodic	i <3 english
18:21	emezeske	Although "nought" is apparently a valid variant of "naught"
18:21	hiredman	http://i.imgur.com/OfN6N.jpg
18:22	aperiodic	^ truth
18:22	emezeske	hiredman: hahahaha
18:22	aperiodic	that's why it's so much fun
18:57	callen	a brief thought, was wondering how other people felt: the emphasis on functions like partition, particular in things like 4clj and clojure-koans feels a bit contrived and cutesy. What do you think?
19:00	dnolen	callen: perhaps, but partition is pretty useful when writing macros that involving the binding forms.
19:00	emezeske	callen: A bit contrived? Things like 4clj and clojure-koans are completely contrived, that's the point.
19:00	emezeske	callen: If you want non-contrived, give yourself a project and build it.
19:01	technomancy	non-contrived exercises don't scale
19:01	technomancy	actually "non-contrived exercise" is probably self-contradictory
19:01	emezeske	technomancy: Exactly
19:03	TimMc	callen: OK, we can write some nice little exercises involving JDBC or shelling out.
19:43	doomlord	the return value of a map is a lazy sequence? i'm trying to concatenate the results of maps on vectors. the return value prints as (a b)(c d)(e f) ... it doesn't seem to let me concat to get (a b c d e f) wheras concat does work on [a b][c d][e f]. Whats the solution?
19:43	doomlord	do i need to turn the result of the map into an actual vector first or something
19:43	emezeske	doomlord: Maybe use mapcat?
19:44	emezeske	,(mapcat (fn [x] [x x]) [1 2 3 4])
19:44	clojurebot	(1 1 2 2 3 ...)
19:44	doomlord	didn't know that existed. looks like what i want. I'm surprised its a special case ; is there another way to do this from primatives
19:44	gfredericks	doomlord: almost certainly your issue is not about the difference between vectors and seqs
19:44	emezeske	,(apply concat (map (fn [x] [x x]) [1 2 3 4]))
19:44	clojurebot	(1 1 2 2 3 ...)
19:45	doomlord	mapcat works ; but i'd like to know what went wrong before
19:45	emezeske	see above ^
19:45	emezeske	mapcat is not really a special case, it's just a convenience function because that's so common
19:45	emezeske	I'm guessing you left off the apply before?
19:46	doomlord	ok apply works too. yes i left of the apply. why is the apply needed
19:46	gfredericks	concat is variadic
19:47	doomlord	(i knew of "apply" from common-lisp being apply a function to an arg array)
19:47	gfredericks	same thing, it sounds like
19:47	doomlord	ok silly me i see it now.
19:47	emezeske	,(concat [1 2] [3 4] [5 6])
19:47	clojurebot	(1 2 3 4 5 ...)
19:48	emezeske	,(apply concat [[1 2] [3 4] [5 6]])
19:48	clojurebot	(1 2 3 4 5 ...)
19:48	doomlord	i was doing (concat [ other lists]). idiot.
19:48	gfredericks	a side effect is that concat with one arg does nothing, which is confusing
19:48	emezeske	doomlord: Not idiot! Just learning.
19:49	doomlord	the repl is an awesome tool for breaking down what the expressions do, i should have been able to figure that out :)
19:54	callen	TimMc: JDBC? Who do you think I am, Mark Pilgrim?
19:56	callen	my deeper point is that using things like partition isn't the hard part, although knowing the idioms is useful.
20:01	amalloy	if 4clojure could effectively teach whatever you think is "the hard part" of programming, we'd have venture capital by now
20:03	emezeske	amalloy: Please make 4clojure download a decade of experience into my brain, matrix style.
20:03	emezeske	"I know computer science."
20:04	doomlord	foreclosure ?!
20:04	TimMc	doomlord: No, 4clojure!
20:04	doomlord	:)
20:05	doomlord	"i know economic collapse"
20:05	TimMc	haha
20:05	TimMc	callen: Are you envisioning purely algorithmic stuff?
20:06	emezeske	aperiodic: After "Show me," I prove that there is no halting problem after all, and compress something below the threshold of its Kolgomorov complexity.
20:06	aperiodic	as Math crumbles around you like so many cathode-green glyphs
20:07	emezeske	:)
20:07	aperiodic	i'd watch that movie
20:07	TimMc	What if I told you...
20:08	TimMc	...that you can parse XML with one regular expression?
20:10	emezeske	*irregular expression
20:16	TimMc	(I have used a regular expression to parse HTML, but I also had a stack, so it was OK.)
20:19	Raynes	I once parsed a regular expression with a regular expression.
20:20	TimMc	Is that...
20:20	TimMc	No, you can't. OK.
20:21	TimMc	Whoa! I think this is right: Regular expressions are a context-free language at minimum, and CFGs are regular at minimum.
20:21	ivan	too often I've replaced >< with >\n< and used a state machine to pull something out
20:21	ivan	maybe because I was working in a language with a vulnerable XML parser
20:24	aperiodic	TimMc: doesn't the second statement follow from the equivalences between regular langs & DFAs and CFGs & stack machines (which are just DFAs with memory)?
20:24	ohpauleez	By the code of the internet we must tell the obligatory RE joke:
20:25	ohpauleez	You have a problem and you think, "I'll solve this with a regular expression." Now you have two problems.
20:25	gfredericks	aperiodic: those things aren't equivalent
20:25	gfredericks	(RE == DFA) != (CFG == SM)
20:25	gfredericks	or maybe you meant that
20:26	gfredericks	now that I distinguish 'and' and '&'
20:26	aperiodic	yeah, two pairs of equivalances
20:29	amalloy	i don't think i understand TimMc's "whoa" at all. though i'm totally in favor of his "what if i told you"
20:30	aperiodic	amalloy: i think TimMC is positing that anything that can describe a regular language in the way regular expressions do (what does that mean?) is necessarily a CFG
20:30	TimMc	aperiodic: RE's are a way of describing regular languages, but I am claiming the set of RE's is non-regular, and is in fact context-free.
20:30	TimMc	Nope, not saying anything about the relationship between regular langs and context-free langs.
20:30	gfredericks	are non-balanced parens legal?
20:31	aperiodic	oh, ok
20:31	gfredericks	,#"(6"
20:31	clojurebot	#<ExecutionException java.util.concurrent.ExecutionException: java.lang.RuntimeException: java.lang.ExceptionInInitializerError>
20:31	gfredericks	&#"(6"
20:31	lazybot	java.util.regex.PatternSyntaxException: Unclosed group near index 2(6 ^
20:31	TimMc	,#"\Q(\E"
20:31	clojurebot	#"\Q(\E"
20:31	TimMc	10 days later
20:31	gfredericks	TimMc: is there anything else non-regular aside from the parens?
20:32	TimMc	gfredericks: Not that I'm aware of.
20:32	gfredericks	of course programming language regular expressions do not describe only regular languages
20:32	amalloy	TimMc: to be clear: you're saying that to recognize whether something is a valid regex, you need a context-free grammar?
20:33	TimMc	amalloy: Yes.
20:33	gfredericks	that's been my assumption too
20:33	TimMc	Recognizing ((((())))) requires a stack.
20:33	amalloy	i wouldn't be astonished to discover you need a context-sensitive grammer
20:34	gfredericks	&#"()\4"
20:34	lazybot	⇒ #"()\4"
20:34	gfredericks	^ is that a back-reference?
20:34	TimMc	I can never remember what each re parser uses for back references.
20:34	clojurebot	That's not my job.
20:34	TimMc	Mine neither.
20:35	gfredericks	in any case it's the back-references that allow the regexes to describe non-regular languages
20:35	amalloy	&(re-seq #"(a)\4" "aba\\4")
20:35	lazybot	⇒ nil
20:35	amalloy	&(re-seq #"(a)\1" "aa\\4")
20:35	lazybot	⇒ (["aa" "a"])
20:36	gfredericks	so java at least allows out-of-bounds backrefs
20:36	amalloy	i think #"(a)\4" is a regex recognizing nothing, because the \4 backref can't ever be matched
20:44	TimMc	amalloy: I never learned CSGs. :-/
20:45	TimMc	Anyway, CFGs would be fine for REs.
20:51	Hodapp	damn, why do I always code myself into corners where I run into problems like "Well, this is easily solved if I just MAKE A MAP REFER TO ITSELF RECURSIVELY!"?
20:55	emezeske	Hodapp: Because cyclic graphs are useful? :)
20:58	Hodapp	emezeske: Out of curiosity, can I make a map refer to itself sanely?
20:59	emezeske	An immutable map technically can't have a reference (i.e. a pointer) to itself
20:59	emezeske	But...
20:59	emezeske	It could contain a sort of path to a submap
20:59	emezeske	e.g. {:a {:b {:c [:a :b]}}
20:59	gfredericks	maps of refs of maps of refs of maps of atoms of futures to themselves!
21:00	TimMc	emezeske: Unless you're in Haskell.
21:00	Hodapp	since I know Clojure allows lazy sequences, but lazy maps are another matter I suppose
21:00	emezeske	TimMc: Is that so?
21:00	TimMc	I think so.
21:00	gfredericks	Hodapp: https://github.com/fredericksgary/lazy-map
21:00	TimMc	some kind of delayed binding
21:00	emezeske	TimMc: I guess if you know where the map will be before you create it?
21:00	emezeske	Ah
21:01	TimMc	http://www.haskell.org/haskellwiki/Tying_the_Knot
21:01	emezeske	TimMc: Thanks, looks like a neat read!
21:01	TimMc	And I believe there are some Schemes wherein a future/promise/delay/what-have-you is actually replaced by its referent when realized.
21:02	Hodapp	the situation I've found myself in is that I'm trying to make an API that mimics what is available in CF3 (Context Free version 3), in which you have definitions like "blah { ... }" which include instantiations of other definitions, and it's normal for a definition to refer to itself recursively
21:02	gfredericks	&(let [p (promise) coll (lazy-seq [@p]), m {:selfs coll}] (deliver p m) m)
21:02	lazybot	java.lang.StackOverflowError
21:02	hiredman	I wlll trot out letfn-
21:03	Hodapp	hiredman: to me?
21:03	hiredman	https://gist.github.com/1179073
21:04	Hodapp	in other words, CF3 lets you express a scene graph that can be, er, recursive.
21:04	Hodapp	and rendering simply bails out when things become too small
21:05	Hodapp	or your computer catches on fire.
21:11	amalloy	hiredman: i can't tell, is letfn- just a clojure-side implementation of letfn, so that it wouldn't have to be a compiler special?
21:11	TimMc	&(let [p (promise) s (lazy-seq (cons @p nil))] (deliver p s) (identical? s (first s)))
21:11	lazybot	⇒ true
21:12	hiredman	amalloy: yes
21:12	TimMc	Oh hah, I ended up with the same solution.
21:12	hiredman	all immutable like
21:19	nightfly_	If I use clojail to blacklist clojure.java.shell/sh does that mean any function that is called within the sandbox that contains calls to clojure.java.shell/sh will also fail?
21:20	TimMc	I don't believe so. Raynes?
21:21	Raynes	nightfly_: No. We can't look inside of functions like that. It will block any macros that call it though, since they can be expanded.
21:21	xeqi	nightfly_: no, it just prevents forms with that
21:22	nightfly_	Raynes: Cool, thank you
21:22	Raynes	You could, theoretically, write a tester function that looks up the source and looks for it now that I think about it. That'd be pretty slow though, since you'd have to run that source code through the sandbox too.
21:23	cemerick	Raynes: can't you hook the blacklisted var to throw when it's called?
21:23	Raynes	That is an option to.
21:23	Raynes	too*
21:23	Raynes	It still isn't great though.
21:23	cemerick	sounds like a future feature
21:24	Raynes	I don't think I'd do it. It's too unreliable.
21:24	Raynes	A function might refer to the actual function and not the var.
21:24	quidnunc	Are Ubuntu packages sufficiently current?
21:24	Raynes	It'd be easy enough to break, seems like.
21:24	Raynes	You'll want leiningen and not any 'clojure' packages.
21:25	Raynes	I'd suggest not using Ubuntu packages for anything Clojure related though.
21:25	TimMc	quidnunc: And don't get leiningen via APT.
21:25	cemerick	Raynes: as long as the blacklist was "applied" prior to the untrusted code being loaded, that shouldn't matter
21:26	Raynes	cemerick: Fair enough. It'd be a nice feature. It'd have to be optional though, because clojail sandboxes aren't meant to be destructive by default.
21:26	xeqi	it'd be inconsistent with blacklisting namespaces or java calls
21:26	xeqi	quidnunc: leiningen.org since no one else plugged it
21:38	Sgeo	antares... isn't here
21:38	Sgeo	darnit
21:56	ivan	,(persistent! (let [x (transient {})] (conj! x [1 (persistent! x)])))
21:56	clojurebot	#<RuntimeException java.lang.RuntimeException: java.lang.IllegalAccessError: Transient used after persistent! call>
21:59	S11001001	ivan: cheeky but ultimately futile attempt to tie a knot
22:00	S11001001	ivan: gotta exploit CLJ-893 for that nonsense
22:01	ivan	hah, neat
22:08	gfredericks	S11001001: lazy-seq + promise
22:09	ivan	&(let [a (to-array [nil]), v (vec a)] (do (aset a 0 v) (v 0)))
22:09	lazybot	java.lang.StackOverflowError
22:10	gfredericks	yes you can also do it with mutable data structures
22:10	gfredericks	&(let [a (to-array [nil])] (aset a 0 a) a)
22:10	lazybot	⇒ #<Object[] [Ljava.lang.Object;@4d031d>
22:10	S11001001	gfredericks: that's patently mutating an immutable data structure
22:11	gfredericks	S11001001: lazy-seq and promise is?
22:11	S11001001	gfredericks: no, the 893 method
22:11	gfredericks	oh okay
22:11	S11001001	gfredericks: but sure, use mutables in documented ways and you get knots
22:11	gfredericks	not so much mutable as lazy
22:12	gfredericks	it's like half-mutable
22:12	ivan	I was kind of hoping immutables were actually immutable
22:12	S11001001	ivan: me too
22:12	TimMc	gfredericks: mutate-once
22:12	S11001001	mutable/immutable is a dichotomy
22:12	TimMc	mutate-before-read
22:12	gfredericks	TimMc: that still sort of implies it had two different values
22:13	gfredericks	S11001001: I claim lazy-seqs do not obviously fall on either side, intuitively. So to make it a dichotomy you'll have to contrive some definition.
22:13	callen	TimMc: far from it @ envisioning purely algorithmic stuff
22:13	callen	TimMc: more like, "how do I cope with the fact that I am surrounded on all sides by terrible Java libraries?"
22:13	ivan	you can blow up just about anything with the recursive vector constructed above
22:19	ivan	if someone really likes the current behavior maybe it could be hidden behind an unsafe-alias true binding
22:21	ivan	(do Java arrays have some internal version number?)
22:21	ivan	eh, who am I kidding, there's no way
22:22	gfredericks	what would that be used for?
22:22	ivan	throwing an exception if it's mutated and you didn't expect it to be mutated
22:23	gfredericks	but this is java. not expecting your array to be mutated is like not expecting your program to be run.
22:23	ivan	right :)
22:23	gfredericks	....and that would be pretty bad for perf
22:23	Hodapp	gfredericks: If you get away without having your program run at all, I'd say that's a fairly substantial optimization
22:24	Khaoz	Hi all.
22:24	TimMc	gfredericks: Maybe you could use it to detect if your array hasn't been mutated recently, and throw in that case.
22:24	TimMc	Clearly, something must be wrong, right?
22:24	gfredericks	lol
22:24	ivan	Warning: code has no side effects
22:25	TimMc	cons¡
22:26	Khaoz	I'm trying to setup nrelp on my emacs, but without success
22:26	Khaoz	i have lein 2 installed using brew and it's under /usr/local/bin
22:26	aperiodic	TimMc: lol!
22:26	Khaoz	i also have this in my init.el: (add-to-list 'exec-path "/usr/local/bin")
22:26	TimMc	aperiodic: A destructive lol.
22:27	Khaoz	and also tried to setup nrepl-lein-command to the full path of lein (/usr/local/bin/lein)
22:27	Khaoz	but ... error in process sentinel: Could not start nREPL server: /bin/bash: lein: command not found
22:27	Khaoz	:(
22:28	technomancy	Khaoz: there are other ways to fix the path fiasco on OS X
22:28	technomancy	I think maybe the clojure-mode doc/ directory covers that?
22:28	Khaoz	humm
22:29	Khaoz	technomancy: thank. I will take a look
22:30	Khaoz	Emacs Setup on OS X - todo
22:31	technomancy	supposedly the "proper" fix involves editing plist xml files
22:31	technomancy	so even if I knew the answer I couldn't tell you it with a clean conscience
22:32	Khaoz	no problem
22:33	callen	Khaoz: I've used nrepl quite a bit on my mac.
22:33	callen	Khaoz: and I have a rather elaborate emacs setup that I use between Linux and Macs regularly.
22:33	callen	Khaoz: can I help?
22:33	Khaoz	but did you installed it using homebrew ?
22:33	callen	did I installed what?
22:33	Khaoz	lein
22:33	callen	there are like, 10 possible components you could be talking about here.
22:33	callen	why the hell would you do that?
22:33	callen	don't do that.
22:33	Khaoz	humm
22:34	Khaoz	it's a good start
22:34	callen	technomancy went to the trouble of making a nice magical script.
22:34	callen	go use that.
22:34	Khaoz	i have everything installed using homebrew
22:34	Khaoz	clojure, lein..
22:34	technomancy	I would say use whatever's consistent with the rest of your system, except homebrew is kind of clown shoes
22:35	callen	Khaoz: stop using homebrew for stuff like that, it's not like apt or yum.
22:35	technomancy	definitely don't need to bother installing clojure
22:35	callen	Khaoz: reserve it for libraries and random odds and ends.
22:35	callen	Khaoz: and like technomancy said...if you're using lein, you don't gotta install clojure.
22:35	technomancy	OTOH installing stuff by hand is kind of skooky
22:35	callen	lein just treats "clojure the language" as yet another jar. because it is.
22:35	callen	technomancy: okay, I gotta know. skooky?
22:35	technomancy	but in this case it's probably the lesser of two evils
22:35	callen	it's not really a big deal.
22:35	technomancy	callen: sketchy; wacky; silly
22:36	callen	fair.
22:36	cemerick	lol @ clown shoes
22:36	callen	it really is clown shoes though.
22:36	technomancy	every time I have to install something by hand for a reason other than "the version in apt/nix isn't new enough" I feel like a Windows user
22:36	callen	to install clojure on my mac, I just make certain the java install is sane, install my emacs environment, then install lein.
22:36	callen	Everything from there is managed by lein.
22:36	Khaoz	install things by hand is not a issue for someone that have used slackware for 4 years :)
22:36	cemerick	package management doesn't exist on the mac, period.
22:36	technomancy	cemerick: but but but... macports =D
22:36	Khaoz	i will review what i have made and try another method
22:36	cemerick	~guards
22:36	clojurebot	SEIZE HIM!
22:37	technomancy	clojurebot: macports?
22:37	clojurebot	Huh?
22:37	callen	technomancy: macports is even worse.
22:37	technomancy	clojurebot: macports is not a package manager, it's a satire about package management.
22:37	clojurebot	Ack. Ack.
22:37	callen	fink is worse still.
22:37	callen	lol ^^
22:37	cemerick	oh gawd fink
22:37	technomancy	hahaha
22:37	callen	Khaoz: get your emacs sane, check your java install, then just install lein as the plain vanilla script.
22:37	technomancy	cemerick: supposedly you can get nix working on OS X
22:37	callen	Khaoz: then just go from there.
22:37	cemerick	I was all about fink for about 3 weeks
22:37	cemerick	nix?
22:37	callen	Khaoz: stop futzing around with homebrew for things its awful at.
22:37	clojurebot	nix is a purely functional package manager exhibiting many similar characteristics to Clojure's persistent data structures or git commit trees: http://nixos.org/nix/
22:38	callen	gotta say. Not a fan.
22:38	cemerick	oh, whatever
22:38	Khaoz	callen: ok. Thanks! :)
22:38	xeqi	leiningen isn't in the mac app store?
22:38	callen	...
22:38	xeqi	missing out on an oppertunity there
22:38	callen	technomancy: start charging $4.99 for leiningen on the app store.
22:38	callen	technomancy: you'll make...hundreds.
22:39	cemerick	I'll bet cygwin has a better package manager than bew.
22:39	cemerick	brew*
22:39	quidnunc	packaging has become mental again.
22:39	callen	cemerick: it does, they at least have the benefit of binaries.
22:39	quidnunc	After a few years of sanity
22:39	technomancy	callen: would that even be legal?
22:39	technomancy	I kind of hope not
22:39	callen	technomancy: doesn't it belong to you?
22:40	callen	technomancy: leiningen that is.
22:40	xeqi	just parts of it
22:40	cemerick	It's all EPL; anyone could package it up and sell it.
22:40	callen	oh wow, time to go make some beer money
22:40	cemerick	Except, it requires the JVM, so the app store is out anyway.
22:40	callen	cemerick: nah you could wrap it up in a GUI.
22:41	aperiodic	hot on the mac app store: leiningen wizard
22:41	cemerick	That downloads and exec's the JVM? Probably not.
22:41	Hodapp	Cygwin's package manager appears to basically be "run the installer again and add packages"
22:41	callen	cemerick: is that a bet?
22:41	callen	technomancy: will you get offended if I start selling a GUI wizard to leiningen for $4.99?
22:41	callen	just to make cemerick mad.
22:41	cemerick	hardly :-P Just what I remember of the app store policy w.r.t. the JVM
22:42	callen	Mac App Store is looser than iOS App Store.
22:42	cemerick	shrug
22:42	callen	tons of non-cocoa+obj-c stuff on there.
22:42	tali713	callen: that sounds like a great idea.
22:42	doomlord	are there converters to compile jvm code -> native
22:42	doomlord	(some subset of )
22:42	callen	tali713: what? the graphical leiningen wizard?
22:43	tali713	yes.
22:43	cemerick	Hodapp: you can't run apt or something with a different set up cygwin-only repos?
22:43	callen	tali713: you really don't want to encourage my craziness.
22:43	Hodapp	doomlord: Yes.
22:43	tali713	I enjoy good jokes made into tools.
22:43	callen	I enjoy tools that make good jokes.
22:45	Hodapp	doomlord: Look up NestedVM.
22:45	cemerick	Apropos of nothing, I predict seeing more Surface Pros at confs in 2013 than laptops running linux.
22:46	callen	cemerick: wat.
22:46	cemerick	callen: heard it here first. :-P
22:46	aperiodic	it seems like that would be heavily dependent on which conferences exactly you mean
22:47	callen	everybody forgets the incorrect predictions...
22:47	cemerick	eh, the ones I care about. :-) Strange Loop, Conj, etc.
22:47	callen	no need to buy a surface, the laptop was perfected long ago thanks to the Thinkpad.
22:47	tali713	i don't, but i should keep a journal for better verification.
22:49	aperiodic	huh. seems bold, then. but maybe i'm just skewed because i've only been to one conference, and hardly saw technomancy away from his laptop the entire time
22:49	technomancy	callen: the app store is an offence to all that is good about computing
22:49	technomancy	it would grieve me deeply to see my code on it
22:50	callen	technomancy: I feel similarly. I'm sorry for even evoking such agony.
22:50	tali713	technomancy: what about a gui that just uses exec to run lein?
22:50	cemerick	Hasn't Ubuntu already rebranded its apt GUI as an "app store"?
22:50	technomancy	callen: most of my non-clojure code is copylefted in such a way as to prevent that
22:50	callen	technomancy: did you know I use leiningen as the "good child" when larting other people over the head for their package/project management tools?
22:51	technomancy	EPL is still copylefted, but I'm not sure it's enough to prevent that kind of abuse
22:51	technomancy	callen: haha; nice
22:51	technomancy	tali713: well I have no idea as to its legality, but it would still bum me out
22:52	callen	technomancy: most recent example was in a github issue I filed with cabal. Their shit broke for the 4000th time and I did a compare/contrast with what it was like to solve a similar problem in a project.clj
22:52	tali713	technomancy: oh, well, I was more curious than anything. I pretty much loathe the appstore.
22:52	callen	technomancy: for people who are obsessed with reliability/stuff-not-breaking, the haskell community has an incredibly unreliable package manager.
22:52	technomancy	cemerick: yeah, that's a big part of why I've moved away from ubuntu. the labeling on the installer tricked me into installing flash.
22:53	technomancy	cemerick: it's all "do you want other packages that may have distribution issues, such as mp3 codecs?" and I'm like "yes, screw US patent law"
22:53	quidnunc	technomancy: What do you use now?
22:53	technomancy	and I was in for a nasty shock
22:53	technomancy	quidnunc: debian stable plus nix
22:53	quidnunc	technomancy: Why stable?
22:53	callen	technomancy: you definitely don't do web dev if you don't want flash :P
22:53	technomancy	quidnunc: because it never ever breaks
22:53	technomancy	callen: thank goodness =)
22:54	technomancy	quidnunc: and because nix gets me everything I need that's newer and not in stable
22:54	callen	technomancy: not that proper web devs use flash for anything serious, but you still have to be running it.
22:54	technomancy	and nix never ever breaks either, but for a different reason
22:54	callen	last time I used flash was for a websockets fallback.
22:54	cemerick	I don't think I've deployed an flv for ~4 years
22:54	technomancy	at least it never breaks in a nontrivial way
22:55	quidnunc	Was nix written by the guy who wrote ion? I remember he proposed this type of package manager many years ago
22:55	Hodapp	what is nix?
22:55	quidnunc	http://nixos.org/nix/
22:55	callen	nix?
22:55	clojurebot	nix is a purely functional package manager exhibiting many similar characteristics to Clojure's persistent data structures or git commit trees: http://nixos.org/nix/
22:55	callen	Hodapp: ^^
22:56	technomancy	nix is amazing
22:56	technomancy	the papers they have on it are pretty accessible too
22:56	technomancy	immutable package installation <3
22:57	callen	the weird package management/linux distro I was excited about was GoboLinux
22:57	callen	I was so sad it didn't take off :(
22:58	quidnunc	Only 2500 packages :(
22:58	xeqi	technomancy: have you tried nixOS?
22:58	callen	you people ask strange questions.
23:00	technomancy	xeqi: haven't taken the plunge yet; I get the feeling it doesn't have critical mass yet
23:01	technomancy	the millions of man-hours poured into Debian stable really count for something, and for stuff like hardware compatibility I don't think the immutable approach helps all that much; for that problem you just need a huge amount of effort put into finding all the edge cases.
23:03	callen	so, serious question. Who's thickey?
23:03	TimMc	Rich's brother, I believe.
23:04	callen	I'm incapable of seeing "hickey" without thinking of our glorious leader.
23:04	TimMc	Tom.
23:04	callen	oh lord, I was half-kidding.
23:04	callen	so he's a programmer too?
23:04	TimMc	I think so.
23:04	TimMc	ISTR he wrote some Clojure library...
23:05	callen	TimMc: http://tomhickey.com/ ?
23:05	technomancy	the enclojure logo is really great
23:05	callen	this twitter picture is hilarious.
23:07	callen	TimMc: well, guessing from his twitter, Rich's brother is quite the character.
23:08	Khaoz	Well.. good night guys.
23:09	Khaoz	thanks for the help. I will continue tomorrow :)
23:11	jodaro	so
23:12	jodaro	is clojurescript one still the best place to start for a not really front end web guy?
23:13	jodaro	i've been noir'ing but i figured i should see what all the kids are doing
23:14	callen	jodaro: clojurescript is a tower of abstraction you don't want to get into unless you already understand frontend.
23:14	callen	jodaro: stick with plain DOM + jQuery until you feel comfortable, then add a layer.
23:15	jodaro	callen: cool, sounds like a good idea
23:15	jodaro	i'm putting together a pretty basic app for a friends company
23:16	technomancy	is clojurescript one still maintained?
23:17	technomancy	I got the impression it was pretty wonky from a tooling perspective
23:18	callen	hasn't been touched in 9 months.
23:18	callen	I really think a single-language stack built on cljs for a web app is for people who are going to be comfortable unraveling the rest when it inevitably breaks.
23:19	callen	so many great clojure presentations, yet they're all on InfoQ
23:20	callen	I...really hate infoQ.
23:20	callen	technomancy: wait a second, how do you watch Hickey's talks if you don't use flash?
23:21	jodaro	yeah, i started to watch the clojurescript one video and started to get a little lost
23:22	jodaro	and then noticed it hasn't been committed upon for a while
23:22	jodaro	so was just wondering if there was a latest and greatest
23:22	jodaro	but
23:22	callen	jodaro: just don't.
23:22	jodaro	i'll just stick with noir+hiccup+twitter bootstrap
23:22	jodaro	which is what i've started with
23:22	callen	...close enough.
23:23	callen	I'd skip the bootstrap unless you need something tolerably attractive ASAP
23:23	callen	if your goal is to learn.
23:23	jodaro	i want it to be reasonably attractive, yeah
23:23	technomancy	callen: view source, occasionally preceded by M-x user-agent android
23:23	jodaro	because its an internal tool for non technical types
23:23	jodaro	but it doesn't have to do much
23:24	technomancy	a lot easier to work with than bootstrap if you don't know what you're doing
23:24	technomancy	for me anyway
23:24	jodaro	cool, i'll look at that too
23:24	callen	jodaro: internal tools are the best use-case for bootstrap, but just keep in mind that overriding bootstrap's somewhat idiosyncratic behavior can turn into more work than it saves you.
23:24	jodaro	bootstrap seems to be working for me so far
23:24	jodaro	i don't think i'm going to use a ton of features
23:24	callen	my frontend guy at my company has been flipping his wig over ripping bootstrap out for awhile now.
23:25	technomancy	yeah, I felt like I was fighting it every step of the way
23:25	callen	technomancy: that works on InfoQ?
23:25	callen	technomancy: setting your user-agent to android?
23:25	technomancy	callen: yeah
23:25	technomancy	one of the sites you have to grab it from the rss; I forget which
23:26	TimMc	hyPiRion: I've got a working skeleton for "rest" in curje.
23:26	TimMc	hyPiRion: https://gist.github.com/gists/3916042
23:26	callen	if only I'd known! :(
23:26	technomancy	I think one of the talks he gave at an oracle conf. isn't available for download, but the rest are easy
23:27	callen	technomancy: it says a lot that the state of "usability" for talks is so poor that you have to fight it to get at the content.
23:28	callen	how much trouble would I get in if I started just ripping and transcribing all of hickey's talks and putting them on a site?
23:28	callen	(I can't do audio, need text)
23:28	technomancy	callen: well, just follow the money
23:28	technomancy	getting a good AV setup doesn't come for free
23:28	technomancy	and it doesn't happen out of the goodness of their hearts
23:29	callen	so am I liable to get sued in htis hypothetical scenario?
23:30	jodaro	hypothetical federal prison
23:31	callen	wat.
23:31	callen	it's content that's freely accessible on the web, when did scraping become illegal?
23:31	aperiodic	callen: well, what's the usual answer to "what happens if I try to distribute copyrighted content for free in the United States"?
23:31	callen	it's freely accessible.
23:31	callen	just difficult to access.
23:32	aperiodic	doesn't matter
23:32	callen	well, guess I better host the servers in somalia.
23:32	aperiodic	it's still copyrighted, and they can still choose to exercise their copyright through a variety of legal avenues
23:33	technomancy	aperiodic: well, the video is copyrighted
23:33	callen	I expressly do not want the video.
23:33	callen	Despise the video/audio.
23:33	callen	Synchronous communication is slavery.
23:33	technomancy	I think the derivative work status of the words it contains is debatable
23:33	aperiodic	oh, sorry, missed the transcription bit
23:33	technomancy	probably depends on the wording of the release form he signed
23:34	technomancy	infoq sends "please don't do that" notices to blogs that explain how to do the user-agent/view-source download trick
23:34	technomancy	they don't seem overly hostile, but they do pay attention
23:35	quidnunc	callen: Hosting servers abroad didn't seem to work for Kim Dotcom
23:36	callen	No comment.
23:37	hiredman	infoq is great
23:37	hiredman	tons of content for free
23:38	callen	hiredman: I'm going to find it hard to take your word for it if you tell me a library is great in future :(
23:38	ChongLi	doesn't the ADA require video be accessible to the deaf and hard of hearing?
23:38	callen	that's my issue, I need text/image only.
23:38	ChongLi	I wonder if that'd put a wrinkle into the copyright issue
23:40	technomancy	doesn't apply broadly commercially
23:40	TimMc	Right, there could be a fair use defense.
23:40	TimMc	Maybe.
23:41	callen	reproductions of copyrighted content intended to enable accessibility by groups covered under the ADA are exempted from copyright liability/crime
23:41	callen	17 USC 121
23:41	callen	Section 1201 DMCA exception
23:42	callen	Video Accessibility Act 2010
23:42	callen	seems good enough to me.
23:42	callen	90% odds they rip my genitals off, but say sorry afterwards.
23:43	aperiodic	suddenly, the number of deaf/hard-of-hearing clojurians skyrockets
23:43	callen	aperiodic: makes sense to me, why else all the def and defn'd people?
23:43	TimMc	urgh
23:44	aperiodic	that was awful
23:44	aperiodic	(inc callen)
23:44	lazybot	⇒ 2
23:44	callen	aperiodic: thank you, I take great pride in that. You should try sharing an office with me.
23:45	aperiodic	there are few things better in life than truly awful puns
23:50	callen	What is good in life? To defeat your enemies and drive them before you, to hear the groans of their women at your bad jokes?
23:50	callen	s/\?/\!/g
23:59	TimMc	gfredericks, hyPiRion: Yay, I made 'rest in swearjure! https://gist.github.com/3916042